Independent Technical Assessment & SEO Oversight

Where things stand

The first Phase 1A review for Diamond Braces, KidCo and Smilify. A working portal, not a finished report: it stays live and grows as each phase is delivered. Start with the three minute version directly below.

Phase 1A · First reviewPrepared for Dr. Oleg DrutSystems read 8 and 9 Sep 2026Diagnostic · no systems changed

Phase 1ADiscovery & access manifestYou are here Phase 1BMeasurement auditNot started · what 1A surfaced Phase 1CBaseline assessmentNot started · what 1A surfaced Phase 4Independent SEO oversightRunning · open ObservationsConcerns across every phase, with your feedbackOpen

boltPhase 1A · the three minute version

Where do your patients come from? Today, nothing on the estate can answer that.

This is how far the trail goes before it stops. Each bar is a real figure read from GA4 over 28 days, 11 August to 7 September. The last bar is the finding: no system on the estate joins any recorded action to a consultation, a treatment start, or a named patient.

How far a marketing touch can be followed

Each stage is a count, not a rate. The trail stops at the fourth bar.

block

Ortho2 online bookings sit on a pending list unlinked to source and disappear once converted. Every call routes through one call center that logs the number dialed, not the reason for the call. So the fourth bar is zero by construction, not by omission.

What we know today

Four things the read established, in plain terms.

storefrontYour Business Profile listings account for 26.1% of all recorded booking actions from 9.7% of visits. Among the sources measured, they convert at the highest recorded rate. They are also the channel measured least completely.
public_offOnly 26.0% of the reported audience sits in the four states where you have an office, and about half of the recorded booking actions come from outside those rows, with no established cause.
visibility_offKidCo and Smilify are not being measured. One property recorded no data at all; the other recorded seven users in a week.
travel_explore1,350 of 2,249 known pages are reported not indexed by Google. Which of them should be indexed has not yet been established.

of reported active users are in NY, NJ, PA or CT

16,546 of 63,588

of recorded booking actions come from those four states

7,879 of 15,691

of pages Google knows about are indexed

899 of 2,249

how_to_voteWhat needs a decision from you

Smilify. The website sells an in-office visit at $2,800, which does not match the briefing. Nothing can be measured or mapped until the business model is stated. Observation 04
Data retention. GA4 keeps event history for two months. Every month it stays there is a month a measurement audit cannot look back on. Observation 03
Account ownership. No company-domain account appears in the measurement edit history. An administrator-led inventory is the first step. Observation 02
The dashboards decision. Building on today's feeds would publish these gaps with a chart on top. Observation 05
SEO capacity. Two specialists cover 75 listings, three websites and content. Narrow the scope, add capacity, or reset priorities. Observation 10
The platform move. A migration without a release gate repeats the pattern the last one is described as having had. Observation 11

Everything below this is the evidence behind those four statements, with its source, its read date and how far it goes. All twelve cross-phase concerns, each with a proposed action, are on the Observations page, where you can agree or disagree with each one.

insightsPhase 1A · the bottom line

What we can prove, and what we cannot prove yet

Read directly from the systems over four weeks, 11 August to 7 September, three things stand out: the reported audience is much larger than the four states an office sits in, the Business Profile channel is the least well measured of the sources compared here, and two brands have measurement gaps. Each of the six conclusions below is backed by an exhibit further down, with its source, its read date and how far the evidence goes. Where a conclusion rests on an inference rather than a screen, it says so.

Top channels by visits

Sessions by GA4 default channel group, 28 days. Volume only; hover for engagement and key events. What each channel does after the click is further down.

GA4, Diamond Braces, Traffic acquisition, 11 Aug to 7 Sep 2026, read 9 Sep 2026. The ten channel rows sum to 82,261 against a GA4 total of 82,086; GA4 computes the total row separately and the two never reconcile exactly. Certain

Key-event rate by campaign

Share of sessions that clicked through to Ortho2 or submitted a form, by session campaign: the four busiest listing slugs, the pooled Profile bucket, and the untagged campaigns. Read as measured today; it can change once tagging is standardized.

GA4, Session campaign, same period, read 9 Sep 2026. Site average marked. Certain

Sessions, Diamond Braces, 28 days

Of which 28.3% (23,229) are classified Unassigned, with three seconds of average engagement time

Business Profile session key-event rate against the site rate, three largest source rows

33.95% to 40.06% against 13.71%. Weighted 38.6% across those three rows, which hold 97.7% of consolidated Business Profile sessions. The rates of the remaining 182 sessions are not supplied.

Of active users are in NY, NJ, PA or CT

16,546 of 63,588 active users. Those four state rows carry 7,879 key events, 50.2% of the property total.

Of known Diamond Braces URLs are reported not indexed

1,350 of 2,249 in Search Console. Which of them should be indexed is not yet established.

CONCLUSION 01

The three largest Business Profile source rows record higher session key-event rates than the rows compared with them, and this is the channel measured least completely.

Across those three rows, 33.95% to 40.06% of sessions produced a key event, against a site rate of 13.71%, a Direct rate of 18.88% and a Google organic rate of 11.33%. One session in ten arrives from a listing; one key event in four is attributed to one. That traffic reaches GA4 under 29 different source and medium spellings and two campaign conventions, so no standard report shows it as a single channel, and a listing that links straight to Ortho2 booking loads no diamondbraces.com page, so website analytics record nothing for it. The comparison covers the rows named and the period read; it is not a verdict on the channel.

So: the tagging and the measurement gap need resolving before any channel comparison is treated as settled. What that work takes in effort, dependencies and testing has not been assessed here.

CONCLUSION 02

26.0% of reported active users were assigned to the four service-area states, and about half of the recorded key events fall outside those rows.

Of 82,086 sessions, 23,229 are classified Unassigned with three seconds of average engagement time. Of 63,588 active users, 16,546 were assigned to the four states an office sits in; the remaining 47,042 includes users GA4 could not place, so it is not all confirmed out-of-area. Of 15,691 key events, 7,467 (47.6%) were attributed to US users outside those four states, and the property remainder outside the four-state rows is 7,812 (49.8%). Several out-of-area states show reported user key-event rates that overlap the service-area range, which spans 24.06% to 27.96% for New Jersey, Pennsylvania and Connecticut, while California, Texas and Illinois sit under 6%. A key event is a recorded click through to Ortho2 or a form submission, not a booking, and none of this establishes the size of new-patient demand.

So: a target or a vendor claim resting on the top-line session or key-event number rests on figures whose composition is not explained. The out-of-area pattern needs a cause before any conversion total is quoted, and the segments worth managing should be validated before they are adopted as KPIs.

CONCLUSION 03

Two brands have measurement gaps, and one of them also has an unconfirmed business model.

The KidCo property showed no recorded data in the property inspected. Smilify's property recorded seven users, 24 events and no key events in the stated week, no client-side analytics request was observed on its homepage, and no Search Console property for it was visible in the accessible account. No Tag Manager container for either brand was visible in the account read, though Tag Manager is not required to deploy analytics. The Smilify website describes an in-office visit at $2,800, which does not match how the brand was described in the briefing.

So: neither brand can currently be reported on from these properties. Whether the tags are absent, carry a different measurement ID, or report somewhere else has to be checked at the site and in any other accounts; the Smilify business model is a question only leadership can settle.

CONCLUSION 04

A large amount of measurement work sits drafted and unpublished, with no recorded owner.

The Tag Manager workspace holds 11 unpublished changes, the oldest about two years, and no publication was observed after December 2024. Four of the six configured key events recorded no data in the period read. The internal-traffic filter sits in Testing, so it does not exclude matching traffic. Displayed Search Console validation statuses vary, and the 18-URL discovered-not-indexed row shows Passed. Event data retention is set to two months, which limits historical exploration and funnel work; standard aggregated reports are not affected.

So: the workspace state is recorded, not diagnosed. Who owns those changes, what status each was meant to reach, and what the retention setting should be are decisions for the client. Proposed to the sponsor: name a single owner for measurement releases and set a release routine, which Phase 4 can then review.

CONCLUSION 05

Search Console reports 1,350 known URLs as not indexed, and the reasons have not been triaged.

Google knows 2,249 Diamond Braces URLs and reports 899 indexed. The largest classifications are 715 redirects, 165 crawled but not indexed, 152 alternate pages with a canonical, 146 duplicates where Google chose a canonical and 95 not found. Intentional redirects and canonicals are not defects, so the first step is identifying which URLs should be indexed. The team has since narrowed its indexing focus to 41 priority pages. The exact query "diamond braces" accounts for 88,104 clicks in 16 months, about 22.6% of reported clicks; a complete branded and non-branded classification has not been produced.

So: the decline discussed with the team sits outside the 16 months Search Console retains, so it cannot be shown from this data. Establishing which URLs should be indexed, and obtaining the pre-Shopify analytics property, both sit with the client.

CONCLUSION 06

Nothing today connects a marketing touch to a patient.

Ortho2 online bookings sit on a pending list unlinked to source and disappear when converted. All calls land in one call center that logs the number dialed but not where the caller came from. The primary "Get Started" form posts to a system with no analytics on it. 43.9% of consolidated Business Profile sessions carry no listing identity, and a listing that links straight to Ortho2 booking loads no page on the site, so website analytics record nothing for it.

So: until a booking or a call can be traced back to a listing, a page or a campaign, every performance conversation is about traffic rather than patients. Phase 1B assesses whether that attribution is feasible, documents the limits and specifies any remediation, which is authorized separately.

blockWhat cannot be concluded yet

Bookings by sourceNo system joins a session to a consultation or a treatment start. Any figure claiming otherwise is an estimate.
Phone attributionPhone is one of the primary booking routes and no call can be tied to a listing or a page. Its share of bookings is unknown.
SmilifyIts business model, its analytics and its booking flow are all unconfirmed. Nothing about it should be reported as fact.
The declineThe traffic drop the team describes sits outside the 16 months Search Console keeps, so it cannot be shown from this data and its cause cannot be attributed here.
Office-level ROI68 campaign slugs identify a location, but the slug-to-listing crosswalk is not confirmed, 43.9% of Business Profile sessions pool under the word Profile, and booking-linked listings bypass the website. Per-office figures stay partial until all three are resolved.

searchSurfaced in the SEO working session, 8 September

Listing linksThe team described its listings as split between website links and direct booking links, with the booking-linked ones booking more, and said it intends to move all of them to booking links.
IndexingThe team said it has narrowed its indexing focus from roughly 1,000 pages to 41 priority pages.
Reviews239 unreplied reviews, against a target of 50 set by Dr. Drut. Comment-free one-star reviews are being disputed.
CapacityTwo people, plus one person for reviews and social, cover 75 listings, three websites and content production.
Full recordEvery discovery, the evidence behind the gaps, and the weekly session log are on the Phase 4 Oversight page. Concerns that cut across phases are on the Observations page, each with what to do about it and a place to leave feedback. Nothing on the Oversight page is a recommendation yet.

flagThe engagement

One portal for every phase

This is the living record for the whole engagement, not a one-off Phase 1A document. Phase 1A is below in full. Phase 1B and Phase 1C each have a page that holds what 1A has already surfaced for them; they are marked not started, and they fill in only once each is authorized in writing. Phase 4 oversight and the cross-phase Observations page are open now. Nothing here commits Diamond Braces to 1B or 1C.

Phase 1AYou are here

Discovery and access manifest. What each property is, what it measures, and where it is blind. Delivers the conversion path map, the traffic quality analysis and the access manifest. Under the SOW, delivery opens a three business day review window. The version submitted for acceptance, the start of that window and its deadline are not yet confirmed with the sponsor; this build is a working draft. Outstanding in 1A: the slug-to-listing crosswalk, a list of which listings link to booking rather than the website, the Ortho2 and NICE answers, and a line-by-line read of the four weekly reports. Phases 1B and 1C are neither authorized nor invoiced until the sponsor approves proceeding in writing.

Current

Phase 1BNot started

Measurement audit. Whether GA4, Search Console, tag management, cross-domain tracking and conversion tracking are configured correctly, and which KPIs become the source of truth. Delivers the instrumentation audit, handoff assessment, phone attribution assessment, starts attribution model, KPI definitions and a separately priced remediation plan. What 1A has already surfaced for it is on the Phase 1B page.

Not authorized

Phase 1CNot started

Baseline assessment. The objective current condition of all three properties, recorded with date, source and method so it can be measured again: technical baseline, authority baseline, work-in-flight inventory, coverage gaps, and one baseline document. Runs in parallel with 1B. What 1A has already surfaced for it is on the Phase 1C page.

Not authorized

Phase 4Running now

Independent SEO oversight. A capped monthly retainer from 1A kickoff, measured against the 1B KPIs and the 1C baseline once they exist. Weekly working sessions with the SEO team began 8 September; what each surfaces lives on the Oversight page.

Running

Review log

Each sponsor review of this portal, what was shown, what was asked, and what was taken away. Working discussion stays out.

9 Sep 2026First Phase 1A review with IT; the sponsor to review the portal separately. Shown: the portal at v2.7, the top-channel read, the Business Profile finding, the out-of-area key events, the KidCo property with no data, and the toxic link profile. Phase 1A described as about 85% complete, closing by the end of the week, with a full review meeting early the following week.
AskedHow the portal can say that listings linking straight to Ortho2 booking perform better when that click never touches the website. Recorded here so the portal and the conversation match: that statement is the SEO team's own report from 8 September and is graded client-supplied throughout this portal. It has not been independently verified. The check still to be done is one screen: whether a directly-linked listing shows a Bookings row in its Business Profile Performance report. Google documents that metric as completed bookings, available only where bookings run through an integrated booking provider, and documents its separate Booking clicks metric as applying to hotels only. So either these listings already record completed bookings, in which case the team's report is verifiable there and can be compared with Ortho2 for the same period, or they carry a plain link, the interaction counts as a website click, and no booking metric exists for them at all. Joining whatever exists to Ortho2 is Phase 1B work. What GA4 does show, and this portal does say, is that the three largest website-linked Business Profile source rows recorded session key-event rates of 33.95% to 40.06% against a site rate of 13.71%.
Confirmed by ITKidCo does not appear in Tag Manager; only Diamond Braces does. Earlier development teams had tried to address toxic links when the site was on the custom platform, without a dedicated SEO function at the time.
Next stepsPortal credentials sent to IT and the sponsor. IT and the sponsor review the findings and use the feedback controls or the chat to challenge anything. KidCo analytics and Tag Manager configuration to be verified and corrected by IT. The toxic link profile to be analyzed from the tool with history before anything is removed. Review meeting the following week.

Meeting notes and transcript, 9 Sep 2026, 12:00 ET. Client-supplied on what IT confirmed.

Diagnostic throughout: every phase produces analysis, not implementation; findings are source-graded; no change is made to any client system; the review is of the properties, not of any person or team; and each phase ends at a real decision point.

domainInventory

Three properties at a glance

Three brands, three GA4 properties, one Tag Manager container, two Search Console properties. The gaps in this comparison are the first finding.

Diamond Braces

diamondbraces.com

PlatformShopify

GA4 property429695085

Measurement IDG-P58HF4HQZK

Data streams1 live, 1 dead (get.diamondbraces.com)

Tag ManagerGTM-KSXTJSZ

Search ConsoleYes

GBP listings69 (68 open)

Sessions, 28 days82,086

checkMeasured Configuration defects

KidCo Dental

kidcodent.dental

PlatformShopify

GA4 property326433410

Measurement IDG-JNNWTJMJ6Z

Data collectedNone recorded in the property inspected

Tag ManagerNone

Search ConsoleYes (9,190 clicks in 16 months)

GBP listings6 (5 open)

Offices on website5, against 6 listings

closeNo data recorded Website and GBP disagree

Smilify

smilify.com

PlatformCustom build

GA4 property399685631

Measurement IDG-2BGW5RHH82 (unconfirmed)

Data collected7 users, 24 events, 0 key events in 7 days

Tag ManagerNone

Search ConsoleNo property

GBP listingsNone identified in the accessible material

Offer on siteIn-office visit, $2,800

closeEffectively unmeasured Business model to confirm

storefrontFinding 01

Channel performance, as measured so far

Volume and behaviour after the click are different questions. By volume, organic search and direct lead. By what happens after the click, the three largest Business Profile source rows recorded session key-event rates of 33.95% to 40.06% against a site rate of 13.71%, and engagement rates above the rows compared with them. That reading is only as good as the tagging underneath it, which is scattered across 29 spellings, and it covers only the listings that link to the website. Treat the comparison as scoped to those rows and the measurement gap as the finding.

7,973

GBP sessions, 28 days, consolidated from 29 tag variants

9.7% of 82,086. Third-largest identifiable source after Direct and Google organic. Website-linked listings only.

69%

Engagement rate, weighted across the three largest Business Profile source rows

Those rows hold 7,791 of 7,973 sessions, 97.7%. Google organic 58.01%. Direct 43.58%. Site 40.67%. The remaining 182 sessions are not rated here.

2.5x to 2.9x

Session key-event rate against the site rate, three largest Business Profile source rows

33.95% (GMB / (not set)) to 40.06% (GBP / organic) against 13.71%. Consolidated Business Profile traffic carries 4,102 key events, 26.1% of the property total.

Key-event rate by session campaign

Share of sessions that produced a key event (Click to Ortho2 or form_submit). Site average marked. Business Profile campaigns in blue.

GA4, Diamond Braces, Traffic acquisition by Session campaign, 11 Aug to 7 Sep 2026, read 9 Sep 2026. Certain

One channel, 29 source and medium spellings

Sessions by source / medium, exactly as they arrive in GA4, for traffic that originates from a Business Profile. Blue is the one row matching the intended convention, utm_source=gbp with utm_medium=organic; red is every other spelling. Raw values are preserved for provenance.

GA4, Session source / medium, all 154 rows read 9 Sep 2026. Five spellings of the source (GBP and GMB in lower case, GMB in capitals, Google, and blank) and 21 rows where a location name sits in the medium field. GMB / (not set) is malformed: a source with no medium. Certain

Engagement rate by source

Top sources by sessions. GBP rows highlighted.

GA4, Session source / medium, 28 days. Certain

Why it is invisible: two tagging conventions

	Convention A	Convention B	Convention C
utm_source	`gbp`	`GMB`	`GMB` or blank
utm_medium	`organic`	`organic` or missing	a location name, e.g. `Denville-NJ`
utm_campaign	location slug, e.g. `clifton-nj`, `3rd-ave-ny`	the literal word `Profile`	none
Sessions (28 d)	4,363	3,428	119
Attributable to a listing	By hand-mapping 68 slugs; the crosswalk is unconfirmed	No, the campaign value carries no listing	Sometimes, from the medium field

A further 63 sessions arrive under stray spellings (gmb / organic, Google / Profile, GMB / referral). The Profile campaign holds 3,500 sessions and the Convention B source rows 3,428; these are two separate marginal reports, so their overlap and which listings sit in each need a cross-tabulation rather than an assumption. Standardizing the listings on one convention would resolve the spellings at source; the effort, dependencies and testing for that change have not been assessed.

GA4 Session campaign (108 values) and Session source / medium (154 values), 28 days, read 9 Sep 2026; Bronx White Plains Rd listing URL observed directly. Certain

info

What this means so far. The channel the SEO team already manages recorded higher session key-event rates than the other source rows shown here, for the period read. The team reported a related pattern on 8 September: that its booking-linked listings book more than its website-linked ones. That report is client-supplied and unverified, and those listings are exactly the ones website analytics cannot see. The measurement gap is structural rather than a reporting choice: 29 tag spellings, two campaign conventions, no Business Profile product link into GA4, and no route from a listing session to an Ortho2 record. Until those are resolved, no channel ranking should be treated as settled.

person_searchFinding 01, continued

Where booking actions start

The question behind every marketing conversation is "where do our patients come from." Today no system can answer it: Ortho2 drops the source when a booking is converted, and the call center logs the number dialed but not what led to the call. The closest measure that exists is a key event, a recorded click through to Ortho2 or a form submission. This is that, read exactly from GA4, by source and by listing. Two cautions travel with every figure: about half of the key events fall outside the four-state rows and a click is not a booking, and the same person can be counted more than once. Phase 1B assesses whether patient-source attribution is feasible, documents the limits and specifies any remediation, which is authorized separately.

9.7%

of visits come from Google Business Profile

7,973 of 82,086 sessions, website-linked listings only

26%

of key events are attributed to Business Profile tagging

4,102 of 15,691. Direct is 46%; Google organic search 23%.

campaign slugs identify a location

Against 69 Diamond Braces listings. The crosswalk is not confirmed: two matches are inferred and one is ambiguous. 43.9% of Business Profile sessions still pool under the word Profile.

Key events by source

Recorded clicks through to Ortho2 (14,389) and form submissions (1,302), 28 days, 15,691 in total. Direct is largest and includes existing patients reaching the portal and login; the Unassigned traffic that is 28.3% of sessions produces under 2% of key events.

GA4 Session source / medium with the Key events column, all 154 rows, 11 Aug to 7 Sep 2026, read 9 Sep 2026. Rows are grouped here; the 29 Business Profile spellings are one bar. Rows under two sessions are omitted, so the groups sum to 15,690. Certain

Key events by listing, Business Profile traffic only

The pooled Profile bucket against the fifteen busiest location slugs. The full list of 68 is below. Everything under Profile carries no listing identity.

GA4 Session campaign, all 108 values, 28 days, read 9 Sep 2026. Key events read directly. Slugs are mapped to listings by name; bronx-ny and gwb-scan-ctr-ny are inferred matches and flatbush-ny could be either of two Flatbush Avenue offices. Certain on the figures, Likely on the slug-to-listing crosswalk, which is not yet confirmed.

ruleWhat this read changed

The first pass read the ten busiest campaign rows and reported that only three listings could be read individually. The full list of 108 campaign values, read on 9 September, shows 68 location slugs. The earlier statement understated it and has been withdrawn. The derived booking-action figures from that pass have been replaced by the key-event counts GA4 shows directly. What has not changed: 3,500 Business Profile sessions still pool under Profile, the slug-to-listing crosswalk is still unconfirmed, and booking-linked listings are still outside website analytics.

link_offWhat stands between this and "patients by source"

One tagging convention across all 69 listings so the Profile bucket disappears. A confirmed crosswalk from each slug to a listing. A record of which listings link to booking rather than the website, so the unmeasured share can be sized. A join between an Ortho2 booking record and the session that produced it, which needs a persistent identifier that has not been shown to exist; the retained catalog codes may support office and patient-type reporting, but they do not identify a marketing source or a unique session. And a way to attribute calls, since phone is one of the primary booking routes. Assessing those is what Phase 1B is for.

Every readable listing, 28 days

All 68 campaign slugs that identify a listing, sorted by key events. Sessions and key events are read directly; the rate is the share of sessions with at least one key event. Small numbers move a lot week to week; read the rates on rows under 50 sessions with that in mind.

#	Listing	Campaign slug	Sessions	Key events	Session rate

GA4, Traffic acquisition by Session campaign, 11 Aug to 7 Sep 2026, read 9 Sep 2026. Slug sessions total 4,343 and key events 2,427; the pooled Profile campaign adds 3,500 sessions and 1,675 key events with no listing identity. Certain

filter_altFinding 02

Traffic quality: what the headline number is made of

82,086 sessions in 28 days. 28.3% of them are classified Unassigned with three seconds of average engagement time, 26.0% of active users were assigned to the four states an office sits in, and about half of the key events fall outside those four state rows, some at rates overlapping the service-area range. What that mix means for new-patient demand is not established by these figures.

Sessions by default channel group

The Unassigned slice is 23,229 sessions GA4 could not match to a channel definition, with three seconds of average engagement time and a 0.92% session key-event rate.

GA4, Traffic acquisition, 11 Aug to 7 Sep 2026, read 9 Sep 2026. The same unsourced traffic reads 23,229 as a channel group, 23,177 as a source / medium and 23,244 as a campaign; the three cuts differ by under 70 sessions. Certain

Engagement rate and time by channel

Unassigned engages 10.34% of the time for three seconds of average engagement. Automation and instrumentation faults are both plausible; neither has been established.

GA4, Traffic acquisition, same period, read 9 Sep 2026. Certain

Inside vs outside the service area

Active users, 28 days. Service area defined as NY, NJ, PA and CT, the four states with a listing. The outside slice includes 5,288 users whose region GA4 did not set.

GA4, User attributes, Region, 28 days, read 9 Sep 2026: 16,546 of 63,588 active users in the four states. Certain

Top regions by active users, with engagement

Colorado is the second-largest region on an orthodontics site with no Colorado office: 7,941 active users, 3.3% engagement, 1 second, 57 key events.

GA4, Region, 28 days, read 9 Sep 2026. Service-area states highlighted. Certain

Several out-of-area states show key-event rates overlapping the service-area range

GA4-reported user key-event rate, which GA4 calculates on total users; active-user counts are shown separately in the tooltip. Blue is the service area. Red is a group of states with no office whose displayed rate overlaps the service-area range. Gray is the pattern more usual for out-of-area traffic.

GA4, Region with Key events, 28 days, read 9 Sep 2026. The rows shown are reproduced from the report; a full seventeen-state breakdown has not been documented and is not retained in this portal. New Jersey, Pennsylvania and Connecticut span 24.06% to 27.96%. Click to Ortho2 is 14,389 of the 15,691 key events. Certain on the displayed figures. Guessing on the cause: automation, a trigger firing on something other than a hand-off to Ortho2, and a real unaccounted population are all untested, and more than one may apply.

priority_highWhy this matters more than the Unassigned traffic

The Unassigned 28.3% is easy to set aside: it engages for three seconds and produces under 2% of key events. This population behaves differently. It reads as engaged on every displayed measure and it produces recorded key events. A dashboard built on GA4 key events today would report 15,691 conversions, of which about half fall outside the four-state rows, some at rates overlapping the service-area range. Nothing in GA4 establishes whether any of them relate to completed bookings. Ortho2 holds consultation and start counts by location, subject to administrator access, and comparing the two is one of the first checks Phase 1B would specify. No conversion total, conversion rate or cost per conversion should be quoted from this property until the cause is known.

Business Profile traffic is not exempt. Its 4,102 key events arrive through the same event definitions, and no geography validation has been run on them, so it needs the same checks as every other source.

Countries

Active users by country. Singapore shows more active users than New Jersey, with six seconds of average engagement.

GA4, Country, 28 days, read 9 Sep 2026. United States 49,059 users, 97.8% of key events. Certain

The headline number, layer by layer

Layer	Count	Share
Reported total, sessions	82,086	100%
Unassigned, no source, 3s average	23,229	28.3%
Active users, reported total	63,588	100%
Active users outside NY, NJ, PA, CT (incl. unset region)	47,042	74.0%
Active users inside the service area	16,546	26.0%
Key events, whole property	15,691	0.25 per user
Key events, United States	15,346	97.8% of all
Key events, US users outside the four states	7,467	47.6% of all
Key events, property remainder outside the four-state rows	7,812	49.8% of all
Key events, four service-area states	7,879	0.48 per user

Half of all key events sit in the four state rows that hold about a quarter of the active users. The two remainder figures differ by 345 because one is scoped to the United States and the other to the property, so they are not the same population and should not be equated. Any future baseline should be built from validated segments rather than the property totals; the segments themselves need testing before they are adopted, which is the subject of the next card.

GA4 Region and Country rows, read 9 Sep 2026 (NY 5,990 + NJ 1,074 + PA 616 + CT 199 key events; United States 15,346). Certain on the rows; Guessing that the 5,288 unset-region users are largely the same Unassigned traffic, which has not been tested.

groupFinding 04

Existing-patient activity is mixed into the traffic figures

Three independent lines of evidence show existing-patient activity in the numbers. None of them measures its share of users, sessions or key events, so the size of the effect is not established.

Top pages by views, 28 days

Two missed-appointment policy pages rank 3rd and 7th. Together they draw 2.3 times the views of the pricing page.

GA4, Pages and screens, 118,186 views across 1,023 pages, read 8 Sep and re-read unchanged 9 Sep 2026. Certain

The three signals

Pages5,683 views on /pages/missed-adjustments and /pages/missed-appointments combined (4.81% of all views) against 2,429 on /pages/pricing.
Searchdiamond braces patient portal 3,472 clicks at 86.4% CTR; diamond braces login 2,419 clicks at 84% CTR; diamond braces appointment 1,942 clicks at 71% CTR. All three contain the brand name, so all three are branded queries under the definition used here. Portal and login point to existing-patient tasks; an appointment search can equally be a prospect. The patient-status mix has not been established. (Search Console, 16 months.)
Booking modalThe homepage Book Now flow offers two general Ortho2 catalog codes, New Patient epmcoa and Existing Patient itozgj, while each office page carries its own three codes for new, existing and reschedule. All of them fire the same Click to Ortho2 key event, so the count mixes prospects with current patients. Every code in use has to be inventoried before the event can be split.
Locator duplicate/pages/dentaloffices-nyc-pa-nj-ct-1 (6,583) and /pages/dentaloffices-nyc-pa-nj-ct (2,201) are the same office locator on two URLs. Combined, 8,784 views, the second-largest destination on the site.

GA4 Pages and screens; GSC Performance 16 months; site walk 8 Sep 2026. Certain

route1A.1

Conversion path map (1A.1)

Every route from a website visitor to a booking, a call or a form, per brand, as observed on 8 September 2026. Green endpoints are measured. Red endpoints are not.

Diamond Braces

diamondbraces.com (Shopify)

Get Started (main site)→engage.diamondbraces.com/f/virtual-start→No analytics on form

Get Started and Book a consultation (blog posts)→Typeform virtualstart2→Cross-domain configured; outcome unmeasured

Book Now→/pages/dentaloffices-nyc-pa-nj-ct-1 (choose an office)→Office page, Book Now again→Modal: New patient, Existing patient or Reschedule→Ortho2, office-specific code

Four clicks from the header to the Ortho2 page, with the office chosen by the visitor along the way; a Business Profile listing that links to booking reaches the same page in one. Each office page carries its own three Ortho2 codes (Clifton: kdkima, cqeuht, wzsjsz), so Ortho2 receives the office and the patient type from the link even though GA4 records one event for all of them. The homepage modal uses the general codes epmcoa (new) and itozgj (existing). Re-walked 9 Sep 2026.

Call Now→tel:201-308-8181→NICE CXone, no source

Location page→Office phone, e.g. 856-210-8560→Unattributed

GBP listing (most)→/pages/<slug>/?utm_source=gbp…→Website, then Book Now

GBP listing (some)→Ortho2 booking directly

Ortho2 catalog 43b5890c-d058-47ef-abb2-06c8b6390eb8 at edgebooking.ortho2.com. Cross-domain measurement is configured for ortho2.com and typeform.com. Typeform is used by blog-post calls to action (Book a free consultation, and the Get Started button on blog templates), so the site has two different Get Started destinations depending on the template. Location pages exist under three URL patterns (/pages/bronx-ny/, /pages/west-berlin-nj-office, /locations/…).

Site walk and network traces, 8 Sep 2026. GBP link targets: one listing observed directly; the mix of website and booking links described by the SEO team on 8 Sep 2026, which said it intends to move all listings to booking links. Certain Client-supplied

KidCo Dental

kidcodent.dental (Shopify)

Header Book→Ortho2 code eflvpd

Location cards→Bay Ridge ggwzzt · Bed Stuy tdgpna · Flatbush eflvpd · Williamsburg hgocwx

Jackson Heights→On-site page

Phone→(929) 249-4440, all offices→Unattributed

The header link and the Flatbush card share one code, so header bookings are pooled with Flatbush. The website lists 5 offices; Business Profile has 6 listings. The website and Business Profile disagree on both postal codes: Bay Ridge shows 11226 on the site against 11209 in Business Profile, and Flatbush shows 11206 against 11226. Bay Ridge's website value matches Flatbush's Business Profile value; Flatbush's website value is a third number. This is not a reciprocal swap.

Site walk 8 Sep 2026; GBP dataset 7 Sep 2026. Nothing here reaches GA4 because the property has no data. Certain

Smilify

smilify.com (custom)

In Office Visit, $2,800→Zip entry→Closed without result (x2)

Am I a candidate?→Unmeasured

Phone→(862) 330-1500→Unattributed

Site walk and network trace, 8 Sep 2026. Certain on what was observed; zip flow behaviour to be confirmed manually.

Phone numbers observed across the three sites

Number	Where	Attributable to marketing source
201-308-8181	Diamond Braces main Call Now	No
856-210-8560	Diamond Braces West Berlin NJ page ("Coming Soon")	No
929-388-4488	Diamond Braces Bronx White Plains Rd (GBP)	No
929-249-4440	KidCo, all offices	No
862-330-1500	Smilify	No

All calls terminate in NICE CXone, which logs the dialed number but not the marketing source (client-supplied, 31 Aug 2026). Client-supplied

Can the same patient be counted twice?

The conversion path map is required to say, for each route, whether one person can appear in more than one count. Three separate risks are distinguished here: repeat counting of the same event, overlapping records across systems, and mixed patient types inside one measure. Different funnel stages are not duplicate patients unless they are added together as though they were.

Where it happens	How	Which counts overlap
Book Now, new versus existing patient	Every Ortho2 catalog code fires the same Click to Ortho2 key event.	Mixed patient types inside one measure. A classification problem, not a duplicate count.
Online booking and a phone call	A patient who books online and also calls is on the Ortho2 pending list and in the NICE call log, with nothing linking the two.	Ortho2 and NICE each count one; neither knows about the other.
Business Profile click, then Book Now	The listing click is recorded in Business Profile Insights; the session is recorded in GA4; the booking in Ortho2.	Three systems, three counts, no join.
Get Started, then Book Now	A form submission lands in the engage CRM (main site) or a Typeform (blog); a later booking lands in Ortho2.	Two records for one person, in two or three systems, never reconciled.
A return visit	A visitor arriving first from a listing and later typing the address is a Business Profile session and a Direct session.	The same person is two sessions in two channels.

Site walks 8 Sep 2026; client statements on Ortho2 and NICE, 31 Aug and 1 Sep 2026. Certain on the mechanics; how often it happens cannot be measured until the systems are joined. Client-supplied

mapService area

Every Business Profile listing across NY, NJ, PA and CT

75 Business Profile listings transcribed from GBP Manager on 7 September 2026 (no export exists), placed by postal code on a map drawn from U.S. Census county boundaries. Hover a pin for the listing and click it to run a Google Maps search for that office. Filter by brand or group. Two listings are not open.

Diamond Braces (69)KidCo Dental (6)Not openDetail area, right

New York City, Hudson County and lower Westchester detail. 57 listings.

Listing register

Brand	Group	Address	City	State	Zip	GBP	Status	Google

Each row opens a Google Maps search built from the listing name and address. These are generated search links, not verified listing URLs; verified links need each listing's Place ID. GBP Manager screenshots, 7 Sep 2026, 75 rows. Group counts matched every group header; no duplicate store codes. Store codes are unverified transcriptions; addresses are the reliable field. Pins sit at postal-code centroids and are offset slightly where several listings share a zip, so they show the area of each office, not the door. Boundaries: U.S. Census Bureau cartographic boundary files (counties, 1:10m). Certain on count and addresses.

location_onFootprint

GBP listings by group, brand and state

The footprint as GBP sees it. This is the yardstick for the service-area analysis above and the population that a single tagging convention needs to cover.

Listings across 11 GBP groups

69 Diamond Braces, 6 KidCo

Open

Mt Laurel NJ permanently closed (still published). KidCo 793 Manhattan Ave temporarily closed.

Verified

1 listing shows Published rather than Verified. Nearly every listing carries a red indicator on its edit icon, recorded without a diagnosis; its meaning is unconfirmed.

Listings by GBP group

Groups as named in GBP Manager.

GBP Manager, 7 Sep 2026. Certain

Listings by state and brand

New York holds 54 of 75. Connecticut has one.

GBP Manager, 7 Sep 2026. Certain

info

Cosmetic inconsistencies noted for the 1C baseline, not assessed here. Two Queens listings use the name order "Invisalign & Braces" where the other 67 use "Braces & Invisalign". KidCo casing varies between the literal spellings Kidco and KidCo. A permanently closed office remains published. Listings are split between website links and direct Ortho2 booking links; the SEO team said on 8 Sep 2026 that its booking-linked listings book more and that it intends to move every listing to a direct booking link.

settings_input_componentFinding 05

Instrumentation: what fires, what is drafted, what is unpublished

The Diamond Braces property is measured, and five separate pieces of measurement work sit in a drafted or unpublished state. This records the observed condition of the property. Who owns each piece, and what status it was meant to reach, is not established here.

Key events configured on the Diamond Braces property

Two recorded data in the period read: Click to Ortho2 (14,389 in 28 days) and form_submit (1,302). Four recorded none. Purchase records nothing and is not marked as a key event. Configured, published and observed-firing are three different states; the table separates them.

Event	Key event	Firing	Corresponding GTM tag
Click to Ortho2	Yes	Yes	Live (v25). A modified version sits unpublished in the workspace.
form_submit	Yes	Yes	Enhanced measurement / Form Submission trigger; trigger modified, unpublished.
EmailClick	Yes	No data in period	Tag and trigger added about 1 year ago; not in the published container.
GetStartedClick	Yes	No data in period	Tag and trigger added about 1 year ago; not in the published container.
PhoneClick	Yes	No data in period	Tag and trigger added about 1 year ago; not in the published container.
Scroll	Yes	No data in period	Tag and trigger added about 1 year ago; not in the published container.
purchase	No	No data in period	None. $0 revenue recorded on the property.

GA4 Admin, Events and Key events, read 8 Sep 2026; counts from Traffic acquisition, 11 Aug to 7 Sep 2026, read 9 Sep 2026; GTM workspace and live version 25 of GTM-KSXTJSZ, read 8 Sep 2026. Behaviour outside the period read was not assessed. Certain

Tag Manager container GTM-KSXTJSZ

Account 6074672983, container 103537973. The only container visible in this account; none for KidCo or Smilify. Other Tag Manager accounts were not accessible.

Unpublished workspace changes

9 added, 2 modified

624

Days since last publish

v25 "GW 0.3", 23 Dec 2024

Live versionv25 "GW 0.3", published 23 Dec 2024. 5 tags, 3 triggers, 6 variables.
Unpublished versionv26 "GW v0.4" created 623 days ago; no publication of it was observed.
DestinationsTwo Google tags: DB Main Site (G-P58HF4HQZK / GT-NGSCZKJV) and G-2BGW5RHH82, whose owning property is not identified. 6 tags each.
EditorsFour accounts with edit history: two personal Gmail addresses, one external vendor domain, one shared Gmail that is also the GBP primary owner. Company-domain accounts: none observed.
TimelineForm Submission trigger and HubSpot cookie variable modified 2 years ago; four click and scroll tags added about 1 year ago; Click to Ortho2 modified more recently. None appear in the published container.

GTM workspace and version history, read 8 Sep 2026. Edit history is not evidence of current permissions, and an email domain is not proof of account ownership. Recorded as a property condition, not an assessment of any individual. Certain on what the screens show.

GA4 configuration defects

Data retention2 months Historical exploration and funnel work is limited to that window; expired event-level data is removed through monthly processing, and standard aggregated reports are unaffected. 14 months is the longer option available.
Internal traffic filterTesting Defined but not active, so it does not exclude matching traffic. Its rules and the volume affected are unconfirmed.
Cross-domainortho2.com and typeform.com configured. Ortho2 confirmed in use; Typeform is used by blog-post calls to action, while the main site sends Get Started to engage.diamondbraces.com.
Enhanced measurementOn: page views, scrolls, outbound clicks and four others. URL query parameter redaction off.
Product linksNone Search Console, Google Business Profile and Google Ads all unlinked.
Dead streamget.diamondbraces.com stream (10035584299) still attached to the property.
Access tierLikely Editor, not Administrator (no Property access management entry rendered), so the administrator list was not visible. Satisfies the SOW; confirm with an administrator.

GA4 Admin, read 8 Sep 2026. Certain except access tier.

Google tag coverage

Tag quality reads "Needs attention". 1,655 pages seen.

The report classifies these as untagged: /locations/, two /locations/… office pages, /collections/braces, /healthform, two Invisalign pages, /pa and /pages/terms-of-service. Classified as no recent data: /appointment, /book-online and //checkout/. These are the report's classifications; the live URLs have not been re-tested.

Google tag, Tag coverage, read 8 Sep 2026. Certain

Drafted, not activated

Five independent pieces of measurement work, each configured but not in a live state.

GTM11 workspace changes unpublished, oldest ~2 years.
GTMVersion 26 created 623 days ago; no publication of it observed.
GA4Internal traffic filter created, left in Testing.
GSCDisplayed validation statuses vary; the 18-URL discovered-not-indexed row shows Passed. Two unused verification tokens.
GBPRed indicator on nearly every listing's edit icon. The client reads it as unpublished edits. Its meaning is unconfirmed and it is recorded without a diagnosis.

The pattern across all five is the same: work configured, not published, and with no owner recorded. Whether that reflects a decision, a handover or an oversight is not established from these screens. Proposed to the sponsor: name one owner for measurement releases and adopt a publish-and-verify routine; Phase 4 can review that routine and its evidence, and client management would operate it. All 43 register items, which mix confirmed observations, suspected defects and questions needing validation, are itemized in the register on the Phase 1B page, which is the working list a measurement audit would be scoped from.

travel_exploreFinding 06

Search Console: indexation and demand

Google knows about 2,249 URLs on diamondbraces.com and indexes 899 of them. Demand is dominated by the brand name. KidCo ranks nationally for toothpaste questions, not locally for dentistry. Smilify has no Search Console property at all.

Diamond Braces: indexed vs not indexed

899 indexed (40.0%), 1,350 not indexed (60.0%).

GSC, Pages report, diamondbraces.com, read 8 Sep 2026. Certain

Why 1,350 URLs are not indexed

Redirects are the largest classification at 715. A further 152 are alternate pages with a canonical and 146 are duplicates for which Google chose a canonical; different URL patterns can still hold valid distinct canonicals, so these need inspecting before any is called a defect. On 8 Sep the SEO team described narrowing its indexing focus to 41 priority pages; redirects for the rest were raised as an option to look into.

Search Console, Page indexing report, reasons, read 8 Sep 2026. Validation status varies by row: "Discovered, currently not indexed" (18) shows Passed. Per-row statuses and timestamps have not been exported. The ten reason rows sum to 1,354 against the 1,350 headline; Search Console refreshes the two at different times, and the difference is preserved rather than reconciled. Certain on the displayed rows.

389K

Clicks, 16 months

40.2M impressions, 1.0% CTR, avg position 15.8

76.8K

Clicks, last 3 months

8.75M impressions, 0.9% CTR, avg position 11.2

88,104

Clicks on the exact query diamond braces

406,777 impressions, about 22.6% of reported clicks. This is one query, not the full branded share; a complete branded and non-branded classification has not been produced.

Top queries, diamondbraces.com, 16 months

Clicks, with queries shown exactly as Search Console records them. Portal, login and appointment queries carry 71% to 86% CTR; all three contain the brand name, so all three are branded under the definition used here, and their patient-status mix is unestablished.

GSC Performance, 16 months (the maximum the tool holds). The post-migration decline discussed with the team most likely predates this window and cannot be shown from Search Console. Certain on the rows; Likely on the window.

KidCo Dental: toothpaste queries lead the displayed sample

9,190 clicks, 2.01M impressions, 0.5% CTR, position 12.8 over 16 months; traffic begins around March 2026. The three toothpaste queries shown total 311 clicks, 3.4% of the total; the rest of the distribution and its geography have not been retrieved.

GSC Performance, kidcodent.dental, 16 months. Indexing: 141 indexed, 111 not (alternate canonical 49, crawled not indexed 21, redirect 19, 404 12, robots 9, other 4xx 1). Certain

info

What the blog is producing. Blog content on both sites ranks for informational queries (wire poking, kids' toothpaste, baby teeth), which points to informational rather than local intent. Blog-specific geography and key-event performance have not been established: the region data on this portal is property-wide, not page-level, so nothing here measures what blog sessions do. Reporting informational traffic separately from local acquisition is the way to find out; today the two are one number.

checklist1A.3

Access manifest (1A.3)

The systems the engagement needs, what has been granted, and what remains open. Tick items as they are resolved. When shared storage is connected, every viewer sees the same ticks and each change is logged with who made it.

Progress

2.1 Required before Phase 1A

2.2 Required for Phase 1B and oversight

2.3 Tools, documents and data

2.4 Written answers

Accounts observed with access to the properties

Recorded for the access review. Whether each should retain access is a client decision.

System	Observation
Google Business Profile	Primary owner is a shared Gmail account (bracesnow@…), which also holds GTM edit history.
Google Tag Manager	Four editing accounts in history: two personal Gmail, one external vendor domain (growthwizard.com), one shared Gmail. No company-domain account observed.
Search Console	Two unused verification tokens present on the diamondbraces.com property.
GA4	Consultant tier appears to be Editor. Administrator list not visible at this tier.

Read 8 Sep 2026. Identities are recorded as account types, not individuals. Certain

helpStill open

Open questions and remaining gaps

Everything Phase 1A could not close without the client. Each item names the role that can answer it. None of these block the 1A read-out; several block 1B scoping. Items that cut across phases are also on the Observations page, where you can respond to each.

#	Question	Why it matters	Who	Urgency

menu_bookEvidence

Sources and method

Every number on this portal names the source it came from, with its period and read date. Source names alone do not reproduce a calculation, so where a figure is derived the exhibit gives its numerator, denominator and scope, and where a supporting figure was not retained it says so. Listing records and a small number of chart constants sit outside the central data block and are noted where they appear. Nothing was exported by the client; everything was read by the consultant under the access granted.

Source register

Source	Scope	Period	Read
GA4, Diamond Braces - Shopify	Acct 304358041, prop 429695085. Admin, key events, data streams, filters, retention, product links, tag coverage read 8 Sep. Traffic acquisition (channel group, all 154 source / medium rows, all 108 campaign rows, key events by row), Region, Country and Pages re-read in full 9 Sep; every acquisition figure on this portal is from the 9 Sep read.	11 Aug to 7 Sep 2026 (28 d); 7-day views where stated	8 and 9 Sep 2026
GA4, KidCo Dental	Acct 236931102, prop 326433410.	All time	8 Sep 2026
GA4, SmilifyMe - GA4	Acct 81242300, prop 399685631.	7 days	8 Sep 2026
Google Tag Manager	GTM-KSXTJSZ, workspace, versions, admin.	Current state	8 Sep 2026
Search Console	diamondbraces.com and kidcodent.dental: Performance, Pages, settings.	16 months, 3 months	8 Sep 2026
GBP Manager	11 location groups, 75 listings, screenshots (no export available).	Current state	7 Sep 2026
Semrush	Diamond Braces project read directly on 9 Sep: Backlink Audit, Position Tracking, AI Visibility, Site Audit, On-Page SEO Checker, GA4 and Search Console connections. Listing Management and the KidCo project were not reached in the tool; those figures come from screenshots supplied by the SEO team on 8 Sep, whose capture date and settings are unconfirmed. Dated exports and historical runs are still outstanding.	Current state	8 and 9 Sep 2026
Public websites	diamondbraces.com, kidcodent.dental, smilify.com, engage.diamondbraces.com, edgebooking.ortho2.com. Page walks and network traces.	Current state	8 Sep 2026
Client statements	IT (Alex): Ortho2, NICE CXone, API, GBP link targets. SEO team (Eddie, Nicolas), 8 Sep working session: GBP link mix, indexing focus, review targets. Engagement briefing: Smilify model.	31 Aug to 8 Sep 2026	as dated

Method and limits

Diagnostic only. Every system was accessed in read mode under the consultant's own login. No setting, tag, listing, page or filter was changed. The one non-observational act was pressing Book and entering a zip code on public websites, which any visitor can do.

Grading. Certain means read from the system screen, and it covers the observation only, never a cause or a consequence drawn from it. Likely means inferred from what was read, with the inference stated. Guessing means an untested hypothesis offered for verification. Client-supplied means stated by the client and not independently observed. The Business Profile link targets and their reported booking performance, the NICE CXone behaviour, the Ortho2 pending-list behaviour and the Tag Manager container inventory are client-supplied.

Known limits. GA4 computes report totals separately from the rows beneath them, so rows can sum to slightly more or less than the total (82,261 channel sessions against a total of 82,086); figures are quoted as GA4 shows them. Session counts read on 8 Sep moved by about 1% when re-read on 9 Sep as GA4 finished processing; the 9 Sep figures are used throughout. GA4 region and country figures are active users, not sessions. Region includes a Not set row of 5,288 users, so the outside-the-service-area remainder contains unknown geography as well as known out-of-area users; the two are not separated here because the underlying breakdown was not retrieved. GA4 reports its user key-event rate on total users, which is a different denominator from the active-user counts shown alongside it. GBP Manager would not render the Diamond Braces location groups for the consultant, so listing-level detail comes from screenshots the client provided plus one live listing observed directly. The Smilify zip flow was tested by automation and should be re-tested by a person. Search Console cannot show anything older than 16 months.

Properties, not people. Account names appear only where they are the identity of a system role (a Business Profile owner, a Tag Manager editor). Findings describe the condition of the properties. A source screen cannot establish global absence, historical behaviour, current permissions, anyone's competence or intent, or a patient outcome; where a claim would need one of those, the text states the accessible scope and the check still to be done instead.

historyVersion history

What changed, and when

The build number, date and time on the cover tell you at a glance whether you are looking at something new since you last opened this. Every uploaded file carries its own number; nothing is reused. Newest first.

v2.189 Sep 2026, 18:38 ET

A three minute version now opens the Phase 1A page, for anyone who will not read the whole thing. It answers the question the engagement exists to answer, and shows how far a marketing touch can actually be followed before the trail stops: 82,086 sessions, 15,691 recorded booking actions, 7,879 of those from the four states with an office, and none of them traceable to a booked patient. Three dials carry the shares that matter, and a short list names the six things that need a decision from Dr. Drut, each linked to the observation behind it. On the Observations page, every proposed action is now a numbered checklist with the role that carries out each step, and each block is tinted by who has to move first: pink for a sponsor decision, blue for work the team can start now, green for what waits on Phase 1B. Presentation fix: status notes in the systems diagram were drawn as pill shapes around text that wraps onto several lines, which looked broken; they are now rounded panels, and the phase badge on narrow screens was corrected the same way.

v2.179 Sep 2026, 17:55 ET

Sign-in is now immediate. Every sign-in used to wait for a round trip to the shared storage service before checking the email and code, even though the invited reviewers are recognized by the page itself, so a slow first response left the button sitting there with nothing on screen. Invited reviewers now open the portal with no network call at all. The button shows a spinner and disables itself while it works, so a second press cannot stack another request. The one case that still needs the shared list, an email the page does not recognize, says "Checking access" while it waits and now gives up after nine seconds instead of waiting indefinitely.

v2.169 Sep 2026, 16:08 ET

The Business Profile booking question is now specific instead of open. Google documents two different metrics: Bookings, which counts completed bookings and is available only where bookings run through an integrated booking provider, and Booking clicks, which applies to hotels only. So a listing here either records completed bookings or records nothing about bookings and counts the interaction as a website click. Observation 06, the review log and the register row now carry that distinction, and the proposed action is the single screen that settles it: whether a directly-linked listing shows a Bookings row in its Performance report. This matters before the listings are switched, because if the metric is present the SEO team's report that booking-linked listings book more becomes directly verifiable.

v2.159 Sep 2026, 15:25 ET

Full correction pass after a second, independent adversarial review of v2.13, which raised 82 items. Every channel verdict is now scoped to the rows and metric it was measured on, so no channel is called best or highest. Out-of-area key events are reported as two separate populations, 7,467 for the United States and 7,812 for the property, which are not the same set. The claim that demand was overstated fourfold is withdrawn and replaced with the arithmetic. The register carries a stable identifier, an evidence and method line and one of four grades on every row, and it is described as 43 items rather than 43 proven defects. Business Profile Insights is no longer described as reporting booking-button clicks per listing; the metric available for these listings and this provider has to be checked first. The KidCo postal codes are stated as two disagreements rather than a swap. Recommendations on the Observations page are labelled as proposed client actions, subject to feasibility and separately authorized implementation, and every effort, cost and no-risk guarantee has been removed. Build times: the clock stamps on v2.7 through v2.13 could not be verified against a record, so they have been removed and the date kept; v2.14 is corrected to the time the file was actually written. A full disposition of all 82 items and the list of checks still needing live evidence accompany this build.

v2.149 Sep 2026, 14:02 ET

Wording tightened after adversarial review. Out-of-area key events are now described as unexplained booking-button clicks that Ortho2 could resolve, not as clicks that "cannot be bookings." Conclusion 02 now states the arithmetic (service-area audience about a quarter of the headline) rather than a fourfold overstatement of demand. Every reference to listings that link straight to Ortho2 booking now uses one formulation: GA4 records nothing for them by construction, and which listings do so is the SEO team's report; the register row is graded accordingly.

v2.139 Sep 2026

Semrush read directly for the Diamond Braces project and every screenshot figure replaced or dated: Position Tracking visibility 27.82% (was quoted at 30.2% from an older screenshot), top-3 55, top-10 81, top-100 150; AI Visibility 316 mentions and 451 cited pages with the per-assistant split; Backlink Audit potentially toxic corrected to 6.9%, audit dated 27 Aug 2026, Authority Score 38 added; Site Audit found to be a year old (20 Sep 2025) and On-Page Checker twenty months old (21 Jan 2025), both now dated; the GA4 property Semrush connects to identified by name as Diamond Braces - Main Site, reporting only the UltiPro recruiting portal, upgraded from Likely to Certain.

v2.129 Sep 2026

Book Now route re-walked on the live site: the office locator, an office page and its booking modal. The modal offers New patient, Existing patient and Reschedule, each with an office-specific Ortho2 code, so Ortho2 receives office and patient type from the link; GA4 still records one event for all three. Route text updated to match.

v2.119 Sep 2026

Review log wording aligned with the follow-up note to IT: booking-linked listing performance is verified in Business Profile Insights, with bookings in Ortho2.

v2.109 Sep 2026

Review log added to the Engagement roadmap with the 9 September review: what was shown, the question raised about how booking-linked listings are measured and the precise answer, what IT confirmed, and the next steps. The Book Now route now shows all four clicks to Ortho2. Register row for Tag Manager containers and Observation 08 updated with what IT confirmed.

v2.99 Sep 2026

Observations page made prescriptive: every one of the twelve items now ends with what to do about it, the role that does it and when, including the handling of the toxic link profile (export and review first, copyright removals for the scraped copies, a narrow disavow for the manipulative domains, a monthly check). The cover wording changed to match: the actions are recommendations; the consultant implements none of them.

v2.89 Sep 2026

Chart labels now use proper names (Spring Valley NY, GBP / organic, Direct) with the literal GA4 tag shown on hover; the readable-listings table gained a Listing column beside the slug. Observations cover text corrected to white on the gradient.

v2.79 Sep 2026

Smilify note removed from Three properties and the site-walk wording neutralized. The two channel charts moved to the top of the Phase 1A page, directly under the cover. Card bottoms aligned across every grid. Browser tab title corrected. From this release every file that is uploaded carries a new sequential number and a time stamp, so the cover always shows which build is live.

v2.69 Sep 2026

Every acquisition figure re-read in full from GA4 on 9 September and reconciled: all 154 source / medium rows, all 108 campaign rows, key events by row, region and country. Derived booking-action figures replaced with the key-event counts GA4 shows directly. Correction: 68 of 69 listings carry a readable campaign slug, not three; the earlier statement came from a ten-row read and is withdrawn. New finding: half of all key events come from users outside the four states, at New York conversion rates. Business Profile traffic shown across all 29 tag spellings. New pages: Phase 1B and Phase 1C, marked not started, holding what 1A surfaced for each; the misconfiguration register (now 43 rows) and the 1B conditions moved to the Phase 1B page; Semrush baseline material on the Phase 1C page. New Observations and concerns page with a feedback control on every item that reaches Jeremy in Slack. Written for the sponsor throughout.

v2.59 Sep 2026

Phases separated across the top, matching the SOW: Phase 1A and Phase 4 are open pages; Phase 1B and 1C show as not authorized and cannot be opened. The SEO discoveries, gaps and session log now live on the Phase 4 Oversight page with the oversight scope from the SOW. Aligned to the SOW: the three business day review window on the roadmap, a double-counting assessment on the conversion path map, the 1B section reworded as conditions rather than fixes, open questions reworded as questions, fix owners shown as roles rather than names, and the access manifest annotated with the permission tiers the SOW requires plus the platforms 1C needs.

v2.49 Sep 2026

Sign-in now holds for the browser session only, and the gate no longer flashes before opening. Top channels by visits and booking-action rate by source added at the front. Channel finding reworded as measured so far rather than a verdict. New SEO group: discoveries from the working sessions, gaps in the SEO function with six posts read closely, and a weekly session log. Every Business Profile listing now links to its Google listing. Typeform found in use on blog templates; the register and open questions corrected. GBP and Google capitalized throughout.

v2.39 Sep 2026

New section, Everything misconfigured in one register: every defect found between Business Profile, the website, Tag Manager, GA4, Search Console and reporting, with severity, owner and a chain diagram showing where each hop is broken. Filter by system or severity.

v2.29 Sep 2026

New section, Where booking actions start: booking actions by source and by office for GBP traffic, with the honest gap between booking actions and booked patients stated up front. Derived figures are marked as derived and will be replaced with exact GA4 key-event counts, plus a source-by-state view, in the next read.

v2.18 Sep 2026

Reframed as the first Phase 1A review and the living home for every phase, with a phase roadmap on the cover. More space throughout, a clearer top-line, and the version history moved to the end. Shared chat and checklist wired to the live Google Apps Script backend, notifying the same Slack channel as iCare.

v2.08 Sep 2026

Full redesign in Diamond Braces branding. Access now requires an email and an access code, and the owner can add or remove people from Manage access. Ask Jeremy chat added with Slack notification. Service map regenerated from public map boundaries with no third-party map service. Findings updated with the 8 September SEO working session: GBP listings are mixed between website and booking links, and the team said it intends to move all of them to booking links.

v1.08 Sep 2026

First portal build. Six conclusions, three-property comparison, blind spots, GBP evidence, traffic quality, conversion paths, service map, GBP listing charts, instrumentation, Search Console, access manifest, open questions, 1B scope, and the source register.

Prepared by Jeremy Mays, Transmyt, for Dr. Oleg Drut, Diamond Braces. Independent Technical Assessment and SEO Oversight engagement. Portal . The sign-in on this page is a courtesy lock for a working draft, not access control: this file carries its content in the page itself, so anyone with the address can read it. Authenticated hosting is a deployment prerequisite before this portal holds anything confidential. Please do not forward the address or the code beyond the named recipients.

lock_openNot started · not authorized · not invoiced

Phase 1B · Measurement audit

What Phase 1A has already surfaced for 1B

Phase 1B has not started. This page exists so that what the 1A discovery turned up about measurement is not lost between phases. Everything here surfaced while reading the properties for Phase 1A; it is recorded, not resolved, and it will be addressed in depth once Phase 1B is authorized in writing.

Authorizable after written approval of 1ASix deliverables, 1B.1 to 1B.6Runs in parallel with 1CNothing here is implemented

info

Not officially started. Phase 1B is neither authorized nor invoiced. The material on this page surfaced during Phase 1A discovery and is held here so it is not lost. It is a preview of the ground a measurement audit would cover, not the audit itself; the audit goes deeper on every item once the phase is authorized in writing.

list_altWhat 1B covers

The six deliverables, and what 1A has already touched

The SOW defines Phase 1B as a measurement audit with six deliverables. For each, this is what the 1A read has already put on the table. None of it counts as 1B work delivered; under the SOW nothing is repeated across phases, so 1B starts from here rather than from zero.

1B.1Instrumentation audit

Whether GA4, Search Console, Tag Manager, cross-domain and conversion tracking are configured correctly. 1A read every admin screen and produced the register of 43 items below: 2 rated critical, 17 high, and 32 that appear to need no website code change. 1B would verify each with the administrators, add what 1A could not see at Editor tier, and establish causes for the two population problems (28.3% of sessions classified Unassigned; about half of key events falling outside the four-state rows).

1B.2Handoff assessment

Where a visitor is handed from one system to another and what survives. 1A mapped the routes (website to engage.diamondbraces.com, to a Typeform, to Ortho2, to the call center) and showed that the same patient can be counted twice at every join. 1B would test each handoff with real submissions and trace what arrives at the far end.

1B.3Phone attribution

Whether calls can be tied to a listing, page or campaign. 1A recorded that all calls route through NICE CXone, that the dialed number is logged but not attributed, that no call-tracking numbers exist, and that five public numbers span the three sites. 1B would establish what NICE can export and whether per-listing numbers are practical with one shared call center.

1B.4Starts attribution model

A defensible way to connect a treatment start back to a source. 1A recorded that Ortho2 online bookings sit on a pending list unlinked to source and disappear at manual conversion, that a read-only API exists, and that every catalog code fires the same GA4 event, including three per office page and two general codes on the homepage modal. 1B would inventory every code in use, test whether the pending list retains one, and assess whether any session-level join is possible. Office and patient-type codes identify neither a marketing source nor a unique session, so a join needs a persistent identifier that has not been shown to exist.

1B.5KPI definitions

Which numbers become the source of truth. 1A shows why the current ones cannot serve yet: 28.3% of sessions are classified Unassigned, 26.0% of active users were assigned to the four service-area states, and about half of key events fall outside those rows without an established cause. Provisional candidates are in-area sourced sessions, Business Profile key events, and Ortho2 consultations and starts by location; each carries open questions of its own, including how much valid activity an exclusion would discard and whether the replacement measures still rest on the same untested event. 1B validates and defines them before any is adopted as a sponsor KPI.

1B.6Priced remediation plan

What to fix, in what order, at what cost, approved separately. 1A supplies the raw material: the register below, sorted by severity, with 32 of 43 items that appear to need no website code change. Effort, dependencies, testing and price are assessed in 1B, not here. Nothing has been fixed and nothing is priced on this page.

buildToward 1B.1 · instrumentation audit

Everything found, in one register

Every item Phase 1A found between Google Business Profile, the website, Tag Manager, GA4, Search Console and the reporting layer, in one place. The 43 rows mix confirmed observations, suspected defects and questions that need validation, so a row is not by itself a proven defect. Each carries a stable identifier, what was seen, why it matters, how serious it looks, the role proposed to verify it, the evidence and method behind it, and one of four grades. This is the working list the 1B instrumentation audit would be scoped from. Severity: Critical means no usable data exists; High means data that exists changes a decision; Medium and Low are trust and hygiene.

Observations, suspected defects and questions to validate, across six systems and the links between them

critical

A property or brand with no usable data

high

Wrong or missing data that changes a decision

appear to need no website code change

Settings, tags, links and listing edits. Effort, dependencies and testing are still to be assessed.

How the systems connect today

Follow a visit from a Business Profile listing to a report. Every hop has a defect.

Business Profile75 listings29 tag spellings, 2 campaign conventions

→UTM tags on links

WebsiteShopify site13 URLs classified untagged; no analytics request seen on the Get Started form

→Google tag via GTM

Tag Manager1 container visible in this account11 changes unpublished; no release observed after Dec 2024

→2 destinations, 1 unknown

GA43 propertiesKidCo no data, Smilify 7 users; retention 2 months; no product links

→Connected property

ReportsSemrush, weekly reportsConnected GA4 property shows recruiting-portal pages

Search Console and Business Profile are not linked to GA4, so query data and profile actions do not arrive. Google documents aggregated reporting limits on the Business Profile link even once it is in place.

ID	System	What was found	Why it matters, and the evidence behind it	Severity, grade and verification owner

Each row carries its own evidence and method line: the system, report, period and read date, what was observed, what is inferred, and what has not been checked. Grades: Certain read from the screen named, covering the observation only; Likely a stated inference from what was read; Guessing an untested hypothesis; Client-supplied reported by the client and not independently observed. Owners are the roles proposed to carry out verification, not judgments about people. Items marked "no code" appear to need no website code change; effort and testing are assessed in 1B. Nothing here has been changed by the consultant; under the SOW remediation is priced in the 1B plan and approved separately.

arrow_forwardConditions and questions

What 1A has shown a measurement audit would have to resolve

This is not a proposal. It is the list of conditions 1A has shown cannot be resolved with the properties as they stand, and the questions only a measurement audit can answer. Each maps to a finding on the Phase 1A page. Phase 1B becomes authorizable only after written approval of Phase 1A.

Conditions 1A has surfaced for 1B to take up

GBP tagging29 source and medium spellings and two campaign conventions, so the channel does not appear as one line in any standard report and 43.9% of its sessions carry no listing identity. The slug-to-listing crosswalk is unconfirmed. A listing that links straight to Ortho2 booking loads no page on the site, so website analytics record nothing for it.
RetentionGA4 event retention at two months, which limits historical exploration and funnel work to that window. Whether to change it now is a client decision listed under Open questions.
FiltersAn internal-traffic filter in Testing that does not exclude matching traffic, 28.3% of sessions classified Unassigned, and about half of all key events falling outside the four-state rows.
Tag ManagerEleven unpublished changes, an unpublished version 26, and a second Google tag destination with no identified owning property.
LinksSearch Console and Business Profile not connected to GA4.
Key eventsOne event for every Ortho2 route, including new, existing and reschedule; four configured events that recorded no data in the period read.
FormsTwo Get Started destinations, engage.diamondbraces.com and a Typeform, neither measured through to a booking.

These are the conditions, not the fixes. Under the SOW the fixes are priced in the 1B remediation plan and approved separately; nothing here is implemented inside an assessment.

Questions only 1B can answer

Key eventsWhy several out-of-area states show user key-event rates overlapping the service-area range, which spans 24.06% to 27.96% for New Jersey, Pennsylvania and Connecticut. Automation, a trigger firing on something other than a hand-off to Ortho2, and a real unaccounted population are all untested, and more than one may apply. The full seventeen-state breakdown still has to be documented.
AttributionWhether any session-level join to an Ortho2 booking is feasible at all, what identifiers are actually retained, and how much of the pending list is lost at manual conversion. 1B assesses feasibility and documents the limits; any remediation is specified and authorized separately.
CallsWhether NICE CXone dialed-number logs can be exported, and whether per-listing or per-page numbers are practical given the shared call center.
KidCoWhy the GA4 property shows no recorded data: a different measurement ID on the theme, an absent tag, collection reporting elsewhere, or an abandoned property.
SmilifyWhat the business actually sells today, where G-2BGW5RHH82 fires, and whether the zip flow works.
IndexationWhich of the 715 redirect URLs, 152 alternate canonical pages and 146 duplicates should be indexed, since intentional redirects and canonicals are not defects, and what the pre-Shopify traffic baseline was (needs the old GA4 or Universal Analytics property).
DemandWhether a local-demand measure can be built at all: in-area, sourced, non-patient sessions by office against Ortho2 consultations and starts. How accurately patients can be identified, and how much valid activity any exclusion discards, has to be tested before such a measure is used.

lock_openNot started · not authorized · not invoiced

Phase 1C · Baseline assessment

What Phase 1A has already surfaced for 1C

Phase 1C has not started. It records the objective condition of all three properties, with date, source and method, so that the same measurement can be taken again later and compared. While reading the properties for Phase 1A, a good deal of baseline material came into view. It is held here, dated, so it is not lost; it becomes a baseline only when 1C is authorized in writing and the measurements are taken deliberately and completely.

Authorizable after written approval of 1AFive deliverables, 1C.1 to 1C.5Runs in parallel with 1BNothing here is a recommendation

info

Not officially started. Phase 1C is neither authorized nor invoiced. The figures on this page were read incidentally during Phase 1A discovery. The Diamond Braces Semrush project was read directly in the tool on 9 September; the Listing Management and KidCo figures still come from screenshots the SEO team supplied, whose capture date and settings are unconfirmed. They are dated and graded so they can be checked, but they are not the baseline; the baseline is taken once, completely, with export access, when the phase is authorized.

list_altWhat 1C covers

The five deliverables, and what 1A has already touched

The SOW defines Phase 1C as a baseline assessment with five deliverables. For each, this is what came into view during 1A. Under the SOW nothing is repeated across phases, so 1C would start from here.

1C.1Technical baseline

Crawlability, indexation, site health, speed and structure, recorded so they can be re-measured. 1A read Search Console indexation for both properties that have one (Diamond Braces 899 of 2,249 known URLs indexed; KidCo 141 of 252), the three URL patterns location pages live under, the duplicated office locator, and a Semrush Site Audit health score of 82% on a crawl capped at 500 pages. 1C would crawl every page of all three sites with a proper tool and record the full technical state.

1C.2Authority baseline

The link profile and brand demand, recorded so movement can be seen. 1A saw a Semrush Backlink Audit rated High toxicity (278 toxic referring domains, 16.1%), a search demand profile in which the exact query diamond braces accounts for 88,104 clicks over 16 months, about 22.6% of reported clicks, and a Position Tracking visibility of 27.82%. Those were read directly in the tool on 9 September; what 1C still needs is dated exports and historical runs, plus the properties Semrush does not track.

1C.3Work-in-flight inventory

What the SEO team and its vendors are doing now, so later movement can be attributed to it. 1A heard, in the 8 September session, the current priorities: 41 priority pages for indexing, listing links moving to direct booking, a review backlog being worked down, blog output of several long posts a week. Six recent posts were read and four weekly reports received. 1C would inventory every open workstream with owner, start date and expected effect.

1C.4Coverage gaps

Where the properties simply have nothing. 1A found no Search Console property or Tag Manager container for Smilify in the accessible accounts and no analytics request on its homepage; toothpaste queries leading the displayed KidCo sample rather than local dental queries; AI-related measurements present in GA4 and Semrush but not confirmed to be in routine reporting; and no visible medical reviewer credit on the sampled health posts. 1C would list every gap against a defined standard.

1C.5Baseline document

One dated record everything later is compared against. Nothing on this page is that document. It is the material a baseline would be built from, and a note of which figures came from screenshots rather than exports.

straightenBaseline material already in hand

Figures read during 1A that a baseline would start from

Each figure below is dated and graded. Search Console and GA4 figures were read directly. Semrush figures for the Diamond Braces project were read directly from the tool on 9 September; the Listing Management figures still come from the team's screenshot of 8 September. Export access with history remains an open item on the access manifest.

Indexation

Diamond Braces2,249 known URLs, 899 indexed (40.0%). 715 classified as redirects, 152 as alternate pages with a canonical, 146 as duplicates where Google chose the canonical, 165 crawled but not indexed, 95 not found, 31 blocked by robots.txt. Validation status varies by row; the 18-URL discovered-not-indexed row shows Passed.
KidCo252 known URLs, 141 indexed (56.0%). A similar classification mix at smaller scale; its exclusion reasons sum to 111.
SmilifyNo Search Console property. Nothing can be said.

Search Console, Pages reports, read 8 Sep 2026. Certain

Search demand

Diamond Braces389K clicks and 40.2M impressions in 16 months; 76.8K clicks in the last three. The exact query diamond braces is 88,104 clicks, about 22.6% of the total. Several other prominent queries also contain the brand name and concern portal access, login and appointments; a complete branded and non-branded classification has not been produced.
KidCo9,190 clicks in 16 months; the three toothpaste queries displayed total 311 clicks, 3.4% of that. The rest of the query distribution and its geography have not been retrieved. Traffic begins around March 2026.
AI assistants342 sessions from the AI Assistant channel in 28 days (ChatGPT 327), producing 72 key events. GA4 and Semrush both carry AI-related measurements; whether routine reporting uses them has not been checked.

Search Console Performance, read 8 Sep 2026; GA4 Traffic acquisition, read 9 Sep 2026. Certain

Authority and link profile

ToxicitySemrush Backlink Audit rates the Diamond Braces profile High toxicity: 278 toxic referring domains (16.1%), 120 potentially toxic (6.9%), about 1,300 non-toxic (77%), from 1,732 referring domains and 10,209 backlinks analyzed; audit last run 27 Aug 2026. Authority Score 38; 89% of referring domains sit in the 0 to 20 authority band.
PatternThe domains Semrush scores most toxic are throwaway web.app subdomains carrying what appear to be copies of the blog's own articles, plus one aged-domain seller scored 100. Scraping and a deliberate campaign are both untested hypotheses: a tool flag establishes neither intent nor ranking harm, and ownership of the copies has not been verified. Current remediation status is unknown, and nothing should be submitted before the history and any existing disavow file are inspected.
VisibilityPosition Tracking visibility 27.82% on 9 Sep 2026 (down 2.16 points over 90 days); 55 tracked keywords in the top 3, 81 in the top 10, 102 in the top 20, 150 in the top 100. "braces cost" ranks first. The tracked set is dominated by "can I eat X with braces" questions.

Semrush, Diamond Braces project, read directly 9 Sep 2026; audit last run 27 Aug 2026. Certain on the figures as displayed; Guessing on cause, intent and any ranking effect, none of which has been tested.

Local presence

Listings75 Business Profile listings across 11 groups: 69 Diamond Braces (68 open), 6 KidCo (5 open). 74 verified. Nearly every listing shows a red indicator in GBP Manager whose meaning is unconfirmed.
SemrushThe Listing Management screenshot shows 22 of 37 tracked listings needing fixes and an online presence rated Bad. The KidCo screenshot shows a "KidZdent" entity in Old Bridge NJ, matching none of the six KidCo listings. Whether the project is configured against that entity, and therefore scoring a different business, still has to be confirmed in the project settings.
Data qualityThe website and Business Profile disagree on the KidCo Bay Ridge and Flatbush postal codes, and the two values are not a reciprocal swap; one closed Diamond Braces listing is still published; two Queens listings use a different name order.

GBP Manager screenshots 7 Sep 2026; Semrush Listing Management screenshot supplied by the SEO team 8 Sep 2026, capture date and settings unconfirmed. Certain on the listing counts and addresses; Likely on the reading of the Semrush project.

AI visibility

Diamond BracesSemrush AI Visibility score 35; 316 mentions and 451 cited pages across ChatGPT (94 mentions, 150 cited pages), AI Overviews (58, 289), AI Mode (94, 249) and Gemini (70, 125).
KidCo14 mentions, from the client-supplied screenshot received 8 Sep 2026, not read in the tool. Capture date and settings to confirm.
Why it is hereSessions are already arriving from this channel and the tools already carry measurements of it. Whether routine reporting includes them is unchecked. A baseline taken now can be compared later; one taken later cannot be backfilled.

Diamond Braces figures read directly in Semrush 9 Sep 2026 Certain. KidCo figure from the SEO team's screenshot of 8 Sep 2026 Client-supplied.

On-page and site health

Site AuditHealth score 82% on a crawl capped at 500 pages, with 14 errors and 1,863 warnings, last run 20 September 2025, so a year old and directional at best; Google knows 2,249 URLs.
On-Page Checker189 ideas across 30 pages, last run 21 January 2025; the pages it ranks highest to optimize are /retainers/ and /careers/dental-assistant-career/, and the second is a recruiting page.
CautionSemrush's organic traffic estimate for the domain (48.4K, down 15.7% month on month) is a model, not a measurement, and should not be quoted beside the GA4 figures. The Semrush Search Console widget, which is a measurement, shows 25.5K clicks in the last 30 days.

Semrush, read directly 9 Sep 2026. Certain

constructionWork in flight, as observed

What is being worked on now, so later movement can be attributed

A baseline is only useful if the work that follows it is recorded too. These are the workstreams 1A has heard about or observed. They are recorded as heard or as observed, and labelled which; nothing here is an assessment of the work.

IndexingFocus narrowed from roughly 1,000 pages to 41 priority pages; redirects for the unindexed long tail raised as an option to look into. Heard, 8 Sep 2026.
Listing linksEvery Business Profile listing to be moved to a direct Ortho2 booking link; the team said booking-linked listings book more. Once done, Business Profile sessions in GA4 will fall while bookings may rise; both have to be read together. Heard, 8 Sep 2026.
Reviews239 unreplied reviews being worked toward a target of 50 set by Dr. Drut; comment-free one-star reviews being disputed. Heard, 8 Sep 2026.
ContentSeveral posts a week of 2,000 to 4,500 words, topics chosen by keyword research. Six posts sent 8 Sep and read on the live site 9 Sep; see the gaps evidence on the Oversight page. Heard and observed.

ReportingWeekly SEO and Business Profile reports, on Fridays by email with the consultant copied; four received (7 to 28 Aug), not yet reviewed line by line. Received, 8 Sep 2026.
Smilify tagInstalling analytics on smilify.com sits with IT. In the session the SEO team described the property as having recorded nothing for over a month; the property itself showed seven users, 24 events and no key events in the seven days read on 8 September. Both are recorded; the two statements have not been reconciled. Heard, 8 Sep 2026; read in GA4, 8 Sep 2026.
Location pagesCreation of the 69 location pages paused pending an infrastructure decision, with a move off Shopify under discussion. Client decision, 1 Sep 2026.
MeasurementNothing in Tag Manager has been published since December 2024; eleven changes wait in the workspace. No measurement work is known to be in flight. Observed, 8 Sep 2026.

info

What this page is not. It is not the baseline and it is not an assessment of the SEO team. It is the material that came into view during Phase 1A, held with its dates so that when Phase 1C is authorized the baseline can be taken quickly and nothing that was seen has to be found again.

Phase 4 · Independent SEO oversight · running

Oversight

A periodic, independent look at whether the highest-impact work is being done and whether it is producing measurable results. This page holds what the weekly working sessions with the SEO team surface, the gaps the evidence shows, and a running log for the sponsor. It records; it does not direct anyone's work and it does not evaluate any individual.

15 hours reserved per monthOne written note per review, by default monthlyWeekly sessions since 8 Sep 2026, within the reserved hours90-day initial term from 1A kickoff

visibilityWhat oversight covers

Scope, and what it is measured against

Oversight reviews work in progress against the KPIs defined in Phase 1B and the baseline recorded in Phase 1C. Neither exists yet, so for now this page records observations and holds them until there is something objective to measure against. That is by design: it means the first months of oversight are informed by watching the work, not by reading reports about it.

In scope

Review of work in progress against the KPIs and baseline, once they exist
Independent validation of reported performance: what moved, versus what was measured differently
Identification of coverage gaps on the properties, with the reasoning
Review of high-risk technical decisions before implementation: URL structure, templates, schema, tracking at scale
An independent technical opinion when internal and external parties disagree
Connecting SEO activity back to calls, consultations, appointments and starts

Not in scope

Managing or directing the internal team
Evaluating individual performance
Implementing changes to any system
Replacing any existing function

Cadence and reporting

By default, one working session and one written note per month; cadence is confirmed after Phase 1B
Since 8 September, weekly working sessions with the SEO team, held within the 15 reserved hours
Each note records what was examined, what was found, what would be changed and what would be left alone; the session log below is that note

searchSEO · discoveries, not conclusions

What the SEO working sessions have surfaced

What the weekly sessions with the SEO team and the systems themselves have surfaced about how search is being managed today. Each item is recorded as heard from the team or as read from a system, and labelled which. None of it is a recommendation; those come once the picture is complete.

link

Listing links are mixed, and the team already knows which kind books

The team described its Business Profile listings as split between links to location pages, chosen deliberately for search value, and direct Ortho2 booking links where a listing had the scheduling feature enabled. It said the booking-linked listings book more, and that it intends to move all listings to booking links.

Heard, 8 Sep 2026. Consistent with the 29 tag spellings read in GA4, and it means the booking-linked listings are the ones GA4 cannot see.

travel_explore

Indexing has been narrowed to 41 pages; validation status is not tracked per reason

The team said it reduced the pages it asks Google to index from roughly 1,000 to 41 priority pages. In Search Console 1,350 of 2,249 known URLs are reported not indexed, and the displayed validation statuses vary by reason row; the 18-URL discovered-not-indexed row shows Passed. Whether fixes or individual indexing requests were made is not established by that status.

Heard, 8 Sep 2026; read in Search Console, 8 Sep 2026.

reviews

239 unreplied reviews against a target of 50

Dr. Drut set the target. The team is working the backlog down and disputing comment-free one-star reviews for removal. One person, based in Colombia, handles reviews and social for every location and is also the contact for video.

Heard, 8 Sep 2026. Semrush shows 22 of 37 tracked listings needing fixes.

article

Content output is high, and its search and booking performance has not been measured

Topics are chosen by keyword research and posts run 2,000 to 4,500 words. The six-post review below documents what each post contains and how it renders. Their target keywords, current rankings and key-event performance were not supplied and have not been analyzed, so nothing here establishes what the blog produces. The team described Shopify as the constraint on adding call-to-action boxes and formatting images.

Heard, 8 Sep 2026; six posts read on 9 Sep 2026, see Gaps below. A dated ranking and key-event analysis is still to be done.

history

The team describes a history of migrations without a QA step

Asked about the site's condition, the team described repeated platform migrations built by outsourced developers, without an internal or external quality-assurance step, as the source of recurring failures that cost time and set the SEO work back. The live blog posts, which carry unrendered template tokens, are consistent with that account.

Heard, 8 Sep 2026; observed on live pages, 9 Sep 2026.

visibility_off

Two brands have measurement gaps the team was also unaware of

The team described the Smilify GA4 property as having recorded nothing for over a month, which was new information in the session; installing the tag sits with IT. The property read on 8 September showed seven users, 24 events and no key events in seven days, and the two statements have not been reconciled. The KidCo property showed no recorded data. The team acknowledged that KidCo draws national rather than local traffic and that Diamond Braces draws a large share of international traffic.

Heard, 8 Sep 2026; read in GA4, 8 Sep 2026.

mail

Weekly reports existed but were not reaching the review

The team has produced a weekly SEO and Business Profile report since at least 7 August. Four of them arrived by email on 8 September, along with six recent posts, once delivery was moved from Teams to email with the consultant copied.

Heard, 8 Sep 2026; four reports received by email.

groups

The function is two people, plus one for reviews

Two specialists cover 75 listings, three websites, content production and weekly reporting, with a third person on reviews and social. The team described the workload as beyond what the current headcount can manage well.

Heard, 8 Sep 2026.

reportSEO · observations

Gaps in the SEO function, as evidenced

Where the evidence shows coverage falling short on the properties, grouped by the kind of gap. It is drawn from the systems and from the team's own account, and it is here so the sponsor can see the shape of the problem before any recommendation is made. Under the terms of this engagement it records the condition of the properties and the function around them; it does not evaluate any individual.

Foundation

The platform and structure the team is asked to work on.

PlatformShopify, described by the team as the constraint on URL structure, calls to action and image handling.
Index60% of known URLs not indexed after the migration; location pages under three URL patterns; the office locator on two URLs.
TemplatesPlaceholder tokens and empty image sources visible on published posts, observed on 9 September; when they first appeared is unknown.
MeasurementTwo brands with no usable data in the properties read; one Tag Manager container visible for three brands, with no publication observed after December 2024.

Capacity

What the headcount is being asked to cover.

Scope75 listings, three websites, content production and weekly reporting across two specialists and one reviews and social lead.
Backlog239 unreplied reviews; listing links not yet inventoried one by one; indexing validation statuses varying and not tracked per reason.
CadenceBlog output of several posts a week while the items above wait.

Execution

Where what shipped does not match the stated process.

TaggingOne channel tagged three ways and two campaign conventions, from the team's own listing links.
Alt textThe team stated that every image gets alt text; three of six posts read have images with empty alt text or an empty image source. Empty alt text is correct for a decorative image, so which of these are informative still has to be identified.
AuthorityNo visible medical reviewer or credentialed byline on the sampled health-condition posts. That is not proof no review took place, and any search effect is unmeasured.
FreshnessFive of six posts carry the same "Updated on May 5, 2025" stamp, which points to a template value; whether it reflects real updates is unverified.
StructureOne post carries 22 near-identical question headings; a recruiting post sits in the patient blog with patient posts recommended beneath it.

Process and quality assurance

What lets defects reach the public and stay there.

ReleaseNo QA step on outsourced development, by the team's account; in Tag Manager, 11 changes unpublished and a version 26 with no publication observed. Whether that reflects a decision or an oversight is not established.
ReviewTemplate tokens visible on live posts carrying a May 2025 date stamp, observed 9 September 2026; how long they have been public is unknown.
ReportingWeekly reports produced but not reaching the people reviewing the work until 8 September.
OwnershipMeasurement changes drafted across two years by four editing accounts, with no publication observed and no owner recorded.

Six recent posts, read closely

The six posts the team sent on 8 September, read from the live pages on 9 September. What each one does and does not carry.

Post	Words	Calls to action in the body	Images	Placeholders or broken elements	Byline and review	Date stamp
Crossbite: what it is, types, causes	~4,200	5, to a Typeform quote form	3; diagrams carry filename alt text only	None	Company byline; no reviewer	Updated May 5, 2025
Hawley retainer: cost, care, lifespan	~4,000	5, quote and nearest office	5; all with descriptive alt text	None	Individual byline; no reviewer	Updated May 5, 2025
Invisalign chewie hacks	~2,200	0 in the body	4; all with empty alt text	"PLACEHOLDER" text left in two places; an unrelated iTero scanning passage pasted in; 22 near-identical question headings	Company byline; no reviewer	Updated May 5, 2025
Orthodontic spacers, complete guide	~3,400	2 unrendered tokens where buttons should be	3; two with an empty image source	"[CTA_NEAREST_OFFICE]" and "[CTA_QUOTE]" showing as text; logo image source empty	Individual byline; no reviewer	Updated May 5, 2025
Dental assistant salary and career	~4,300	6, to the recruiting portal	3; all with descriptive alt text	None; a recruiting post inside the patient blog, with patient posts recommended beneath it	Company byline	Updated May 5, 2025
Can you drink with Invisalign	~3,600	4, consultation and quote	3; alt text missing on all three	One section heading appears twice; footer navigation rendered twice	Company byline; no reviewer	Updated Sep 4, 2026

Read from the live pages on 9 Sep 2026; links from the SEO team's email of 8 Sep 2026. Word counts are approximate. Certain on what rendered.

info

What this section is and is not. It records the condition of the SEO function as the evidence shows it, so that the sponsor can see the gaps between capability, capacity, platform and process before any recommendation is put forward. It draws no conclusion about what to do. That comes later, and separately. The concerns that reach beyond SEO, and a place to respond to each, are on the Observations page.

groupsSEO · weekly log

Working sessions

A running record of the weekly working sessions with the SEO team: what was heard, what the sponsor should know from it, and what the team said it would do next. Working discussion that belongs inside the team stays out. What appears here is what helps the sponsor prepare for recommendations when they come.

8 September 2026

Jeremy Mays, Nicolas Cabrera, Eddie Santiago. Alex invited. Cadence moved from every two weeks to weekly.

What was heard

ReportsWeekly SEO and Business Profile reports had been going to Teams and not reaching the review. They now go by email on Fridays with the consultant copied; four arrived the same day.
Listing linksListings are split between location-page links, chosen for search value, and direct Ortho2 booking links where the scheduling feature was enabled. The booking-linked ones book more.
IndexingFocus narrowed from roughly 1,000 pages to 41 priority pages. Redirects for the rest were raised as an option to look into.
Reviews239 unreplied, target 50 set by Dr. Drut. Comment-free one-star reviews are being disputed. One person in Colombia handles reviews and social for all locations and is the contact for video.
ContentTopics come from keyword research; images are compressed with Squash and, per the team, given alt text. The team described Shopify as the constraint on call-to-action boxes and image formatting, and a shortage of real procedural photography.
HistoryThe team described repeated migrations built by outsourced developers without a QA step as the source of recurring failures.
MeasurementSmilify's GA4 property recording nothing for over a month was new to the team; the install sits with IT. KidCo's national traffic and Diamond Braces' international traffic were acknowledged.
CapacityTwo specialists plus one reviews and social lead for 75 listings, three sites, content and reporting; described by the team as more than the headcount can carry well.

What the sponsor should know from it

Links change the numbersMoving every listing to a direct booking link removes the website from that path. Business Profile sessions in GA4 will fall while bookings may rise. Both have to be read together or the change will look like a loss.
Smilify is an IT taskAn analytics tag has to be placed on the site before anything about Smilify can be measured. IT is the proposed owner; the deployment permissions and responsibilities for that site still need confirming.
The review target has a dependencyReaching 50 unreplied reviews depends on one person who also carries social for every location. What that takes in time is for the client to assess.
The QA account is consistent with what shippedThe team's description of development without a quality step is consistent with what the live blog posts show: template tokens visible to the public on 9 September, on posts carrying a May 2025 date stamp. How long they have been public is unknown.

What the team said it would do next

Eddie, NicolasUpdate every Business Profile listing to link directly to booking.
NicolasLook into redirects for the unindexed long tail.
BothSend results and new content by email as they land; keep the weekly session.

What comes back to the next session

From this reviewDone on 9 September: exact GA4 key events by channel, source and listing now replace the derived figures on the Phase 1A page. Still to come: a list of which listings link to booking rather than the website, from the team.
From the six postsThe placeholder and image defects above, for the team to trace to the template.

rate_reviewCross-phase · for the sponsor · with your feedback

Observations and concerns

What matters that does not belong to one phase

Some of what Phase 1A turned up is bigger than any single deliverable: account control, data retention, decisions already taken that will inherit today's measurement gaps, a channel about to become less visible in one system. Each item below states what was observed, why it matters to the business, and a proposed action with the role that would carry it out. Under each one you can tell Jeremy whether you agree, are not sure, or disagree, and add a comment. The proposed actions are the consultant's recommendations for the client to consider; each is subject to feasibility, effort and separately authorized implementation, and the consultant implements none of them.

12 observationsEach dated and gradedFeedback reaches Jeremy when delivery succeedsProposed actions, not decisions

rate_reviewObservations

Twelve things worth your attention

In rough order of consequence. Each one ends with a checklist of the steps it would take, the role that would carry out each step, and the order they go in. The colour of that block says who has to move first.

how_to_voteNeeds a decision from Dr. Drut firstboltThe team can act on this nowscheduleWaits on Phase 1B authorization

Observation 01

Half of the recorded conversions come from users outside the four-state rows, and their nature is unverified

GA4 recorded 15,691 key events in 28 days, 14,389 of them Click to Ortho2 events and 1,302 form submissions. 7,879 were attributed to users in the New York, New Jersey, Pennsylvania and Connecticut rows. 7,467, or 47.6% of the total, were attributed to United States users outside those four states. A further 345 fall outside the United States rows, so the property remainder outside the four-state rows is 7,812, or 49.8%; the two remainders are different populations and should not be treated as one. Several displayed out-of-area states show elevated reported user key-event rates, some overlapping the 24.1% to 29.3% span of the four service-area rows, while California, Texas and Illinois sit under 6%. The full seventeen-state breakdown remains to be documented. A key event is a recorded click on a booking control, not an appointment.

Why it matters. Any report, dashboard or vendor claim that quotes a conversion count, a conversion rate or a cost per conversion from this property is quoting a number of which roughly half is unexplained. Whether those clicks are automated, a tag firing on something other than a booking click, or a real population nobody has accounted for cannot be read from GA4. Ortho2 appointment records by office, once access is granted, could show whether they became appointments.

Root cause to be diagnosed in Phase 1BRead 9 Sep 2026

GA4 Region rows with Key events and User key-event rate, 11 Aug to 7 Sep 2026, read 9 Sep 2026. Certain on the figures as reported; Guessing on the cause; the three candidate causes are untested hypotheses.

boltThe team can act on this now

IT. Test the Click to Ortho2 trigger in Tag Manager against the live booking modal and the office-page buttons, and record exactly what fires it. A trigger fault would explain part of this population, not necessarily all of it.
IT. Build a GA4 comparison for New York, New Jersey, Pennsylvania and Connecticut, so in-area key events can be reported on their own while the question is open.
IT. Investigate the out-of-area rows using the GA4 dimensions that exist, such as device category, browser, landing page and hostname, and if needed server or CDN request logs. GA4 has no network-domain dimension, and more than one cause may apply.
Sponsor. Approve any bot control or filter change before it is applied, with a legitimate-traffic check first. None of it should be inferred from aggregate analytics alone.

Phase 1B diagnoses the event and specifies what a corrected key event would need. Any tag change is separately authorized implementation.

Observation 02

Historical edits to the measurement estate involve accounts outside the company domain

The Tag Manager container's version history shows four editing accounts: two personal Gmail addresses, one external vendor domain and one shared Gmail mailbox. That same shared mailbox appears as the primary owner of the Business Profile account holding the 75 listings. No company-domain account appears in the Tag Manager edit history, and none was visible as a Business Profile owner at the access level granted to the consultant. Two unused verification tokens sit on the Search Console property. The consultant's GA4 tier appears to be Editor, so GA4 administrators could not be listed.

Why it matters. Edit history is not a permissions inventory and an email domain does not prove who controls an account. What it does show is that the company has no documented, company-controlled record of who can change what is measured or who owns the listings. Until an administrator produces that inventory, the exposure cannot be sized.

Decision for the sponsorAccess review proposed for Phase 1B

GTM version history, GBP Manager, Search Console settings, GA4 Admin, read 8 Sep 2026. Certain on what was displayed; current permissions unverified.

how_to_voteNeeds a decision from Dr. Drut first

Sponsor. Ask IT, for GA4 and Tag Manager, and whoever holds the Business Profile organization to document current ownership, roles and verification dependencies across all three brands.
IT. With that inventory in hand, propose company-controlled administrator access, plus any demotions or removals, for the sponsor to approve.
IT. Add the new access before removing any of the old, so nothing is locked out.
IT. Leave the two Search Console verification tokens in place until their dependencies are known.

Effort and dependencies are for IT to assess. This is configuration rather than website code, but it is not free of risk.

Observation 03

Event data retention is set to two months, and only the client can change it

GA4 event data retention on the Diamond Braces property is set to two months. Under Google's documented behaviour, event-level data older than the retention period is removed through monthly processing; standard aggregated reports are not affected, but explorations, funnels and cohorts are limited to the retained window. Any such analysis in Phase 1B would have at most two months of event history behind it. Changing the setting is an administrative choice; under the engagement's no-changes principle the consultant will not touch it.

Why it matters. This is the one item where waiting has a direct cost. Every month the setting stays, another month of event-level history becomes unavailable to a measurement audit.

Decision for the sponsorConfiguration, not website code

GA4 Admin, Data retention, read 8 Sep 2026. Certain on the setting; retention mechanics per Google documentation.

how_to_voteNeeds a decision from Dr. Drut first

Sponsor. Decide whether to raise GA4 event data retention from two months to 14 months.
IT. On approval, change the setting in GA4 Admin under Data settings, and enable reset on new activity.
IT. Record the date of the change, because Phase 1B needs to know where usable event history starts.

The change does not restore data already removed. IT confirms there are no dependencies before making it.

Observation 04

Two brands have measurement gaps, and Smilify's business model is also unconfirmed

The KidCo GA4 property showed no recorded data in its all-time view on 8 September. Smilify's property recorded 7 users and 24 events with no key events in the seven-day view, no client-side analytics request was observed on its homepage during the test, and no Search Console property or Tag Manager container for it was visible in the accessible accounts. The engagement briefing described Smilify as an online-ordering alternative to Invisalign; the live site offers an in-office visit at $2,800 with a phone number and a candidacy check, no cart, and a footer dated 2020.

Why it matters. Nobody can say today what KidCo or Smilify marketing produces. KidCo looks like an instrumentation problem. Smilify needs a statement of what the business is, and where the in-office service is delivered, before any path can be mapped or any tag placed; that is a leadership decision, not a technical one.

Decision for the sponsorInstrumentation diagnosis in Phase 1B

GA4 properties, network traces under the stated test conditions, site walk, read 8 Sep 2026; briefing as supplied. Certain on what was observed; other collection paths not verified.

how_to_voteNeeds a decision from Dr. Drut first

Sponsor. State in one paragraph what Smilify sells, where it is delivered, and how a patient starts. Nothing else about Smilify can proceed without it.
IT. For KidCo, check which measurement ID the kidcodent.dental theme carries against G-JNNWTJMJ6Z, and which accounts hold its analytics.
IT. Propose the KidCo fix, and the Smilify analytics and Search Console setup, for approval. IT reported on 9 September that only Diamond Braces appears in Tag Manager.
Phase 1B. Maps the Smilify conversion path against the stated business model once that model exists.

Effort and dependencies are for IT to assess.

Observation 05

The internal dashboards decision will inherit every gap in the sources

On 31 August the company decided to build its BI dashboards internally. The sources those dashboards would draw from are the ones described on the Phase 1A page: sessions of which 28% are classified Unassigned, recorded conversions that are half unexplained, a Business Profile channel scattered across 29 spellings, and a Semrush connection to a GA4 property named Diamond Braces - Main Site whose displayed top pages are all UltiPro recruiting paths. The team's weekly reports were not reaching the people reviewing the work until 8 September, by the team's account.

Why it matters. A dashboard makes whatever feeds it look authoritative. Built now, it would publish the gaps with a chart on top. The order that protects the investment is: define the KPIs and resolve the sources first, then build.

Sequencing, Phase 1BClient decision of 31 Aug

Client decision record; GA4; Semrush GA4 connection read directly 9 Sep 2026 (top pages only; property ID and hostname distribution not recorded). Certain on what was displayed.

how_to_voteNeeds a decision from Dr. Drut first

Sponsor. Decide whether to pause the GA4-fed part of the dashboard build until Phase 1B has defined the KPIs.
IT. Record the numeric property ID and reporting period of the GA4 property Semrush is connected to.
IT. If that is not the website property, propose reconnecting Semrush to property 429695085.
Ortho2 administrator. Grant reporting access, so consultation and start counts by location can be validated as an interim source.

Ortho2 has not been inspected. Whether it can serve as that source is subject to access and validation.

Observation 06

The Business Profile channel is partly invisible to website analytics, and the planned link change makes it more so

Business Profile traffic that reaches the website is one visit in ten and one key event in four, and its three largest source rows record higher session key-event rates than Direct, Google organic and the site average, as measured so far. GA4 sees only the listings that link to the website. The SEO team said on 8 September that listings linking straight to Ortho2 booking book more, and that it intends to move every listing to a booking link. When that happens, Business Profile sessions in GA4 will fall and the channel will look, from GA4 alone, as though it has shrunk.

Why it matters. The change may well be right for patients. It has to be read alongside Ortho2 booking counts and whatever Business Profile reports for these listings, or it will look like a loss in every report the company has today. Nobody currently owns that reconciliation. Which Business Profile metric is available is not a detail: Google documents a Bookings metric that counts completed bookings and is available only where bookings run through an integrated booking provider, and a separate Booking clicks metric that applies to hotels only. A dental or orthodontic listing therefore either reports completed bookings, if a provider is integrated, or reports nothing about bookings at all and records the click as a website click.

Phase 4 oversightAttribution feasibility in Phase 1B

GA4 read 9 Sep 2026; team statements 8 Sep 2026. Certain on the GA4 rows; Client-supplied on which listings link where and on their booking performance, not independently verified.

boltThe team can act on this now

SEO team. Open one listing that links directly to booking in Business Profile Manager and check whether its Performance report shows a Bookings row. That single screen decides everything below it.
SEO team. Export six months of Business Profile Performance per listing, covering calls, direction requests, website clicks, messages and bookings where that metric is present, before any link changes, so a before picture exists.
Ortho2 administrator. Confirm whether the booking link can carry a per-listing identifier that survives onto the pending list.
Sponsor. If it cannot, decide whether to hold the website links until Phase 1B has assessed a join.
SEO team. Standardize every remaining website link on one tag convention: utm_source=gbp, utm_medium=organic, utm_campaign= the listing slug.
Weekly session. From the week the links change, read Business Profile Performance, GA4 and Ortho2 bookings side by side.

Google documents Bookings as completed bookings, available only where an integrated booking provider is in place, and documents Booking clicks as applying to hotels only. Step one establishes which of the two applies here.

Observation 07

Search volume is heavily branded, and local non-brand demand is not measured

Of roughly 389,000 Search Console clicks in 16 months, 88,104 (about 22.6%) are the exact query diamond braces. The next largest displayed queries are brand-plus-task queries: diamond braces patient portal, diamond braces login and diamond braces appointment, whose patient-status mix has not been established. A complete branded and non-branded classification has not been done. The blog ranks for informational questions such as wire poking and kids' toothpaste; its own geography and key-event performance have not been isolated, and Semrush's tracked keyword set is dominated by "can I eat X with braces" style questions.

Why it matters. The available query sample does not show how much search activity represents new-patient demand, and nothing on the estate measures local non-brand demand by office. Until that is measured, search performance cannot be judged against new-patient acquisition.

Baseline in Phase 1CPhase 4 oversight

Search Console 16 months, read 8 Sep 2026; GA4 Region, read 9 Sep 2026; Semrush Position Tracking, read 9 Sep 2026. Certain on the displayed rows; the full query distribution is not in this portal.

boltThe team can act on this now

SEO team. Report Search Console branded and non-branded query performance separately, with a documented classification rule.
SEO team. Report GA4 geography separately from query data. Search Console has no state dimension, so a query-by-service-area breakdown needs a valid method established before it is attempted.
SEO team. Add a location-scoped keyword set per office in Semrush Position Tracking, using the terms a new patient actually types.
Sponsor with the SEO team. Set content priorities. The evidence here supports separating informational blog traffic from local acquisition in reporting; it does not set a production quota.

Phase 1C takes the location-scoped set as a ranking baseline once that phase is authorized.

Observation 08

The team's own tool flags a toxic link profile; remediation status is unknown

Semrush rates the Diamond Braces backlink profile High toxicity: 278 toxic referring domains, 16.1% of the 1,732 analyzed, in an audit last run on 27 August 2026. The most toxic sources listed are throwaway web.app subdomains that appear to carry copies of the blog's own articles, plus an aged-domain seller scored 100. IT recalled on 9 September that earlier developers had tried to address this on the custom platform, without a dedicated SEO function at the time; what was done then, and whether a disavow file already exists, has not been checked.

Why it matters. A flagged link profile is a risk to assess, not a diagnosis. Intent, ranking impact, and any relationship to the decline the team describes are unverified. Acting on a screenshot or a summary score would be a mistake; so would ignoring it.

Authority baseline in Phase 1CRemediation history unknown

Semrush Backlink Audit, read directly 9 Sep 2026. Certain on the figures as shown; Guessing on scraping or negative SEO as the cause; Client-supplied on the remediation history.

boltThe team can act on this now

SEO team. Export the full Backlink Audit list with scores, first-seen dates and live status, together with the Search Console Links report, any existing disavow file, and any manual-action history.
Consultant. Review that set under Phase 4 before anything is submitted anywhere.
SEO team. For the web.app copies, verify ownership of the copied content before treating them as infringement, then choose between removal requests and no action.
SEO team. For domains linking with manipulative anchors, decide whether a disavow is warranted at all. Google states it ignores most spam links, and a new disavow upload replaces any existing list.
Client. If one is warranted, submit it at the domain level for those specific domains only, never for web.app as a whole, and keep the file under version control.
Oversight. Add a monthly Backlink Audit check to the routine, with the formal baseline taken in Phase 1C.

Observation 09

AI assistants are a recorded traffic source, and their place in routine reporting is unconfirmed

In 28 days, 342 sessions arrived from the AI Assistant channel, 327 of them from ChatGPT, and they recorded 72 key events, a session key-event rate of 13.16% against a site rate of 13.71%. Semrush counts 316 mentions of Diamond Braces across ChatGPT, AI Overviews, AI Mode and Gemini, and 451 cited pages. GA4 and Semrush therefore already measure this; whether the weekly reports include it is unknown because those reports have not been reviewed line by line.

Why it matters. The snapshot is small and its growth has not been measured here. A dated series started now costs little; one started later cannot be backfilled. Patient outcomes from this traffic are unknown.

Baseline in Phase 1C

GA4 Traffic acquisition and Semrush AI Visibility, both read 9 Sep 2026. Certain on the figures; reporting practice unverified.

boltThe team can act on this now

SEO team. Confirm whether the weekly report already carries the GA4 AI Assistant channel and the Semrush AI Visibility score.
SEO team. Add them if it does not, so a dated series exists from now on.
Consultant. Treat this 28-day read as a provisional Phase 1A snapshot, not the Phase 1C baseline.

Indexing of location pages, structured data and plain-language answers on service pages are hypotheses about what improves assistant citation. They need page-level evidence and are not guaranteed fixes.

Observation 10

The scope the SEO function is asked to cover appears larger than its capacity, by its own account and by the evidence

Two specialists, with a third person on reviews and social, cover 75 listings, three websites, content production of several long posts a week, and weekly reporting. The team described that load as more than it can carry well. The evidence is consistent with the description without judging anyone: 239 unreplied reviews by the team's count, indexing validations not started on most reason rows, listing links not yet reviewed one by one, template tokens visible on public posts on 9 September with first appearance unknown.

Why it matters. Oversight can only ask for what the function has the capacity to do. The sponsor decides whether the scope is right, the capacity is right, or the priorities are. This portal records the condition of the function; it does not evaluate individuals, and the consultant does not assign work.

Phase 4 oversightDecision for the sponsor

Working session 8 Sep 2026; live pages 9 Sep 2026. Client-supplied on load and backlog; Certain on what was observed on the properties.

how_to_voteNeeds a decision from Dr. Drut first

Sponsor. Ask the SEO team and IT for estimated effort and measurable outcomes for each main workstream: listings and reviews, indexing fixes, and content.
Sponsor. Choose between narrowing the scope, adding capacity, or resetting priorities, and set the order of work.
Phase 4. Reviews the resulting plan and its evidence. Client management assigns the work and sets output.

One piece of evidence bears on the choice: the Business Profile channel records higher key-event rates than the other measured rows, and both the review backlog and the listing link work sit in that channel.

Observation 11

Another migration without a quality gate carries the same risk the last one is described as having carried

A platform decision is open: location page creation is paused and a move off Shopify is under discussion. Search Console currently reports 1,350 of 2,249 known URLs as not indexed, with 715 classified as redirects and 298 as alternate or duplicate canonicals; location pages sit under three URL patterns and the office locator on two addresses. Which of those URLs should be indexed, and how much of the pattern dates from the migration, has not been established. The team describes previous migrations as built by outsourced developers without a quality-assurance step, and the live blog posts, with unrendered template tokens in public view on 9 September, are consistent with that account.

Why it matters. The platform choice is the company's to make. Whatever it is, release without verification is the pattern to break. A migration plan with a redirect map, a crawl before and after, and someone accountable for sign-off is what protects the search presence through the change.

Phase 4: high-risk decision reviewDecision for the sponsor

Search Console 8 Sep 2026; client decision 1 Sep 2026; team account 8 Sep 2026; live pages 9 Sep 2026. Certain on what was displayed; Client-supplied on migration history.

how_to_voteNeeds a decision from Dr. Drut first

Sponsor. Make a release gate a condition of the developer contract, whatever platform is chosen.
Developer. Build a full URL inventory and 301 map from a crawl of the current site plus the URLs Search Console already knows.
Developer. Sign off a staging crawl before launch.
Developer. Run a launch-day and a day-seven crawl comparing status codes, canonicals and titles against the map.
SEO team. Validate every fixed issue in Search Console.
Sponsor. Name the one person who signs off.

The consultant reviews the plan and the crawls under Phase 4 as a high-risk technical decision, and implements nothing.

Observation 12

Existing-patient activity is mixed into the numbers the company calls traffic and leads, and its share is unmeasured

Two missed-appointment policy pages are the third and seventh most viewed pages on the site, 4.8% of all views, ahead of the pricing page. The largest queries after the exact brand term are diamond braces patient portal, diamond braces login and diamond braces appointment. The booking modal offers new-patient, existing-patient and reschedule routes with office-specific Ortho2 codes, and all of them fire the same GA4 key event. The internal-traffic filter sits in Testing, so it does not currently exclude matching traffic. Campaign names such as button_list_FindaLocation and Site-Promo suggest internal elements carry campaign tags, which can affect event-level attribution.

Why it matters. Existing-patient activity is present in traffic, queries and key events, and its share of users, sessions and key events has not been measured. Until prospects and patients are separated in the measurement, every growth target set against site traffic or key events is partly a target for patient self-service. The separation is a definition problem for Phase 1B's KPI work.

KPI definitions in Phase 1B

GA4 Pages, Session campaign and Admin; Search Console; site walk. Read 8 and 9 Sep 2026. Certain on what was displayed; Likely on the internal tagging, pending theme inspection.

scheduleWaits on Phase 1B authorization

Phase 1B. Inventory every office-specific and general Ortho2 code, so new, existing and reschedule routes can be counted apart.
Phase 1B. Build a provisional patient segment from portal, login and policy-page activity, reported alongside acquisition until its accuracy is validated.
Phase 1B. Confirm the internal-traffic filter’s rules and, if they are correct, propose activating it.
Phase 1B. Confirm in the Shopify theme whether internal elements carry campaign tags, and propose removing them if they do.

All four are specified for separate authorization. Proposed segments stay provisional until event accuracy and exclusion effects are validated; only then would in-area, sourced, non-patient sessions and the appointments they produce become the number the company manages.

infoHow to use this page

Your feedback, and what happens to it

Three choices and a comment

Under each observation, pick Agree, Not sure or Disagree, add a comment if you want to, and press Send. Disagreement is the most useful answer; it is how a wrong reading gets corrected before a proposed action is taken further.

Where it goes

Successful delivery is confirmed on screen, with the observation named, and the response is kept in the shared record with the chat. If delivery fails, the response stays saved on this device and the screen says so. Replies come back in the Ask Jeremy chat at the bottom right of this page.

What it is not

Feedback refines these observations and proposed actions; it does not approve implementation, authorize any phase, or accept the Phase 1A deliverables. Those remain written decisions under the SOW.

Engagement Portal

Where do your patients come from? Today, nothing on the estate can answer that.

How far a marketing touch can be followed

What we know today

how_to_voteWhat needs a decision from you

What we can prove, and what we cannot prove yet

Top channels by visits

Key-event rate by campaign

The three largest Business Profile source rows record higher session key-event rates than the rows compared with them, and this is the channel measured least completely.

26.0% of reported active users were assigned to the four service-area states, and about half of the recorded key events fall outside those rows.

Two brands have measurement gaps, and one of them also has an unconfirmed business model.

A large amount of measurement work sits drafted and unpublished, with no recorded owner.

Search Console reports 1,350 known URLs as not indexed, and the reasons have not been triaged.

Nothing today connects a marketing touch to a patient.

blockWhat cannot be concluded yet

searchSurfaced in the SEO working session, 8 September

One portal for every phase

Review log

Three properties at a glance

Diamond Braces

KidCo Dental

Smilify

Where we are blind

Which office a lead came from

Whether a booking became a patient

Which marketing drove a phone call

Anything about KidCo online

Anything about Smilify

GBP profile actions

Anything older than two months

What the Get Started form produces

Internal and test traffic

Whether a key event is a booking

Channel performance, as measured so far

Key-event rate by session campaign

One channel, 29 source and medium spellings

Engagement rate by source

Why it is invisible: two tagging conventions

Where booking actions start

Key events by source

Key events by listing, Business Profile traffic only

ruleWhat this read changed

link_offWhat stands between this and "patients by source"

Every readable listing, 28 days

Traffic quality: what the headline number is made of

Sessions by default channel group

Engagement rate and time by channel

Inside vs outside the service area

Top regions by active users, with engagement

Several out-of-area states show key-event rates overlapping the service-area range

priority_highWhy this matters more than the Unassigned traffic

Countries

The headline number, layer by layer

Existing-patient activity is mixed into the traffic figures

Top pages by views, 28 days

The three signals

Conversion path map (1A.1)

Diamond Braces

KidCo Dental

Smilify

Phone numbers observed across the three sites

Can the same patient be counted twice?

Every Business Profile listing across NY, NJ, PA and CT

Listing register

GBP listings by group, brand and state

Listings by GBP group

Listings by state and brand

Instrumentation: what fires, what is drafted, what is unpublished

Key events configured on the Diamond Braces property

Tag Manager container GTM-KSXTJSZ

GA4 configuration defects

Google tag coverage

Drafted, not activated

Search Console: indexation and demand

Diamond Braces: indexed vs not indexed

Why 1,350 URLs are not indexed

Top queries, diamondbraces.com, 16 months

KidCo Dental: toothpaste queries lead the displayed sample

Access manifest (1A.3)

Progress