Singulariti Frøya Organics · Native Attribution Audit

CTV Attribution & Cross-Channel Review

Vision Bridge display-native claims, checked against Converge first-party data. 467 claimed orders, every one matched to its real path — and each order a drop-in you can open.
Credit taken: $0.00
Pushback — UTM fix in progress
Data through 2026-09-11
Executive Summary

467 claims. Zero credited.

Vision Bridge reported 467 native orders across eight days. We matched every one of them against Converge first-party data and found no Myosin UTM anywhere in the visible paths. Nothing has been paid out while the fix runs.

467
Orders claimed
Vision Bridge reports, 09/02–09/08, device-level IDs
141
Verified to the minute
Converge journey capture, real order timestamps (ET)
46
With warehouse sessions
UTM-level event stream — the basis of the 27 / 7 / 6 readout
13
Potentially attributable
7 Direct + 6 Organic — no Myosin UTM to prove them yet
27
Last touch: Meta
Meta takes 0% by rule — no credit there
$0.00
Credit taken
Balance with the publisher while the UTM fix runs
$0.00
Credit taken to date. Publishers claimed 467. 46 show a visible pre-order sequence: 27 ended on Facebook — no credit there by the 0% rule. 7 Direct and 6 Organic — 13 that could have been attributed if the UTM map had held. UTMs were being stripped client-side; the fix is in flight. We are pushing back on all publisher credit, tracking a week, and coming back with proper numbers.
Where the evidence sits
141 of 467 verified
141 — order-verified paths, minute precision
382 — confirmed in Vision Bridge's own live order table
326 — journey capture still queued
92% of verified dates land within ±1 day of the report date, and 99 of the 100 verified orders that overlap the VB table match to the day across the two independent sources.
What the data shows

Zero Myosin UTMs across all 467 visible paths — at both the visit level (141 verified) and the warehouse UTM level (46). When a tag fires, the pipeline captures it completely (Google's control sessions show full source/medium/campaign structure). Myosin's simply don't fire.

The claims' timing checks out — verified orders land within a day of the report date. The gap is not fabrication; it's missing tags.

UTMs dropped client-side

33 of the 46 session-having orders carry a UTM — zero carry Myosin's. The native path lands untagged. Client-side stripping, fix in flight. The 13 Direct/Organic are what we could credit if the map held.

Zero balance

27 of 46 end on Facebook — Meta takes 0%. Nothing is owed while we track a week and return with proper numbers.

141 verified to the minute

The journey rail verified 141 orders with real timestamps and full visit paths. 92% within ±1 day of the report date.

The 382: inside the warehouse gap

Every one of the 382 VB-confirmed conversions landed Aug 28 – Sep 9 — exactly the two-week hole in the warehouse copy (full export ends Aug 26; the live sync holds Sep 11 only). Not lost — the copy is behind. When the backfill lands, the page re-pulls the same tables and fills them in.

The orders exist — 382 confirmed

Pulled straight from Vision Bridge's live order table: 382 of 467 claims with ad-exposure date (entry Aug 28 – Sep 8), conversion date and revenue ($41,191). Median 16 hours from exposure to purchase. 99 of the 100 that overlap our verified set match Converge to the day.

The 85 not in VB's table

85 claims don't appear in Vision Bridge's live order table — all device-matched, all from the September file. Each is flagged in the explorer as no live profile / no VB row: the publisher follow-up list.

The rules we score by

Meta — 0% creditGoogle — 25% creditCPA — $65

Every order is a drop-in

Open any of the 467 in Customer Explorer — its live profile, its VB-table entry→conversion row, its verified path, its warehouse sessions, its Myosin UTM check.

Channel Attribution

Two layers, two methods

The 46 orders with warehouse sessions give the UTM-level readout behind the 27 / 13 numbers. The 141 verified journeys give the visit-level census. Different samples, different methods — each labeled, neither overruled by the other.

Layer 1 — Last touch, the 46 orders with warehouse sessions (UTM-level)
Facebook
27
Rule: 0% credit — Meta takes nothing
Direct
7
Potentially ours — could be the native path, no Myosin UTM to prove it yet
Organic
6
Potentially ours — same UTM gap
Klaviyo email
4
Internal email flows — welcome + prospect campaigns
Google
1
Rule: 25% credit if verified — free Shopping (product_sync)
Other
1
13 of the 46 (28%) — 7 Direct + 6 Organic — carry no Myosin UTM and could be the native path. That is the "13" we can defend if the UTM map holds.
Layer 2 — Visit census across the 141 verified paths (1,144 pre-purchase visits)
Meta Ads
507
44% of all pre-purchase visits across the verified paths
Direct
213
Could be native-path traffic — no UTM to prove it
Email
209
Internal Klaviyo flows
Google Ads
66
Rule: 25% credit if verified
Organic Search
63
Long tail: Paid Social 20 · TikTok 19 · Referral 19 · unattributed 19 · AppLovin 11 · Organic Social 10 · Pinterest 5 · Organic Shopping 2 · Organic Video 1. Final pre-purchase visit: Email 46 · Direct 34 · Google Ads 20 · Meta 15 · Organic Search 14 · unattributed 4 · Referral 3 · Paid Social 2 · AppLovin 2 · Organic Shopping 1.
Why the two layers differ

Layer 1 is the warehouse event stream — session-level, UTM fields intact, the basis of the 27 / 7 / 6 / 4 / 1 / 1 readout. It only exists where the warehouse held a matching session (46 orders).

Layer 2 is Converge journey capture — visit-level paths with minute-precision order times (141 orders). Meta shows up larger here because journeys see every touch, not just sessions that logged events.

The control: Google proves the pipeline works

Google Ads sessions in the same stream carry the full structure — utm_source=google, utm_medium, utm_campaign, cvg tags. When a partner's tags fire, we capture them completely. Myosin's tags simply don't fire — the gap is upstream, on the client side, and the fix is in flight.

Customer Explorer

Drop in on any claim

Every one of the 467 orders opens, mapped to where it actually lives: the live Converge profile (identity, first order, order session), Vision Bridge's own order table (ad-exposure date, conversion date, revenue), the verified visit path, and UTM-level warehouse sessions — per order.

The anchor case — 12337452056943
Verified — Converge journey
Actual order Sep 7, 2026, 9:19 AM ET (minute precision). 19 pre-purchase visits — all Meta. Publisher reported Sep 6: one day early, consistent.
Warehouse — UTM-level
8 sessions, Aug 8–11 window, all facebook / paid (cvg=fb). Zero Myosin UTMs in either source.
All 467
Verified 141
Warehouse 46
No profile
All 467 claimed orders

Click any row to open it in the explorer above. VB ✓ = in Vision Bridge's live order table · Entry→Order = ad exposure → conversion (turnaround) · Profile = live Converge profile status · V = verified order time · W = warehouse sessions · UTM ✓ = a Myosin UTM was found (none were).

OrderReportedFileMatchVBEntry→OrderProfileOrder time (verified)VisitsLast touch / visitWH sessionsUTM
What Happens Next

Track a week. Then settle it.

The balance stays at $0.00 while the fix runs. Five things have to land before we score week two against the rules.

01 · Now

UTM fix + tracking

Client-side UTM stripping is being fixed. One tracking week from deploy, then we re-score native vs Meta against the rules.

02 · Backfill

The Aug 27 – Sep 10 gap

The warehouse copy has a two-week hole (full export ends Aug 26; live sync holds Sep 11 only). All 382 VB-confirmed conversions sit inside it. When the backfill lands, these orders get warehouse sessions + order events and the table fills in.

03 · Capture

326 journeys pending

Journey capture is still queued. Every completed order adds a verified timestamp + path — and each one cross-checks against the VB conversion date already on this page.

04 · Publisher

The 85 flagged rows

85 claims aren't in Vision Bridge's live order table (all device-matched). The explorer flags each one — send the list back for line items: order dates, credit, and why the ID isn't in their feed.

05 · Re-pull

Page refreshes

This dashboard pulls from the same tables — rerun the click-path build and redeploy when the backfills land. Nothing is hardcoded.