Measured: 2026-08-28 · Window: 2026-08-14 → 2026-08-28 (14 days) · Accounts: all 5 running v2
A like-for-like comparison of the two Littledata purchase connections in every account that runs them.
Both are read from the same GAQL rows over the same window, at the combined Purchase - Littledata
action only — the New Customer / Returning Customer variants are excluded from both sides. Metric is all_conversions / all_conversions_value.
Attribution is split by segments.conversion_attribution_event_type, whose values are UNKNOWN
(click), ENGAGED_VIEW (10s+ view of a skippable video ad) and IMPRESSION.
metrics.view_through_conversions covers only the last of those three and is not used here.
| Account | v1 action | v1 type | CT | VT | Primary | v2 action | v2 type | CT | VT | Primary |
|---|---|---|---|---|---|---|---|---|---|---|
soda-sense |
Purchase - Littledata |
UPLOAD_CLICKS |
90d | 1d | yes | Purchase - Littledata (Web) |
WEBPAGE |
90d | 30d | no |
becausemarket-prod |
Purchase - Littledata (9255055094) |
UPLOAD_CLICKS |
90d | 1d | no | Purchase - Littledata (Web) |
WEBPAGE |
90d | 30d | no |
jimmy-joy |
Purchase - Littledata |
UPLOAD_CLICKS |
28d | 1d | yes | Purchase - Littledata (Web) |
WEBPAGE |
90d | 30d | no |
gardeners-dream-online |
Purchase - Littledata |
UPLOAD_CLICKS |
90d | 1d | yes | Purchase - Littledata (Web) |
WEBPAGE |
90d | 30d | no |
sparklingjewels-com |
Purchase - Littledata ×2 (both REMOVED) |
UPLOAD_CLICKS |
30d | 1d | — | Purchase - Littledata (Web) |
WEBPAGE |
90d | 30d | no |
| Account | v1 conv | v1 value | v2 conv | v2 value | v2 ÷ v1 value | v2 ÷ v1 conv |
|---|---|---|---|---|---|---|
becausemarket-prod |
160.9 | 16,955 | 231.1 | 24,682 | 1.456× | 1.436× |
soda-sense |
700.6 | 37,209 | 721.4 | 42,532 | 1.143× | 1.030× |
gardeners-dream-online |
4,000.3 | 207,484 | 3,960.4 | 207,331 | 0.999× | 0.990× |
jimmy-joy |
1,398.7 | 103,818 | 1,109.8 | 81,893 | 0.789× | 0.793× |
sparklingjewels-com |
— | — | 9.0 | 562 | v2 only | — |
Median across the four comparable accounts: 1.071×. Values are in each account’s own reporting currency and are not converted, so rows are comparable within an account, not across them.
jimmy-joy is the one account where v2 reports materially less than v1, and its lookback windows rule
out the obvious explanation: v1 runs a 28-day click window against v2’s 90-day, so v2 has the wider
window on both axes and still reports 21% less revenue.
| Account | Click conv | Click value | Engaged-view conv | Engaged-view value | Impression conv | Impression value | View-side % of value | |
|---|---|---|---|---|---|---|---|---|
becausemarket-prod |
v1 | 156.6 | 16,687 | 4.3 | 268 | 0.0 | 0 | 1.6% |
| v2 | 193.2 | 22,044 | 35.9 | 2,596 | 2.0 | 42 | 10.7% | |
soda-sense |
v1 | 700.6 | 37,209 | 0.0 | 0 | 0.0 | 0 | 0.0% |
| v2 | 721.4 | 42,532 | 0.0 | 0 | 0.0 | 0 | 0.0% | |
jimmy-joy |
v1 | 1,394.7 | 103,596 | 4.0 | 222 | 0.0 | 0 | 0.2% |
| v2 | 1,098.9 | 81,165 | 10.9 | 728 | 0.0 | 0 | 0.9% | |
gardeners-dream-online |
v1 | 3,998.1 | 207,390 | 2.2 | 94 | 0.0 | 0 | 0.0% |
| v2 | 3,954.5 | 207,100 | 5.9 | 231 | 0.0 | 0 | 0.1% | |
sparklingjewels-com |
v2 | 9.0 | 562 | 0.0 | 0 | 0.0 | 0 | 0.0% |
v2 takes more view-side credit than v1 in every account that has any: 10× on
becausemarket-prod, 3× on jimmy-joy, 2.5× on gardeners-dream-online.
Against each account’s value difference:
| Account | v2 − v1 value | of which click-side | of which view-side |
|---|---|---|---|
becausemarket-prod |
+7,727 | +5,357 (69.3%) | +2,370 (30.7%) |
soda-sense |
+5,323 | +5,323 (100.0%) | 0 (0.0%) |
gardeners-dream-online |
−153 | −290 | +137 |
jimmy-joy |
−21,925 | −22,431 | +506 |
The two accounts where v2 reports more do so by different routes.
| Account | v1 value / conv | v2 value / conv | Change | v2 ÷ v1 conv | Reads as |
|---|---|---|---|---|---|
becausemarket-prod |
105.38 | 106.80 | +1.3% | 1.436× | difference is in conversions credited |
soda-sense |
53.11 | 58.96 | +11.0% | 1.030× | difference is in value per credited conversion |
jimmy-joy |
74.23 | 73.79 | −0.6% | 0.793× | difference is in conversions credited |
gardeners-dream-online |
51.87 | 52.35 | +0.9% | 0.990× | parity on both |
soda-sense is the only account where the difference sits in value per credited conversion rather
than in the number credited. It concentrates in two search campaigns — ga-us-2025-paid-search-bf-conv-… (+3,316) and
ga-us-2025-paid-search-bf-eng-c… (+1,887).
Shares are of served impressions by segments.ad_network_type, which reports the surface an
impression actually served on; Performance Max campaigns serve largely on search.
| Account | Impressions | Cost | View-eligible surfaces | Tactic tier | Spend by channel |
|---|---|---|---|---|---|
soda-sense |
210,177 | 23,727 | 58.9% | mixed | Search 23,022 (97%) · Demand Gen 699 (3%) |
sparklingjewels-com |
440,786 | 2,077 | 12.7% | search/shopping-led | — |
becausemarket-prod |
840,802 | 13,679 | 12.6% | search/shopping-led | PMax 8,235 (60%) · Search 3,675 · Shopping 1,769 |
jimmy-joy |
493,588 | 14,648 | 2.8% | search/shopping-led | — |
gardeners-dream-online |
7,809,707 | 38,400 | 0.8% | search/shopping-led | — |
The two are not separated by how long v2 has been running. Both crossed into steady state between
2026-08-10 and 2026-08-13 and are measured over the same window; gardeners-dream-online and
jimmy-joy have had v2 live since June and sit at 0.999× and 0.789×.
What the measurements do separate them on:
becausemarket-prod (1.456×) |
soda-sense (1.143×) |
|
|---|---|---|
| Where the difference sits | conversions credited (+43.6%) | value per credited conversion (+11.0%) |
| View-side share of v2’s value | 10.7% | 0.0% |
| View-eligible share of served impressions | 12.6% | 58.9% |
| View-surface spend | PMax 8,235, converting | Demand Gen 699 (2.9% of spend), converting nothing |
v1 action primaryForGoal / inMetric |
false / false | true / true |
View-eligible impression share and view-side credit run in opposite directions here.
soda-sense buys the most view inventory of any account in the cohort — 58.9%, from a Demand Gen
campaign serving YouTube 34.9% and Content 24.0% — and takes none of it: that campaign produces no
Littledata purchase conversions at all, of any attribution type. Every purchase row in the account is
search and click. becausemarket-prod buys the second least and takes 30.7% of its difference from it.
So the distinguishing variable in the measured data is not the presence of view inventory but whether
that inventory converts, with enough spend behind it to move a total. One confound cannot be separated
at n=2: becausemarket-prod is also the only account whose v1 action is non-primary and excluded from
the Conversions metric, which may affect it as an upload target independently of media mix.
scripts/screen-google-ads-v2-candidates.ts applies that criterion across the base, cheapest test
first: the audit’s stored campaign tactic puts enough served impressions on view-eligible surfaces
(free, from auditChecks); those surfaces carry a real share of spend; and they produce purchase
conversions with ENGAGED_VIEW or IMPRESSION attribution. The third test is on purchases
specifically — an account taking view-side credit only on page views and add-to-carts is the
soda-sense case.
223 of 587 not-yet-v2 accounts clear the first test at ≥15% view-eligible impressions. Of the 40 highest-ranked, 29 were readable and 8 pass all three (30 days to 2026-08-28, ≥$1,000 total spend, ≥10% view-surface spend, ≥1 view-side purchase conversion):
| Shop | Ad account | View-side purchase conv | All view-side conv | View spend | Top served networks |
|---|---|---|---|---|---|
mycarpe |
8021532545 | 8,346.9 | 23,175.9 | 1,365,583 | YouTube 94% · Search 6% |
we-are-fussy |
1372164662 | 569.2 | 11,042.2 | 30,601 | YouTube 90% · Search 8% · Discover 1% |
zesty-paws |
7837737546 | 372.2 | 7,528.4 | 102,151 | YouTube 85% · Search 12% · Discover 2% |
uaudio-prod |
1064252781 | 117.8 | 264.7 | 5,251 | YouTube 85% · Search 10% · Discover 5% |
walabot2 |
5771339606 | 46.0 | 355.4 | 130,811 | YouTube 90% · Search 4% · Discover 4% |
chomps |
7161353153 | 21.9 | 2,865.3 | 119,420 | YouTube 90% · Search 10% |
nira-retail |
7961285697 | 14.0 | 5,990.6 | 22,949 | YouTube 91% · Search 6% · Discover 3% |
tiptop-storefront |
1133838314 | 13.7 | 1,970.8 | 14,526 | Content 58% · YouTube 24% · Search 16% |
mycarpe carries more view-side purchase credit in 30 days than the entire current v2 cohort has
recorded, on $1.37M of view-surface spend at 81% of its total. we-are-fussy and zesty-paws are the
next closest analogues to becausemarket-prod at a larger scale.
Eleven of the 40 could not be read: no ld-shopify OAuth token resolved, or the token had expired.
npx tsx scripts/compare-ld-v1-v2-vtc-by-campaign.ts <shopName> 14 [--all-actions]
npx tsx scripts/screen-google-ads-v2-candidates.ts [--days=30] [--top=40]
Three GAQL queries against campaign over the window: conversions segmented by
segments.conversion_action, split by connection version via ldPurchaseConnectionVersion(); an
unsegmented impressions/cost query, because segmenting by conversion action duplicates impressions
across actions; and the same conversions segmented by segments.conversion_attribution_event_type.
--all-actions adds a whole-account table by conversion action and the served-network mix.
Related: conversion connection v2 · UPLOAD_CLICKS view-through attribution