At a glance
A real-time integrity alert that fires when Pinterest conversion tracking breaks. Pinterest measures conversions two ways: the browser-side Pinterest Tag and the server-side Conversions API (CAPI). Since iOS privacy changes, accurate attribution depends on both firing AND being de-duplicated by a shared event identifier so the same purchase is not counted twice. This card watches two failure modes: the de-duplication rate falling (which inflates conversions and produces a lying, too-good ROAS) and conversion volume collapsing against the recent expected baseline (which means events have stopped flowing at all). Either way, every downstream Pinterest number becomes untrustworthy until it is fixed.
| What it watches | Two signals. First, the share of conversions that arrive from both the Tag and CAPI and are correctly matched and de-duplicated. Second, conversion volume versus the recent expected baseline for the time of day and day of week. |
| Why dedup matters | The Tag (browser) and CAPI (server) often report the same purchase. They are matched on a shared event identifier so it counts once. If that identifier is missing or mismatched, the same order is counted twice, conversions inflate, and ROAS reads far higher than reality. A too-good ROAS is as dangerous as a collapse: you scale spend on a number that is not real. |
| Why volume matters | If the Tag is removed from the site, a tag manager publish breaks it, or the CAPI server stops sending, conversion volume falls toward zero. Catching this within minutes prevents a day of blind spend and prevents the optimisation algorithm from learning on bad data. |
| iOS context | Browser-side measurement is heavily degraded on iOS by privacy restrictions. CAPI exists to recover that signal server-side. The two together, de-duplicated, is the only reliable post-iOS setup. This card effectively monitors the health of that recovery. |
| Failure modes caught | Dedup rate falling materially below healthy (doubled conversions), conversion volume falling materially below the expected baseline for a sustained window (events stopped), and a sudden divergence between Tag and CAPI counts. |
| Unit | Count and percentage. The alert list shows the dedup rate, the expected-versus-actual conversion ratio, and the affected window. |
| Time window | RT (real-time-ish, evaluated continuously on the standard ingestion cadence, with a short sustained window before firing to avoid single-blip noise). |
| Alert trigger | De-duplication match rate falling below roughly 70%, OR conversions running below about half of the expected baseline for a sustained window of around half an hour. |
| Sentiment key | pin_alert_tracking_broken |
| Roles | owner, marketing, engineering |
Calculation
Calculated automatically from your Pinterest Ads data. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading.Worked example
A UK home decor brand on Shopify with both the Pinterest Tag and the Conversions API enabled. The alert evaluates dedup health and volume continuously.| Time (UTC) | Dedup match rate | Conversions vs expected | What happened |
|---|---|---|---|
| 09:00, 18 Jun 26 | 94% | 100% | Healthy. Tag and CAPI matched cleanly. |
| 14:20, 18 Jun 26 | 92% | 98% | Healthy. |
| 15:05, 18 Jun 26 | 41% | 188% | Dedup collapsed, conversions doubled. A theme update changed the event payload so the shared event identifier stopped matching. |
| 15:40, 18 Jun 26 | 41% | 191% | Sustained. Alert fires. ROAS now reads ~2x reality. |
- The doubled-conversions case is the silent killer. Conversions jumped to 188% of expected because the same purchase is counted by both the Tag and CAPI without de-duplication. ROAS looks great, so nobody notices, and spend gets scaled on a fantasy. This card catches it precisely because the dedup rate dropped, not because the number looked bad.
- The cause is almost always a deploy. A theme update, a tag manager republish, a checkout change, or an app update altered the event payload so the shared identifier is missing or mismatched. The fix is to restore the identifier on both sides so events match again.
- Volume collapse is the other half. If instead conversions had fallen toward zero, the cause would be the Tag removed or the CAPI server down. Same card, opposite symptom, both equally urgent.
- Every Pinterest number is suspect while this is firing. ROAS, CPA, conversion rate, and all the cross-channel cards inherit the bad conversion data. Treat them as unreliable until the alert clears and a backfill window passes.
- Dedup down + conversions up = double-counting. Restore the shared event identifier on Tag and CAPI.
- Conversions down toward zero = events stopped. Check the Tag is present and the CAPI server is sending.
- Tag count healthy but CAPI count zero = server-side integration broke. Check the CAPI connection and access token.
- CAPI count healthy but Tag count zero = browser tag removed or blocked. Check the theme and tag manager.
- Both healthy again = wait for the backfill window before trusting ROAS, then resume normal optimisation.
Sibling cards merchants should reference together
| Card | Why pair it with this alert |
|---|---|
| Pinterest ROAS | The number this alert protects. A broken tracking setup makes ROAS lie in either direction. |
| Pinterest ROAS Dropped Below Threshold | A volume-collapse tracking break shows up there as a sudden ROAS drop with no demand cause. Rule tracking out first. |
| Pinterest All Conversions | The conversion count this card watches. A spike or collapse here is the symptom. |
| Pinterest Conversion Actions Breakdown | Shows which event types (checkout, add-to-cart, page visit) are mis-firing, narrowing the diagnosis. |
| Pinterest Conversion Lag | Helps separate a real tracking gap from the normal attribution tail. |
| Pinterest Total Revenue | Pinterest-reported revenue depends entirely on the conversion value events this card monitors. |
| Shopify Total Revenue | The ground truth. Compare Pinterest-tagged store revenue to Pinterest’s reported value to confirm a break. |
Reconciling against Pinterest Ads Manager
Where to look in Pinterest Ads Manager: Pinterest Ads Manager > Conversions > the conversion sources and event health view shows Tag events, CAPI events, and the deduplicated total. Pinterest’s own diagnostics flag low match rates and missing event identifiers there. Cross-check with the source platform too: in Shopify or your tag manager, confirm the Pinterest Tag is present on the checkout and that the CAPI app or connection is active and authorised. Why our number may legitimately differ from Pinterest’s UI:| Reason | Direction | Why |
|---|---|---|
| Real-time vs batched diagnostics | Ours fires first | This card evaluates continuously and fires within minutes of a sustained break. Pinterest’s event health view can lag, especially for CAPI server events that are batched. |
| Sustained-window guard | Ours may delay slightly | The card waits for a short sustained window before firing to avoid alerting on a single blip, so a momentary glitch that self-corrects will not fire. |
| Backfill after a fix | Both rise together | Once the identifier is restored, late events backfill and both the UI and this card recover over the following ingestion windows. |
| Account time zone vs UTC | Boundary timing | Pinterest’s health view uses the account time zone; this card uses UTC for window edges. |
| Card | Expected relationship | What causes legitimate divergence |
|---|---|---|
shopify.total_revenue filtered to Pinterest UTM | Pinterest-reported value should sit within roughly 30% of Pinterest-tagged store revenue | A dedup break pushes Pinterest far above the store figure; a volume break pushes it far below. A normal 10-30% gap is attribution timing, not a break. |
google_analytics.ga_revenue_by_channel | GA4 Pinterest channel revenue is an independent third reference | If Pinterest and GA4 both look sane but the store does not, suspect the store tag, not Pinterest. |