Skip to main content
Card class: HeroCategory: Nerve Centre

At a glance

A real-time alert that fires when account or campaign ROAS falls materially below its recent same-day-of-week baseline. The trigger compares the latest reading against the same day of the week over the prior 7 days, so a Monday is judged against recent Mondays, not against the weekend. The hard part on Pinterest: the save-to-purchase lag means today’s ROAS is structurally under-reported, so a naive “ROAS is down” reading throws false alarms. This card watches the leading indicator (save rate, engagement volume) alongside the lagging ROAS number, so a genuine collapse is separated from the normal attribution-tail dip.
What it fires onLatest period ROAS down by more than roughly a quarter versus the prior 7-day same-day-of-week baseline, sustained rather than a single noisy data point. Threshold is configurable per account; the default is deliberately conservative to avoid alerting on the Pinterest attribution tail.
Why same-day-of-weekPinterest traffic and conversion behaviour is strongly weekly. Sunday-evening planning sessions and weekday lunchtime browsing produce different ROAS profiles. Comparing Monday-to-Sunday would fire constantly. Same-DOW comparison removes that noise.
Leading vs laggingROAS is the lagging number. Save rate, close-up rate, and engagement volume are leading. This card reads both: a ROAS dip with healthy saves is usually the attribution tail (will recover); a ROAS dip with collapsing saves is a real problem (creative fatigue, audience exhaustion, tracking break).
Attribution-tail handlingBecause Pinterest credits conversions on a 30-day click plus 30-day engagement window, the most recent days always read low and fill in later. The alert discounts the freshest 1-3 days and weights the same-DOW baseline so the tail does not trigger a false drop.
ScopeAccount-level by default, with per-campaign drill-down. Shopping, Standard Pin, Idea Pin video, Carousel, and Collections each carry their own ROAS profile, so a campaign-level fire is more actionable than an account roll-up.
UnitUnitless ratio (the ROAS itself) plus the percentage drop. The alert list shows both the current ROAS and the magnitude of the fall.
Time windowRT (real-time-ish, on the standard 4-6 hour ingestion cadence) against a prior-7-day same-DOW baseline.
Alert triggerROAS drop of more than approximately 25% versus the prior 7-day same-day-of-week reading.
Sentiment keypin_alert_roas_drop
Rolesowner, marketing, finance

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 running Pinterest Ads. Account currency: GBP. The alert evaluates the latest reading against the same weekday over the prior week.
DayROAS (this week)ROAS (prior same-DOW avg)Save rateFired?
Mon 08 Jun 265.9x6.1xsteadyNo (within noise)
Tue 09 Jun 266.0x6.2xsteadyNo
Wed 10 Jun 264.3x6.0xstill steadyYes, but tail-suspect
Thu 11 Jun 263.1x5.8xdropped 40%Yes, real
How to read it:
  1. Wednesday’s fire is the attribution tail, probably. ROAS fell 28% but save rate held. Pinterest’s 30-day engagement window means recent-day ROAS reads low and backfills. This card flags it but tags it tail-suspect because the leading indicator is healthy. Wait, do not pause.
  2. Thursday’s fire is real. ROAS fell 47% AND save rate fell 40% together. Two signals moving the same way is the pattern that matters. Saves are the leading indicator: if savers are walking away, the lagging ROAS will not recover on its own. Investigate now.
  3. The same-DOW baseline matters. If this card compared Thursday against Wednesday it would misread the normal weekday rhythm. Against recent Thursdays, a 47% fall is unambiguous.
  4. Always open the campaign drill-down. An account-level fire often hides a single runaway campaign (one Idea Pin video set that exhausted its audience) dragging the blended number while Shopping holds steady.
Quick triage:
  • ROAS down + saves steady = attribution tail, wait a few days before acting.
  • ROAS down + saves down = real demand or tracking problem, investigate now.
  • ROAS down + spend up = scaling past the efficient frontier; pull back the newest ad groups.
  • ROAS down + spend flat + clicks down = creative fatigue, refresh the Pin.
  • ROAS down + conversions reporting near zero = suspect a tracking break, check the tag and Conversions API card.

Sibling cards merchants should reference together

CardWhy pair it with this alert
Pinterest ROASThe metric this alert watches. Open it to see the absolute level, not just the drop.
Pinterest ROAS TrendThe daily series. Because Pinterest ROAS back-revises, the trend chart tells you whether a fired drop is recovering.
Pinterest ROAS by CampaignAn account-level fire usually traces to one campaign. Open this to find which.
Pinterest Conversion LagRising lag means a ROAS dip is artificially low right now but will recover. The tail-suspect flag leans on this.
Pinterest Conversion Drop AlertThe leading-indicator companion. Conversions falling confirms a ROAS drop is real, not tail noise.
Pinterest Tag + Conversions API Tracking BrokenA tracking break shows up as a sudden ROAS collapse with no demand cause. Rule it out first.
Pinterest Wasted SpendIf ROAS dropped because spend drifted to dead keywords, this is where you fix it without touching the numerator.
Google Ads ROASCross-platform sanity check. If Google ROAS dropped the same day, suspect a site-wide cause, not a Pinterest one.

Reconciling against Pinterest Ads Manager

Where to look in Pinterest Ads Manager: Pinterest Ads Manager > Reporting > Performance > set the date picker to the recent days in question and add the Spend and Total checkout value columns to derive ROAS. Pinterest Ads Manager has no native “ROAS dropped” alert with a same-day-of-week baseline, so you are reconstructing the comparison manually. Confirm the attribution model in Settings > Conversions matches the account default (30-day click + 30-day engagement) before comparing days. Why our number may legitimately differ from Pinterest’s UI:
ReasonDirectionWhy
Same-DOW baseline vs raw dayDifferent reference pointPinterest’s UI shows a flat day-over-day or period-over-period comparison. This card compares against the same weekday in the prior week, which is the right reference for Pinterest’s strongly weekly rhythm.
Attribution-tail discountingOurs may not fire when the UI looks downThe card down-weights the freshest 1-3 days because they always read low and backfill. The UI shows the raw, still-incomplete number.
Account time zone vs UTCBoundary days offPinterest reports in the ad account’s configured time zone; this card uses UTC for window edges, which can shift a single day’s reading.
Ingest lagOurs lower for last 4-6 hoursThe card refreshes on a 4-6 hour cadence; the UI is closer to real-time.
Cross-connector reconciliation: When this alert fires, confirm the drop against commerce-platform revenue tagged as Pinterest-sourced. A real ROAS collapse shows up as fewer Pinterest-attributed orders in the store; a tail-only dip does not.
CardExpected relationshipWhat causes legitimate divergence
shopify.total_revenue filtered to Pinterest UTMA real drop reduces Pinterest-tagged store revenue tooAttribution windows differ; the store uses last-touch, Pinterest uses a 30-day engagement window, so timing gaps are normal.
google_ads.gads_roasPeer ad-platform ROAS, NOT a reconciliationIf both platforms drop the same day, suspect a site-wide cause (checkout outage, stockout, price change).

Known limitations / merchant FAQs

Why did this alert fire when my ROAS always recovers a few days later? Because Pinterest’s 30-day click plus 30-day engagement attribution means the most recent days always read low and backfill as engagement-attributed conversions arrive. The card discounts the freshest days and tags those fires as tail-suspect, but it still surfaces them so you are aware. If the save rate held steady, treat a tail-suspect fire as informational and wait. If saves dropped alongside ROAS, it is real. What is the difference between a tail-suspect fire and a real fire? A tail-suspect fire is a ROAS drop where the leading indicators (save rate, engagement volume, clicks) are still healthy. The lagging ROAS is low only because conversions have not attributed yet. A real fire is a ROAS drop where leading indicators fell too, meaning fewer people are engaging and the revenue is genuinely not coming. Two signals moving together is the actionable pattern. Why same-day-of-week instead of yesterday? Pinterest behaviour is strongly weekly. Sunday-evening planning, weekday lunchtime browsing, and Saturday-morning project research all produce different ROAS. Comparing today against yesterday would fire on the normal weekly rhythm. Comparing today against recent same-weekdays isolates a genuine change. Can I change the 25% threshold? Yes, the threshold is configurable per account. Lower it (say to 15%) if you run a high-spend account where small swings matter; raise it if you want only severe collapses surfaced. Be cautious lowering it too far on Pinterest, the attribution tail alone can move recent-day ROAS by 10-20%. The alert fired but Pinterest Ads Manager shows ROAS is fine. Who is right? Usually both, measuring different things. The UI shows a raw, still-incomplete recent number and no same-DOW baseline. This card applies the baseline and the tail discount. Re-window the UI to the same dates and confirm the attribution model matches before assuming a discrepancy. Should I pause a campaign the moment this fires? No. On Pinterest, pausing on a recent-day ROAS dip is the single most common mistake. Check the save rate first. If saves are steady, the revenue is on its way. Only pause when ROAS and the leading indicators fall together and the campaign-level drill-down confirms a specific campaign is the cause.

Tracked live in Vortex IQ Nerve Centre

ROAS Dropped Below Threshold is one of hundreds of KPI pulses Vortex IQ tracks across Pinterest Ads and 70+ other ecommerce connectors. Nerve Centre runs the detection layer; Vortex Mind investigates the cause when something moves; Ask Viq lets you interrogate any number in plain English. Start for free or book a demo to see this metric running on your own data.