Each of these is a real decision Pinterest is working through, written so you can make the call yourself — incomplete information, someone who pushes back, no answer key.
Pinterest's products serve hundreds of millions of pinners, creators, advertisers and merchants around the world. Pinterest has a Pinterest Assistant team building a visual-first, AI-powered companion that helps Pinners go from inspiration to action across shopping, search, and discovery. For advertisers, Pinterest supports automatic bidding, which maximizes the number of clicks, conversions and GMV, as well as target average bidding with CPA or ROAS targets for its ads. Pinterest has a CTV advertising platform purpose-built for performance marketers. Pinterest runs shared infrastructure that powers both Generative AI and Recommender Systems across Pinterest's entire product suite.
Your first one is free. No card.
41 live right now.
The rest
Group feedback into traceable themes during Visual Search Ads beta
PMs are treating similar advertiser concerns as separate requests because the evidence arrives in disconnected forms. You must weigh fast theme detection against the review needed before feedback shapes product decisions.
Reconcile catalog dependencies while owners dispute what moved
Conflicting dependency records can delay catalog changes that keep product Pins current for visual planners. You must decide which source changes AI can surface automatically and which judgments stay with program managers.
Find the cause of blank pages on return visits before release review
Visual planners hit blank board pages after certain return visits, and engineers cannot yet agree on the failure chain. You must decide what the system can rank from evidence and what engineers must confirm before shipping.
Choose what to rank first when a gift Pin tops the budget
Planners keep returning to the same gift searches after finding product Pins beyond their budget. You must weigh visual fidelity against retailer-ready alternatives when product details are uneven.
Decide if a 5.1% lift is buyer intent or late reporting
The release candidate shows a 5.1% offline lift, yet 62% of that gain sits in late-reported purchases. You have to weigh seasonal bidding upside against committing budgets to a signal that may be delayed reporting.
Calibrate notification scores before another low-relevance send
Predicted high-intent recipients rose 14% while observed opens fell 9% across three sends. You have to weigh a fast score correction against preserving notification reach and opt-out safety.
Prepare account reviews without guessing at campaign changes
Enterprise account managers are rebuilding campaign context while advertisers ask for explanations they can defend. You must weigh faster preparation against the review needed before a campaign explanation reaches a client.
Build an adoption path for finance advertisers while measurement varies
Financial-services advertisers are withholding performance budgets because campaign choices do not yet connect clearly to qualified customers. You must weigh a repeatable adoption path against the measurement and compliance limits each advertiser brings.
Choose a bidding model before holiday budget locks
The advertiser has five days to lock a holiday acquisition plan while platform and first-party conversions disagree. You must weigh faster automated learning against the proof finance needs to release the remaining budget.
Rebuild catalog sales before Friday’s luxury advertiser review
Catalog sales have fallen below the advertiser’s 4.0x ROAS target after a creative refresh. You have to weigh a fast campaign change against stable delivery and a recommendation the advertiser can trust.
Prepare QBRs without losing the evidence behind each recommendation
Enterprise account leads are finding important campaign questions after QBRs have already started. You must weigh faster preparation against the human review needed to keep client advice credible.
Move 12 retailers into a first test without bespoke consulting
Retail advertisers are interested in Visual Search Ads but cannot translate existing product imagery into a confident first test. You have to weigh fast cohort learning against guidance that stays credible and repeatable.
Move 24 campaigns to SALES once finance can trace the revenue event
Twenty-four conversion campaigns must move while finance cannot reconcile the revenue event in its budget review. You have to weigh seasonal delivery continuity against a measurement claim the client can defend.
Align catalog retargeting while seasonal products miss the results
The campaign is spending, yet its seasonal catalog products are absent from retargeting results. You have to weigh a quick configuration reset against protecting the learning already underway.
Warn advertisers of weak budgets before they pause in week two
Thirty-eight percent of new Performance+ campaigns start below 10 weekly optimization events and often pause before week two. You must weigh a useful early budget recommendation against overclaiming what historical delivery can predict for a new campaign.
Decide if attribution stays in bidding before the 36-hour backfill
Attributed conversions rose 14%, but audited samples disagree across app surfaces and match confidence. You must weigh wider conversion visibility against a reporting history and bidding signal that cannot be cheaply corrected.
Steady ROAS pacing after delayed conversions distort the bid loop
Forty-two target ROAS campaigns are swinging up to 18% away from hourly pacing targets. You must weigh faster response to real demand against treating delayed measurement as a demand change.
Sort PMNC feedback before product evidence goes stale
PMNC feedback keeps arriving in disconnected threads, forcing PMs to reconstruct evidence before they can judge a pattern. You must decide what AI assembles, what a PM verifies, and how the team runs the loop afterward.
Prove new-customer value when finance questions Performance+ results
Performance marketers cannot turn healthy conversion totals into a credible new-customer answer for finance. You must weigh a useful budget-review view against what the available signals can honestly support.
Choose the SALES cutover while campaign outcomes conflict
Active Performance+ campaigns can keep spending while reporting names a different optimization event. You must weigh migration continuity against measurement credibility before legacy objectives retire.
Correct creative warnings while Visual Search budgets are still scaling
Affected Visual Search campaigns receive creative-mix warnings even when their programmatic variants are distinct. You must weigh clearer guidance for buyers against stable quality and diversity guardrails.
Trace attribution gaps without slowing App Promotion releases
App Promotion investigations are producing conflicting explanations for gaps between MMP totals and Ads Manager reports. You have to weigh faster evidence assembly against the review needed before engineers act on it.
Surface signal readiness before advertisers commit to ROAS goals
Advertisers reach ROAS-goal setup without knowing whether their conversion signals can support it. You must weigh a clear setup path against guidance that stays faithful to the underlying signals.
Commit the SALES cutover while reports and delivery disagree
Migrated campaigns can keep spending while reporting a different optimization event. You must weigh a fixed retirement date against a cutover that advertisers can still trust.
Fix publishing for SALES campaigns without weakening validation
Eligible SALES campaigns fail at publication after advertisers add supported conversion types. You must weigh a fast compatibility change against keeping validation trustworthy for every existing objective.
Launch a pilot of purchase reports once privacy review clears
Shopping creators cannot see whether product Pins drive purchases beyond saves and retailer visits. You must weigh early creator value against attribution limits, privacy review, and a pilot that can be trusted.
Decide on Hive retirement while 4 of 38 datasets stay disputed
Four of 38 critical datasets show inconsistent product attribute histories despite a 17% compute reduction. You must weigh irreversible savings against relevance risk that is expensive to detect after cutover.
Align Iceberg readiness before Friday’s migration window closes
Eleven producers appear complete in the dashboard, but only seven pass migration checks. You must weigh a timely wave against publishing a readiness view people cannot trust.
Reveal product evaluation after visual search, not just clicks
Visual search engagement is rising, but no shared dataset shows whether people progress to product evaluation. You have to weigh a narrow, decision-ready first version against a broader journey definition that may take longer to trust.
Decide the cutover date for creator rewards despite an 11% lift
A cutover rehearsal added 11% more modeled purchases, but retailer-confirmed samples do not match the increase. You have to weigh payment timing against evidence strong enough to defend every attributed reward.
Correct Homefeed exposures before the morning experiment review
Exposure counts rose 7.4% after an instrumentation change, distorting active experiment reads. You have to weigh preserving new event detail against restoring trusted daily measurement.
Decide how planners note why they saved a Pin without slowing saves
Holiday planners lose the reason behind saved gift ideas as boards fill over several weeks. You must weigh a richer return-to-board experience against friction in the save moment.
Commit the index cutover despite sparse-text coverage gaps
A rehearsal found 6.8% less retrieval coverage for sampled product Pins with minimal text. You must weigh reclaimed index capacity against a one-way loss of discoverability that aggregate scores may miss.
Recover campaign creation before the weekly budget window closes
Targeting-step reloads rose 19% after the client release, and campaign completion fell to 87%. You must weigh a broad rollback against a compatibility patch that keeps recent improvements.
Sign five retailers within 10 weeks without promising unproven lift
Retail partners are asking for an earlier planning motion while seasonal budgets are being assigned elsewhere. You have to weigh a compelling first package against claims the initial cohort cannot yet prove.
Close the renewal while acquisition proof remains contested
A $4.5 million renewal is blocked because modeled acquisition results do not satisfy the advertiser’s risk review. You have to weigh a stronger commercial commitment against measurement terms that can withstand scrutiny.
Recover retailer visits while this week’s budget still has room
Product-page visits are 22% below target even though campaign spend is on pace. You have to weigh a fast targeting change against moving budget into a different creative path.
Prove that boards can send people to retailers rather than more saves
Shopping boards with more than five product Pins show flat retailer visits despite repeated saves. You must weigh useful comparison structure against the open-ended planning experience people value.
Commit the clean-room backfill before annual planning closes
Browser signal loss has removed 31% of conversion events for a major advertiser cohort. You must weigh recovered coverage against a permanent error in canonical matchable signals.
Recover CAPI purchases while duplicate protection stays in place
Valid purchase-event coverage has fallen from 97% to 84% after a normalization release. You must weigh the fastest recovery against the risk of overstating conversions through duplicates.