What Is an Attribution Window in Meta Ads? (2026 Explainer)

TL;DR: An attribution window is the time period after an ad interaction during which Meta credits a conversion to that ad. The 2026 default is 7-day click + 1-day view. Post-iOS 14, longer windows contain more modeled conversions — the number you see in Ads Manager is a statistical estimate, not a pixel-confirmed count. Choosing the right window changes both reported ROAS and how the algorithm optimizes your campaigns.

If you have ever looked at your Meta Ads dashboard and wondered why conversions there don't match what Shopify or your CRM shows, attribution windows are the most likely culprit. Not a broken pixel. Not a CAPI misconfiguration. The window setting itself.

This guide explains what attribution windows are, how Meta's options work in practice, what iOS 14 and Aggregated Event Measurement changed, and how to pick a window that reflects your actual business. For the full mechanics of how attribution window settings have evolved since 2021, there's a dedicated deep-dive. This piece is the conceptual foundation.

What Is an Attribution Window?

An attribution window is a defined time interval. The rule: if a person interacts with your ad and then converts within that window, Meta records the conversion as caused by the ad.

The word "interact" covers two distinct trigger types:

• Click attribution: The person clicked the ad. The window starts at the moment of click.
• View attribution: The person saw the ad for at least one second but did not click. The window starts at the moment of the impression.

Meta lets you set an independent duration for each trigger. Available options:

The 2026 default — and the standard setting you'll see on most new campaigns — is 7-day click + 1-day view. That's a change from the pre-2021 default of 28-day click + 1-day view, which Meta retired after iOS 14 reduced signal data. Per Meta's official attribution documentation, the 7-day click window replaced 28-day click as the platform default in January 2021.

The practical effect: if someone clicks your ad on Monday and buys the following Sunday, that conversion counts. If they click and buy 8 days later, it does not.

For a quick primer on how Meta's Ad Library surfaces which settings competitors tend to run, the AdLibrary ad detail view shows campaign-level metadata that can hint at measurement philosophy. You can also use the ROAS calculator to model how different attributed conversion volumes change your efficiency numbers before making window decisions.

Why the Window Setting Matters More Than It Looks

On the surface, choosing between 1-day and 7-day sounds like a minor reporting preference. It isn't. The window setting shapes three things simultaneously.

1. What the algorithm optimizes toward. Meta's delivery system optimizes campaigns for the conversion event defined at the ad set level — within the attribution window you've chosen. A 7-day click window means the algorithm hunts for people likely to convert within 7 days of clicking. A 1-day click window means it hunts for people likely to convert today. These are different populations with different behavioral profiles.

2. How many conversions the learning phase sees. The learning phase requires roughly 50 optimization events in a rolling 7-day period. A shorter attribution window means fewer events qualify per week, which makes exiting learning harder and keeps CPAs volatile longer. This is a structural cost of conservative measurement that most advertisers don't account for when switching windows.

3. The ROAS number in your dashboard. A 7-day window will almost always report a higher ROAS than a 1-day window for the same campaigns. Neither is wrong — they measure different things. The gap tells you how much of your attributed revenue comes from buyers who took 2-7 days to convert after clicking. For blended ROAS calculations that combine Meta data with platform-agnostic revenue numbers, the attribution window is the first thing to align before drawing conclusions.

Click vs. View Attribution: The Real Trade-Off

View-through attribution is the most debated setting in Meta Ads measurement. Here's the honest read.

The case for including 1-day view: Display and video ads generate real awareness that influences purchase decisions. Someone who watches 80% of your video ad and buys the next morning is plausibly influenced by that ad, even without a click. Suppressing view attribution entirely undercounts the contribution of upper-funnel creative — especially for Reels campaigns, video in Advantage+ placements, or awareness objectives.

The case against view attribution: Meta's ad delivery system reaches tens of millions of people. Many will buy something you sell within 24 hours of seeing your ad purely by coincidence — especially if you're selling a commodity with broad demographic appeal. The view attribution window captures this coincidental overlap as "conversions," which inflates your reported numbers and sends false optimization signals to the algorithm.

The empirical test: run a holdout test where a control group is shown a PSA instead of your ad. Compare their conversion rate to the exposed group's view-attributed rate. If the lift is small, view attribution is overclaiming. In most direct-response campaigns, it is overclaiming by a meaningful margin. The IAB's cross-media measurement guidelines specifically flag view-through attribution as requiring holdout validation before being used in optimization decisions.

The practical recommendation for 2026: Use 7-day click + 1-day view as the default. Disable view attribution (set to 0-day view) for any bottom-funnel direct-response campaign where you need conservative numbers. Enable it for awareness campaigns where influencing without clicking is the actual mechanism. For incrementality measurement to be meaningful, your attribution window must be as tight as your conversion cycle allows.

How iOS 14 and ATT Changed What Attribution Windows Mean

Before April 2021, Meta's pixel tracked conversions on iOS devices via browser-level cookies and IDFA device identifiers. Attribution was mostly direct: Meta saw the click, saw the conversion event fire, matched them, credited the ad.

The iOS 14 ATT framework changed the permission model. Users had to explicitly opt in to cross-app tracking. Opt-in rates landed in the 20-35% range for most categories. The majority of iOS users became unmeasurable by conventional pixel means.

Meta's response was Aggregated Event Measurement (AEM). Under AEM:

• Each domain is limited to 8 prioritized conversion events.
• Events are reported with a delay (up to 72 hours for privacy thresholds).
• Conversions below statistical confidence thresholds are modeled — Meta estimates them using machine learning on comparable cohorts rather than direct observation.

The attribution window still applies, but a growing fraction of what fills it is modeled rather than pixel-confirmed. For a 7-day click window on an iOS-heavy audience, you might be looking at 40-60% modeled conversions in the reported total, depending on your industry and event quality. Apple's App Tracking Transparency framework documentation details the technical constraints that created this measurement gap.

This is why Meta Conversions API (CAPI) matters so much in 2026. Server-side events sent through CAPI bypass the browser/app permission model. They improve Event Match Quality (EMQ), which in turn reduces the proportion of modeled conversions in your attribution window. A higher EMQ score means the numbers in your window are more observed and less inferred.

For Meta Pixel setups that maximize signal for attribution, running browser pixel + CAPI in parallel is now the baseline, not an enhancement.

Practical implication: for iOS-heavy audiences (gaming, health, finance, apps), treat your 7-day click numbers as directional estimates with a wide confidence interval. Shorter windows have smaller absolute errors because there's less time for modeled events to accumulate.

Attribution Windows and the Learning Phase

The interaction between attribution windows and Meta's learning phase is under-discussed and frequently responsible for unstable performance after account changes.

The learning phase requires 50 optimization events in 7 days at the ad set level. Only conversions that fall within your attribution window count toward the 50.

Consider two scenarios with identical actual purchase volume:

Scenario A: 7-day click + 1-day view window Weekly reported conversions: 65 (algorithm exits learning, stabilizes CPA)

Scenario B: 1-day click + 0-day view window Weekly reported conversions: 22 (algorithm stays in learning, CPA swings ±40%)

The actual business result is the same. The measurement choice determines whether the algorithm can function. This is why switching to a tighter attribution window often "makes performance worse" in the short term — not because the tighter window is inaccurate, but because the algorithm loses its footing.

The right response: either increase spend to generate more conversions within the tighter window, or move to a higher-funnel event (Add to Cart or Initiate Checkout) that fires more frequently, letting the algorithm exit learning on a proxy event while you evaluate downstream CPA manually.

For campaign budget optimization (CBO) accounts, attribution window standardization across ad sets is critical — ad sets with different windows are effectively playing by different scoring rules. Use the CPA calculator to model whether tightening your window and temporarily raising bids is viable given your margins.

The 2026 Attribution Stack: Where Windows Fit

Attribution windows are one layer in a broader measurement stack. Here's how they relate to the other tools.

Meta Pixel + CAPI: The data sources that feed the window. Higher Event Match Quality reduces modeled events and makes the window more accurate. No window setting compensates for bad signal.

Marketing Mix Modeling (MMM): Platform-agnostic, privacy-safe. MMM doesn't use attribution windows at all — it infers channel contribution from spend and revenue time series. Use MMM to cross-check whether Meta's window-attributed ROAS aligns with what econometric regression says Meta's contribution actually is. The Marketing Science Institute's research on attribution model validity shows MMM-vs-MTA gaps widening post-iOS 14.

Holdout Testing: The only way to measure incrementality directly. If 7-day attributed conversions are 100 but holdout tests show only 40 incremental conversions, your window is overclaiming by 2.5×. This gap is your "phantom ROAS."

Post-Purchase Surveys: Ask buyers "How did you hear about us?" If Meta is your largest channel but only 15% of survey respondents name it, your attribution window is amplifying a smaller actual contribution.

SKAdNetwork (SKAN): iOS-native attribution framework for app campaigns. SKAN operates independently of Meta's attribution windows and has its own postback timing rules. For app advertisers, SKAN data is often more reliable than Meta's modeled web-conversion numbers for iOS traffic.

For a comprehensive view of how these pieces fit together, the death of attribution post covers the full evolution and what operators should use in combination.

Choosing Your Attribution Window: A Decision Framework

Here's a practical decision tree for 2026.

Step 1: Identify your conversion cycle. Median time from first click to purchase for your product. Track this in your CRM or Shopify analytics. If median is less than 24 hours, a 1-day click window captures the vast majority of your real conversions. If median is 3-5 days (considered purchases, B2B leads, subscription trials), a 7-day window is appropriate.

Step 2: Assess your iOS exposure. If more than 50% of your audience is on iOS (check the Device breakdown in Meta reporting), your pixel-based attribution has significant gaps. CAPI becomes non-negotiable. Shorter windows reduce accumulated modeling error.

Step 3: Determine your optimization event frequency. Can you get 50 qualifying events per week per ad set with your chosen window? If not, either expand the funnel event or widen the window temporarily until spend is high enough to sustain tighter measurement. The learning phase calculator helps you estimate required spend for a given event frequency and window combination.

Step 4: Run a window comparison. In the same ad account, run identical creative with different attribution settings across separate campaigns. Compare reported ROAS, CPA, and CPM. The spread reveals how much your current window is inflating or deflating numbers.

Step 5: Cross-check with a holdout. Even a 90-day holdout test on 5% of your audience tells you how much of the window-attributed conversion volume is incremental. Build this into your quarterly measurement cadence.

For advertisers running broad targeting with Advantage+ or Andromeda's signal-based delivery, the holdout is especially important — the algorithm's reach expansion can cause it to convert people who would have bought organically, amplifying the phantom attribution problem.

What Competitor Ads Tell You About Attribution Strategy

When a competitor runs the same creative continuously for 90+ days at scale, they're not doing it by accident. They have measurement infrastructure that shows that ad is profitable — and part of that infrastructure is their attribution window choice. Operators who run long-duration ads confidently tend to have conservative measurement settings, because they need accurate data to justify continued spend.

AdLibrary's ad timeline analysis lets you see how long competitor ads have been running and across which platforms. A brand that consistently runs 60-90-day campaigns has a measurement stack they trust. A brand that rotates every two weeks is either testing aggressively or chasing a noisy signal — often from an overly wide attribution window that makes every new creative look like a winner in week one.

The multi-platform coverage matters too: if a competitor runs coordinated campaigns on Meta, TikTok, and YouTube simultaneously, their attribution problem is more complex than yours. They're almost certainly running MMM in parallel because cross-platform attribution with platform-native windows double-counts aggressively.

For competitive research workflows that incorporate measurement signals alongside creative analysis, AdLibrary's platform filters and geo filters let you narrow the universe to the exact market and channel mix you're evaluating.

Meta's free Ad Library API gives you basic creative data for one platform. The moment your research spans TikTok, YouTube, and LinkedIn in the same query — and you need the timeline depth and signal enrichment to draw conclusions about competitor measurement strategy — you need something built for the job. The AdLibrary API (Business tier) covers all eight major ad platforms in a single authenticated endpoint, with richer metadata per ad than Meta's public data returns. See the API access feature if you're running programmatic competitor research at scale.

Frequently Asked Questions

Q: What is an attribution window in Meta Ads?

An attribution window in Meta Ads is the time period after someone interacts with your ad during which a subsequent conversion is credited to that ad. Meta supports click windows (1-day or 7-day) and view windows (1-day). The default since 2021 is 7-day click, 1-day view.

Q: What is the difference between 1-day and 7-day click attribution in Meta?

A 1-day click window credits conversions that happen within 24 hours of clicking your ad. A 7-day click window extends that to a full week. The 7-day window reports more conversions and is better for products with longer consideration cycles; the 1-day window is more conservative and aligns closer to actual ad-driven demand.

Q: How did iOS 14 change attribution windows on Meta?

iOS 14 ATT cut the pixel-based data Meta could collect from Safari and iOS apps. For iOS web events, Meta shifted to Aggregated Event Measurement (AEM), which imposes an 8-event priority cap and models many conversions using statistical inference rather than actual signals. This made longer windows less reliable because more of the attributed conversions are modeled estimates rather than observed events.

Q: Should I use click or view-through attribution in Meta Ads?

Use click-only attribution (7-day click, 0-day view) if you want the most conservative measurement that's hardest to inflate. Enable 1-day view attribution if you run awareness-heavy campaigns or upper-funnel video where view-driven conversions are plausible. Avoid 1-day view for direct-response bottom-funnel campaigns where it typically overclaims credit.

Q: Does changing the attribution window affect campaign delivery in Meta?

Yes. Meta's algorithm uses the attribution window setting to define the conversion event it optimizes toward during the learning phase. A shorter window means the algorithm sees fewer qualifying conversions per week, which can extend or prevent completion of the learning phase (typically requiring 50 optimization events in 7 days). Changing the window mid-flight resets learning.

Common Attribution Window Mistakes That Inflate ROAS

These are the most expensive measurement errors operators make with attribution window settings.

Mistake 1: Accepting the default without auditing your conversion cycle. Meta defaults to 7-day click + 1-day view. For impulse-purchase categories (beauty, apparel under €50, digital subscriptions), the 7-day window captures a meaningful amount of organic conversions. Running a 1-day click test for 4 weeks often reveals that 20-40% of 7-day-attributed conversions disappear — not because conversions declined, but because they were never caused by the ad.

Mistake 2: Changing the window mid-campaign. Switching attribution settings on a live campaign forces Meta's algorithm back into the learning phase. CPA typically spikes for 7-14 days. Always apply window changes to new campaigns or at the start of a new budget cycle.

Mistake 3: Comparing campaigns with different windows. An agency runs a retargeting campaign on 1-day click and a prospecting campaign on 7-day click, then reports blended ROAS. The numbers are meaningless. Window standardization is a precondition for any cross-campaign analysis. For retargeting vs prospecting comparisons specifically, mismatched windows typically make retargeting look 3-5× more efficient than it actually is.

Mistake 4: Using view attribution on broad prospecting. If you're running Advantage+ Shopping or broad targeting campaigns at scale, 1-day view attribution on a large audience means Meta is counting thousands of "converted viewers" who bought from you because they were already in-market. For advanced retargeting segmentation, this mistake compounds: the algorithm feeds warm audiences into broad prospecting, view attribution inflates the numbers, and budget shifts toward the wrong placements.

Mistake 5: Not running a holdout to calibrate the window. Without a holdout test, you have no external reference point. A 7-day attributed ROAS of 4.2 might represent an incremental ROAS of 1.8 — or 3.9. You cannot know which without a holdout. Build holdout testing into your performance marketing cadence.

Attribution Windows and Multi-Platform Measurement

Meta's attribution windows exist in isolation from every other ad platform. Google Ads defaults to 30-day click + 1-day view for Search. TikTok defaults to 7-day click + 1-day view, same as Meta's current default. Each platform claims credit independently.

The result: if a buyer clicks a Meta ad on day 1, clicks a Google Shopping ad on day 3, and purchases on day 4, both platforms record a full conversion. Your total reported conversions across platforms can easily be 40-80% higher than your actual order count. This is the double-counting problem that MMM was designed to solve.

For paid social strategies that run Meta alongside Google, TikTok, and YouTube simultaneously, a unified attribution approach is required:

• Revenue-based MMM: Regression modeling that allocates revenue to channels based on spend patterns, bypassing platform-native windows entirely.
• First-touch or last-touch UTM tracking: Simple, transparent, but misses assisted conversions.
• Third-party MTA platforms: Tools like Northbeam, Triple Whale, or Rockerbox attempt to deduplicate using probabilistic matching.
• Post-purchase surveys: Direct customer self-reporting as a sanity check on platform numbers.

For ecommerce ad tracking software that attempts to unify cross-platform attribution, the window settings you choose on each native platform feed directly into how those tools model contribution. The ad budget planner can help you model spend scenarios where you're deliberately allocating budget to account for known attribution inflation across platforms.

What to Do Next

If you've read this far and aren't sure where to start:

1. Audit your current window settings across every active ad set. Look for inconsistencies across campaigns.
2. Check your iOS traffic share. If above 50%, prioritize CAPI setup before tuning windows.
3. Run a 1-day vs 7-day comparison on your top-performing campaign over 4 weeks. Calculate the gap in attributed conversions and estimate how much is genuine consideration-cycle buyers vs coincidental converters.
4. Set up a holdout group on your highest-spend campaign. Even 5% audience holdout gives you calibration data in 60 days.
5. Study competitor longevity. If your top competitors run 60-day-plus campaigns consistently, they have measurement confidence. Use AdLibrary's timeline analysis to benchmark their creative staying power — which is a proxy for their measurement discipline.

For teams managing multiple clients or running competitive research at scale, the AdLibrary Pro plan gives you the tools to benchmark competitor measurement maturity alongside creative analysis.

Attribution windows don't tell you whether your ads are working. Holdout tests tell you that. What windows tell you is how many conversions to show the algorithm so it can find the next buyer. Get that right, and the algorithm does its job. Get it wrong, and you're optimizing toward a mirage.