Meta Attribution Settings Explained: Click, Engage-Through, View and Incremental Attribution

What are Meta attribution settings?

Meta attribution settings are the rules that decide which ad interaction gets credit for a conversion and how many days that credit window stays open. You choose separate windows for link clicks, engagements and views; Meta then uses the same rules to report results and to pick the signals it optimises delivery for.

In short, this setting defines what the "Results" column in Ads Manager actually means. I have worked with ad accounts since 2012, and a large share of the reporting disputes I see come down to this one setting. For example, when a founder says "Meta shows 40 purchases, my store shows 22 orders", the two systems usually count different things.

In this guide I cover the attribution windows, the recent change to the click definition, the difference between standard and incremental attribution, and why GA4 rarely agrees with Ads Manager. I rely on official Meta and Google documentation. My goal is not to hand you one "true" number; instead, I want to show which number answers which question.

Keep one idea in mind as you read: attribution is not an accounting entry but an estimation rule. Your store tells you how many orders you took. Meta's report, by contrast, flags which of those orders a Meta ad may have influenced, according to its own rules. Treating both as answers to the same question causes most of the confusion.

What do click, engage and view windows count?

When Meta connects a conversion to an ad, it looks for three types of touchpoint. Each type also has its own time limit, which Meta calls a window. A sale that happens after the window closes does not count for that ad, even though the customer really bought and you can see the order in your shop system.

• Click-through: someone clicks the link in your ad and converts within the chosen period. The usual options are 1 day and 7 days.
• Engage-through: someone likes, shares or saves the ad, or watches the video for a minimum time, without clicking the link, and then converts within 1 day.
• View-through: someone sees the ad, does not interact at all, and converts within 1 day.

As a result, the same campaign can look very different under a 1-day click and a 7-day click window. Products with a long consideration phase in particular collect more credit in the wider window. On the other hand, for cheap impulse purchases the gap between the two windows tends to stay small.

One more point matters here. If a user touched several of your Meta ads, Meta credits the conversion to a single ad, so the same sale does not appear twice inside your account. However, Google Ads, email or organic search can still claim that very sale in their own reports.

How did Meta change the definition of a click?

In an announcement dated 3 March 2026, Meta said that for campaigns optimising for website and in-store conversions, click-through attribution would now include only link clicks. Before this change, likes, saves and shares could also act like clicks and produce click-through conversions.

Meta states the reason plainly: its reporting should line up better with third-party tools such as Google Analytics. The company also said that billing does not change. In other words, the update affects how Meta reports your spend, not what you pay. You can read the official statement in the Meta for Business announcement.

In practice, once the change reaches your account, click-through conversions may drop while engage-through conversions rise. Even if the total stays flat, the split moves. Therefore, when you compare periods, I recommend splitting the date range into before and after the change. Otherwise you might think performance fell and pause a campaign that still works.

Meta said the rollout would start later in March and reach advertisers gradually. That also explains why two accounts could behave differently in the same week.

What does engage-through attribution cover?

Engage-through attribution is the new name for what Meta used to call engaged-view attribution, and its scope grew. It now groups the social interactions that are not link clicks: shares, saves, likes and similar actions. For video, Meta also lowered the viewing threshold from 10 seconds to 5 seconds.

Meta shared its own data to justify the lower threshold. According to the company, 46 percent of online purchase conversions with Reels happen within the first 2 seconds. That figure comes from Meta itself, not from an independent study; still, it hints at how people consume short video.

So how much should you trust this category? My approach is simple: track engage-through conversions in their own column, but do not make budget decisions from that column alone. A person who watches a video for 5 seconds and searches for your brand the next day may reflect real influence. That same person, however, might have bought without ever seeing the ad. The tests I describe later help you reduce that uncertainty.

Note that the engage-through window has one option only: 1 day. If you want, you can switch it off by selecting "none" in the settings.

What is the default in Meta attribution settings?

For campaigns that optimise for website conversions, the default combination is a 7-day click, a 1-day engage-through and a 1-day view window. You find this at ad set level, in the section for conversion and performance goal. Depending on your account and campaign type, the layout can differ slightly.

In practice, the default makes a reasonable starting point for most businesses. Still, do not accept it without thinking about your own situation. For example, in a retargeting campaign, view-through conversions can inflate results, because this audience already visited your site and sits close to buying. People who would have completed the checkout anyway end up credited to the ad through the view window.

Conversely, switching off view and engage windows for a cold-audience video campaign can hide genuine impact. That is why I suggest you first place different windows side by side in reporting before you change the default. I explain how to do that further down.

Also remember that your chosen window shapes delivery, not just reporting. The system starts to look for people who resemble the signal it learns to treat as a conversion.

What happened to the 7-day and 28-day view windows?

Meta announced on its developer blog that, from 12 January 2026, the Ads Insights API would no longer return the 7-day view and 28-day view windows. The windows that remain in the API are 1-day click, 7-day click, 28-day click, 1-day engaged view and 1-day view. You can check the details in the Meta for Developers post.

The same update also changed data retention. Unique-count fields and hourly breakdowns now go back 13 months, while frequency breakdowns go back 6 months. Total values stay available for up to 37 months, in line with Ads Manager.

This hit businesses that pull reports into tools like Looker Studio through the API in particular. If your dashboard used to read 28-day view data, that field may now come back empty. So if you noticed a sudden drop in your dashboard after New Year, suspect the connector and the window it requests before you blame the campaign.

My advice: for long-term comparisons, rebuild your historical data using today's windows. Showing two periods that rely on different windows in the same chart remains one of the most common reporting mistakes I find.

How does standard attribution differ from incremental attribution?

Standard attribution credits every conversion that follows an eligible touchpoint inside the window you picked. Incremental attribution asks a different question: would this conversion have happened without the ad? Under this model, Meta counts the conversions it predicts the ad caused, and it optimises delivery towards those conversions.

Meta's help centre article on incremental attribution summarises how the model works. My own observation: the incremental model shrinks the reported number, yet that shrinkage often means you move closer to reality. Especially for brands with strong awareness and healthy organic sales, standard attribution tends to overstate what ads contribute.

Which campaigns can use incremental attribution?

Meta rolls this option out gradually, so it does not appear in every account at once. According to the setup that Meta's Help Centre describes, you find it in the performance goal section of the ad set, in Sales or Leads campaigns with a website conversion location. Account eligibility and conversion volume also play a role.

Before you switch it on, ask yourself these questions:

1. Is your weekly conversion volume high enough for the model to learn?
2. Does the campaign aim to win new customers or to bring back existing ones?
3. Are you ready to explain lower reported numbers to your team and management?
4. Do you have a control campaign running on standard attribution in the same period?

I consider the fourth question the most important. If you compare an incremental campaign against an older standard campaign, you compare apples with oranges. Instead, judge both through total store revenue and profit.

Also expect learning to reset once you select incremental attribution, so some volatility in the first days is normal. I explain how to handle that period in my article on the Meta Ads learning phase.

How do attribution choices affect delivery and learning?

Still, many advertisers treat attribution as a pure reporting switch. In reality, Meta also uses the window you choose as its optimisation signal. With a 1-day click window, the system looks for people similar to those who click and buy on the same day. With a 7-day click window, it also accounts for people who need a few days to decide.

This has two consequences. First, a narrow window produces fewer conversion signals, so small accounts can struggle to leave the learning phase. Second, a wide window produces more signals, but some of them may reflect purchases that had nothing to do with the ad. The system can then drift towards people who would buy anyway.

For example, in the ecommerce accounts my team and I manage, we keep the 7-day click window for products with a high basket value, and we judge the view window per campaign role. We stay sceptical about view-through conversions in retargeting; my guide on strengthening the sales funnel with remarketing explains why those audiences inflate numbers.

Finally, I recommend testing a new window in a new ad set instead of editing a live one. That way you can compare both setups cleanly.

Why do Meta attribution settings and GA4 show different numbers?

Put simply, the two systems count the same sale with different logic. Meta recognises people who touched its ads through their logged-in accounts, and it can include views. GA4 only sees sessions that reach your website; it knows nothing about someone who saw an ad and never clicked. That single difference explains a large part of the gap.

In addition, the two tools share credit differently. GA4 sees every channel in a purchase journey, including Google Ads, organic search, email and Meta, and it splits credit according to its own model. Meta, by contrast, only sees its own touchpoints and takes full credit for any conversion inside the window. Consequently, if you add up the sales every ad platform reports, you end up above your real store revenue.

Timing forms the third difference. Meta usually reports a conversion on the day of the ad impression or click, whereas GA4 reports it on the day it happens. So a product someone clicks on 28 September and buys on 2 October lands on different days in the two reports.

Data loss adds a fourth gap: cookie consent, ad blockers and incomplete tagging affect each side to a different degree. I cover the basics of GA4 in my Google Analytics 4 guide.

Which attribution models does GA4 still offer?

Google retired the first click, linear, time decay and position-based models in GA4 in November 2023. Today you can choose data-driven attribution or rules-based last click options. You find the controls in the GA4 Admin area under data display, in the attribution settings. The Google Analytics Help page covers them in detail.

• Data-driven: splits credit by modelling how much each touchpoint contributed. It serves as the default and suits most cases.
• Paid and organic last click: ignores direct traffic and gives all credit to the last channel.
• Google paid channels last click: gives credit only to Google ad clicks.

The lookback window matters just as much. For acquisition key events the default is 30 days, and you can reduce it to 7. For all other key events the default is 90 days, and you can pick 30 or 60 instead. That span runs far longer than Meta's 7-day click window, so factor it into any comparison.

One more detail: a model change in GA4 applies to historical data too, but a lookback change only affects data going forward. Note the date whenever you change it. Otherwise you will waste time months later hunting for the cause of a break in your reports. Also, if your property links to Google Ads, conversions you import from GA4 follow this setting as well.

How should you read Meta and GA4 numbers side by side?

Instead of making the two systems compete, give each one a job. The Meta report works well for choosing between campaigns and creatives. GA4 suits the balance between channels and on-site behaviour better. The final referee, though, is the order and margin data in your shop system.

Here is a practical rule. To see your Meta traffic correctly in GA4, use consistent UTM parameters on every ad link; our UTM builder makes that quick. Then watch the ratio between the two numbers for a few weeks. If the ratio stays stable, all is well; if it suddenly breaks, something in your tracking changed.

Which Meta attribution settings suit which business?

No single window fits everyone, but each sales cycle has sensible starting points. The list below follows the order I use when I think through an account:

1. Low-priced impulse products: compare 1-day click with 7-day click in reporting. If the gap stays small, the short window gives a cleaner signal.
2. Products with a long decision time: for furniture, education or high basket values, a 7-day click window sits closer to reality.
3. Service businesses that collect leads: forms often arrive quickly, but what matters is whether a lead turns into a sale. So track lead quality in your CRM as well.
4. Strong brands and retargeting-heavy accounts: treat view-through numbers with caution and consider testing incremental attribution.

While you decide, also check whether the ads really make a profit. With the ROAS calculator you can enter the revenue each window reports and compare it against your break-even point. That way you see in concrete terms which window pushes you towards false optimism.

How can you compare windows before changing anything?

Ads Manager lets you view several windows side by side in reporting without touching the campaign setting. When you customise columns, use the option to compare attribution settings and open, for example, 1-day click, 7-day click and 1-day view results as separate columns. Menu labels change with interface updates, so look for this option in the column settings.

In this comparison I focus on three questions:

• What share of conversions arrives on day one, and what share on later days?
• How much of the total comes from view-through conversions?
• Do these ratios differ clearly between campaigns?

For instance, if most conversions in a campaign come from the view window, I do not raise its budget before a separate test confirms its real contribution. Conversely, if most conversions come from first-day clicks, I assume the campaign captures direct demand.

Next, you can run the same analysis at creative level. Some creatives trigger instant clicks, while others work after someone watches them. My article on Meta Ads creative testing explains how to set up those tests.

How does data loss distort Meta attribution settings?

However carefully you choose your attribution setting, no model can credit a conversion that never reaches Meta. So check your tracking before any attribution debate. The browser-side Meta Pixel can miss some events because of cookie restrictions, ad blockers and in-app browsers.

To close that gap, Meta offers the Conversions API, a server-side connection. When the Pixel and the Conversions API run together, you need deduplication with an event ID so that the same event does not count twice. Sending hashed customer data such as email and phone also improves match quality. I cover the setup basics in my Meta Pixel guide.

The effect of data loss on attribution feels subtle. It does not spread evenly across campaigns because mobile and in-app browser traffic usually suffers more. Consequently, a campaign that looks weak in reporting might simply fall victim to a tracking gap.

I use a simple check: I compare one week of store orders with the purchase events Meta reports. The ratio gives a rough but useful indicator of how healthy your tracking is.

What mistakes do advertisers make with attribution reports?

In account audits I meet the same mistakes again and again. Most have little to do with technology; instead, they come from reporting habits.

• Treating platform totals as real sales: adding up Meta, Google Ads and email reports and comparing them with revenue always overshoots.
• Mixing windows: showing last year's 28-day view data next to today's 1-day view data in one chart.
• Missing the definition change: blaming a campaign for falling click-through conversions without noticing that the click definition narrowed.
• Deciding too early: judging a campaign on a 7-day window after two days, before part of its conversions arrives.
• Ignoring currency and value events: if the purchase event sends the wrong currency or includes shipping, ROAS misleads you no matter how accurate attribution looks.
• Overrating retargeting: reading retargeting results inflated by view-through conversions as your best campaign.

The fix sounds simple: write the window and model on top of every report. When everyone on the team uses the same definition, the discussion moves from numbers to decisions.

Also watch agency reports closely. If a report only shows the most flattering window, that might reflect a deliberate choice. You have every right to ask which window, model and date range it uses.

How can you validate incrementality with your own tests?

No attribution model replaces a well-designed experiment. Meta's own experiment tools include conversion lift studies, which hold back a control group that does not see the ads and measure the real effect in eligible accounts. If you lack access to that tool, simpler methods still give you useful evidence.

• Geo test: pause ads in one of two similar regions, keep them running in the other, and compare total sales.
• Time test: pause a specific campaign for a short, controlled period and track the change in store sales.
• Budget steps: raise the budget in planned steps and note the change in total sales at each step.

Pay attention to statistical significance in these tests, because small differences are often noise. Our A/B test calculator helps you check whether a gap between two groups holds up.

Also match test length to your sales cycle. A one-week test misleads you for a product that customers take two weeks to decide on. For awareness campaigns, brand lift studies make a better yardstick than sales.

How do my team and I handle Meta attribution settings?

When we take over an account, we start with tracking health, not with the attribution setting. We check Pixel and Conversions API events, deduplication and the ratio to store orders. Then we define each campaign's role: new customer acquisition, retargeting or awareness.

Once the roles feel clear, we choose windows accordingly and open comparison columns in reporting. In management reports we state clearly which window and model produced each number. As a result, the "what does Meta say versus what does the shop say" debate turns into a better question: which campaign contributes to profit?

In eligible accounts we test incremental attribution in a controlled way and judge the outcome by total sales and margin, not by the Meta report alone. If you want support that also covers tracking for your social ads, take a look at our social media management service; for store data and margin structure, see our ecommerce consulting page.

My closing thought: attribution is not a contest about accuracy but a decision tool. Once you know which question you ask, you also know which number to look at.