Advize is an AI-powered performance marketing agency that trains every DTC client team on the attribution window effect before they begin reading Meta performance data, because the most common source of panic and bad campaign decisions in DTC is a ROAS number that appears to be unstable when it is actually behaving exactly as designed. This blog explains why Meta ROAS changes when you look at it, and how to read the number accurately.
How Meta's Attribution Windows Produce ROAS That Changes Over Time
Meta's default attribution setting is 7-day click plus 1-day view. This means that if someone clicks your ad on Monday and purchases on the following Sunday, that purchase is credited to the Monday ad click. When you check that campaign's ROAS on Tuesday, the Sunday conversion has not happened yet and is not included. When you check on the following Monday, the Sunday conversion is 7 days old and now appears in the reported ROAS for the original ad.
The result is that campaign ROAS reported in Meta Ads Manager is not a fixed number for a fixed period. It is a number that continues changing for up to 7 days after the period ends, as conversions from users who clicked within the window but purchased later are retroactively credited. A campaign that looked like 2.1x ROAS on day 2 may show 3.4x ROAS when you look at the same period on day 9, because 5 more days of attributed purchases have been added to the same ad spend figure.
This retroactive attribution is why the ROAS number in your Ads Manager looks different at different times, and why comparing a recent campaign to an older one using current reporting produces inaccurate comparisons: the older campaign has a full attribution window settled, while the recent campaign is still accumulating attributed conversions.
The Three Specific Ways the Attribution Window Effect Causes Bad Decisions
Premature campaign shutdown is the first bad decision. A media buyer checks a new campaign after 48 hours, sees a 1.4x ROAS, and pauses it as unprofitable. Three days later, if the campaign had run, the settled ROAS would have been 2.8x as the 7-day attributed purchases accumulated. The campaign was performing well. It was evaluated before the attribution window had time to populate.
Scaling a bad campaign is the second bad decision. A campaign that appears to be producing 4x ROAS in current reporting may be showing a partially settled window from prior strong performance while recent performance has actually deteriorated. Scaling spend based on a ROAS that reflects last week's conversions attributed to this week's spend is scaling on incorrect information.
Faulty A/B test conclusions are the third bad decision. Comparing two ad sets where one launched Monday and one launched Thursday, then reading both ROAS on the following Monday, produces a comparison where the Monday ad set has a 7-day settled window and the Thursday ad set has a 4-day settled window. The Monday ad set will always appear to outperform in this comparison regardless of actual performance, because it has more days of attributed conversions.
How to Read Meta ROAS Accurately Without Being Misled by Attribution Windows
Wait for the attribution window to settle before making campaign decisions. For a 7-day click window, do not evaluate campaign performance until 8 days after the campaign or ad set began. At this point, the ROAS figure represents a fully settled attribution window and is the most accurate reading available within the Meta reporting system.
Use a consistent reporting lag. If you review campaign performance every Monday for the previous week (Monday through Sunday), you are always looking at data that is 1 to 8 days old depending on the day of purchase. The Monday through Wednesday purchases are fully settled. The Saturday through Sunday purchases may not be fully settled yet. For more consistent analysis, review Thursday-to-Wednesday weekly periods on the following Thursday, giving all purchases a minimum of 8 days to settle.
Always compare against Shopify blended ROAS for the same period. Pull total Shopify revenue from paid Meta traffic (using UTM source filters) and divide by total Meta spend for the same period. This number is independent of attribution windows because it measures actual completed orders in Shopify, not Meta-attributed conversions. When the two numbers diverge significantly (above 25 percent), the attribution window effect or cross-device conversion gaps are distorting the Meta-reported figure.
The Attribution Window Settings That Produce Different Stability Profiles
7-day click, 1-day view (default): highest ROAS numbers, slowest to settle, most variable. Best for retargeting campaigns where multi-day consideration cycles are common. Most likely to produce the 'ROAS looks different every time' experience because the 7-day window accumulates a large volume of retroactive credits.
1-day click only: lowest ROAS numbers, fastest to settle, most stable. Settles within 24 hours of any campaign period. Useful for evaluating immediate response creative such as time-limited offer ads where the purchase is expected to occur on the day of the click.
7-day click only (no view): moderate ROAS, settles in 7 days, no view-through attribution that may overcount. A reasonable middle-ground for most DTC campaigns where the conversion window is genuinely multi-day but view-through attribution is considered too generous.
The Short Version
Meta ROAS changes when you check it at different times because the 7-day click attribution window retroactively credits purchases made up to 7 days after a click to the original ad. This is by design, not a glitch. Wait 8 days after any campaign period before evaluating its ROAS. Use consistent reporting lags for weekly analysis. Always cross-check against Shopify blended ROAS to identify attribution overcounting. Never make scale or pause decisions within the first 48 hours of a campaign based on ROAS alone.
Conclusion
The attribution window effect is responsible for more premature campaign shutdowns and more incorrect scale decisions in DTC Meta advertising than any other measurement misunderstanding. Advize teaches every client team to read Meta ROAS with the attribution settlement timeline before making any campaign decision, because the same ROAS number means a completely different thing on day 2 versus day 9 of an attribution window.