Why the default windows exist
Meta's default 7-day click, 1-day view window was not designed by engineers optimizing for advertiser accuracy. It was designed by a growth team optimizing for reported ROAS. A wider window captures more conversions within the attribution period. More conversions means higher reported ROAS. Higher reported ROAS means advertisers spend more.
Google's default last-click attribution has the same structural incentive. By assigning 100 percent of credit to the last click, Google search gets full credit for purchases that were influenced by a Meta video ad two weeks prior. Every platform's default setting maximizes its own reported contribution. This is not a conspiracy. It is a structural misalignment of incentives.
How to choose the right window for your product
The right attribution window mirrors your actual customer purchase cycle. A brand selling a 30-euro skincare serum with repeat purchase behavior has a different cycle than a brand selling a 800-euro piece of furniture. The serum brand might see 70 percent of Meta-influenced purchases happen within 3 days of the first ad interaction. The furniture brand might see that same figure take 21 days.
You can measure this empirically using your own Shopify order data. Look at the time between first touch and purchase for customers who came from paid channels. The distribution of those intervals tells you where your natural attribution window should sit.
View-through attribution: the most abused setting
View-through attribution gives a channel credit for a conversion that happened after someone saw an ad but did not click it. Meta's default includes a 1-day view window. This means that if someone scrolls past your Meta video ad at 9am and then buys your product through a direct search at 5pm, Meta claims the conversion.
For awareness campaigns running video at scale, some view-through credit is legitimate. A customer who watches 30 seconds of your product video and buys three hours later was probably influenced by that video. But a customer who saw a 2-second scroll-past of a static image in their feed and converted a day later via a Google search is almost certainly not attributable to that impression.
The practical rule: enable view-through attribution only for video campaigns where you can verify completion rates above 25 percent. Disable it for static image campaigns. Check this setting per campaign type, not account-wide.
Cross-channel attribution windows
The window problem compounds when you run multiple channels simultaneously. If Meta claims 7-day click, Google claims last-click, and TikTok claims 7-day click plus 1-day view, a single customer who touches all three channels before purchasing generates three separate claimed conversions. Your total platform-reported revenue for that customer is three times your actual revenue.
The solution is to set shorter windows on channels where your data shows shorter purchase cycles, and to use your server-side event data to apply a single consistent attribution logic across all channels. This is not possible inside any individual platform's reporting. It requires a unified data layer.
What to actually report internally
Report three numbers for each channel internally: platform-reported (what the platform claims, useful for optimization inputs), Shopify-matched (orders where the customer came from that channel based on UTM and first-party data), and incremental (estimated from holdout testing). These three numbers will rarely agree. The gap between them is the signal worth understanding.
Reporting only platform numbers is like a sales team reporting pipeline in CRM without ever checking which deals actually closed. The number exists but it is not the truth.
---