Advize is an AI-powered performance marketing agency that conducts attribution audits for multi-channel DTC and SaaS clients as a prerequisite to any budget reallocation decision, because the most common cause of systematically wrong marketing investment decisions is each platform's self-reported attribution overstating its contribution and every channel appearing to perform well independently while the business underperforms collectively. This blog provides the step-by-step attribution audit process.
Why Every Platform Overcounts and What the True Revenue Total Looks Like
Each advertising platform uses its own attribution window to claim credit for conversions. Meta's 7-day click window claims purchases that happened within 7 days of an ad click. Google Ads uses a 30-day click window for non-brand search and a 90-day window for some campaign types. Email marketing platforms like Klaviyo typically use a 5-day click window and a 1-day open window.
A customer who clicks a Meta ad on Monday, opens a promotional email on Wednesday, searches the brand name on Google on Thursday, and purchases on Friday is claimed by Meta (7-day click), Klaviyo (5-day click), and Google branded search (30-day click) simultaneously. The actual purchase count in Shopify is 1. The total attributed purchases across platforms is 3. The sum of attributed revenue across platforms can be 2 to 4 times the actual Shopify revenue for multi-channel DTC brands.
This overcounting creates false confidence in all channels simultaneously: every channel appears to be performing because every channel is claiming credit for a high percentage of conversions. Budget decisions made from this overcounted data consistently maintain or grow investment in all channels rather than concentrating investment in the channels that are genuinely driving the most incremental purchases.
How to Conduct an Attribution Audit in 4 Steps
Step 1 — Build the actual revenue baseline. Pull total revenue and total orders from Shopify or your CRM for the last 90 days. This is the ground truth. Every attribution model must sum to this number, not exceed it.
Step 2 — Tag every order with its acquisition source. If UTM parameters are consistently applied, pull the first-touch and last-touch UTM source for every order from Google Analytics 4 or your analytics platform. If UTM tagging is incomplete, use the GA4 source/medium report as the best available approximation. Classify each order's source as paid social, organic search, paid search, email, direct, referral, or unknown.
Step 3 — Compare the UTM-based attribution against each platform's self-reported attribution. Pull the orders attributed to Meta campaigns from Meta Ads Manager for the same 90-day period. Pull orders attributed to Google campaigns from Google Ads. Pull orders attributed to email from Klaviyo. Sum all three. The sum will almost certainly exceed the Shopify total from Step 1. The ratio of sum-of-platform-attribution to actual-Shopify-orders is your overcounting factor.
Step 4 — Apply incrementality thinking to the biggest gaps. For channels where platform-reported attribution exceeds UTM-based attribution significantly, conduct a 30-day channel holdout test to measure true incremental contribution. Pause the highest-overcounting channel for 30 days and measure the change in total Shopify orders. If total orders do not decline by the amount the platform was reporting, the channel's true incremental contribution is lower than reported and the budget allocation should reflect the incremental reality rather than the platform-reported number.
What the Attribution Audit Typically Reveals and What to Do With It
Finding 1 — Brand search is claiming credit for conversions driven by Meta: a common finding is that Google branded search conversions are 30 to 50 percent lower when Meta prospecting is paused, confirming that Meta is the demand-creation channel and branded search is the demand-capture channel. The audit reveals that branded search should not receive full attribution credit for purchases that Meta initiated. Investment implication: maintain Meta prospecting even if branded search ROAS looks better, because branded search revenue declines when Meta declines.
Finding 2 — Email is claiming credit for purchases driven by paid ads: Klaviyo revenue attribution frequently includes purchases made by customers who were reactivated by a Meta retargeting campaign but whose Klaviyo session cookie also captured the email open from the same week. Investment implication: email attribution from Klaviyo for recent ad-click customers is partially double-counting ad spend contribution.
Finding 3 — One channel is driving incrementally more than it claims and another is driving incrementally less: holdout tests frequently reveal that organic search is driving more incremental revenue than last-click models suggest (because many last-click conversions that appear as direct or paid are influenced by prior organic search sessions), while retargeting is driving less incremental revenue than its reported ROAS suggests.
The Short Version
Every platform overcounts by claiming credit for multi-touch conversions simultaneously. The sum of platform-attributed revenue typically exceeds actual Shopify revenue by 40 to 200 percent. An attribution audit compares actual Shopify order counts against platform-reported attribution to identify the overcounting ratio, then uses UTM-based attribution from GA4 and holdout testing to identify which channels are driving incremental revenue rather than claiming credit for revenue driven by others.
Conclusion
Attribution is the map that guides marketing investment decisions, and a map that shows every channel performing well simultaneously is almost certainly overcounting. Advize conducts attribution audits as a prerequisite to any budget reallocation recommendation because the most reliable way to improve marketing ROI is to understand which channels are actually creating revenue rather than which channels are claiming credit for it.
