Reconciling a 38% Conversion Gap Before Scaling Retail Spend
Fabrikam Home spends $45,000 per month on Microsoft Advertising e-commerce campaigns with a 600% target ROAS. The platform reports 1,240 conversions and $312,000 revenue for the last 30 days. Shopify reports 1,980 orders and $356,000 revenue for the same period. The UET conversion tag fires on /checkout/confirmation, but a second tag also fires on /cart. The conversion window is 30 days and view-through conversions are enabled at 1 day. Enhanced conversions are not enabled. A call vendor's offline conversions are imported weekly by CSV: 180 calls last month, but 30% have no Microsoft click ID. Bidding uses Maximize Conversions without a target CPA. Platform attribution is last-click; the analytics team uses data-driven attribution and reports 1,510 conversions. The CMO wants to raise budget 40% next month.
1. Which conversion count should the team use for bidding and executive reporting, and why? Show how you would reconcile the three numbers.
Read the worked answer
Use the platform-reported number for bid management, but do not treat it as revenue truth. The bid algorithm only learns from conversions it can see; feeding it Shopify's 1,980 would be a different signal, and the 1,510 analytics figure uses a different attribution model. For executive reporting, use Shopify orders matched to Microsoft clicks where possible, and disclose the unmatched remainder. Reconcile in this order: (1) confirm the purchase goal fires once per order on /checkout/confirmation and that the /cart tag is not counted as a purchase; (2) align time zone, currency, and 30-day window; (3) separate view-through from click-through conversions; (4) check consent, ad blockers, and cross-device; (5) match the 180 offline calls and remove the 30% without click IDs from platform CPA. The 740-order gap is too large for normal attribution noise. A plausible pattern is under-tracking on the Microsoft side plus Shopify counting all channels, with analytics counting modeled or multi-touch conversions. Report a reconciliation table: Shopify orders, Microsoft-click-matched orders, platform conversions, and unexplained gap. Do not present any single number as 'true' ROAS until the gap is under 10%.
2. What must be fixed before a 40% budget increase, and what would you monitor in the first 30 days?
Read the worked answer
Fix tracking before scaling: enable enhanced conversions, remove the duplicate /cart purchase tag, consolidate to one purchase goal, set the account time zone and currency to match Shopify, and document the 30-day window. Move offline call imports to the Microsoft Advertising Conversion API or at least require a click ID; exclude unmatched calls from bidding. Set a tROAS near 600% only after 30 days of clean data-Maximize Conversions without a target is not appropriate when the CMO expects efficiency. Run a 30-day holdout or geo test to estimate incrementality. Monitor daily: platform vs Shopify order gap, conversion latency, revenue by campaign, new-customer rate, return rate, and cost per acquired customer. Scale in 20% steps, not 40%, and pause if the gap widens or CPA rises more than 15%.