Service answerHow can we tell which marketing is bringing orders, not just views?
Start with the store's actual orders, refunds and marketing costs. Then compare those records with tagged campaign visits and the conversions each platform reports. Differences can come from attribution rules, timing, modeled results and missing signals. Adding platform totals together can count the same order more than once.
Use a consistent review period and retain unknown sources. For a small brand, a useful report ends with a decision: repeat a creative idea, improve a page, adjust an offer or stop a test. More views alone are not a reason to increase spending.
Choose one reporting window and reconcile spend, recorded purchases, refunds and the store order total before evaluating channels.
Trace a view through to a real business event
Brand advertisinge.l.f. × Tinuiti: define the conversion
Historical TikTok campaign · account looks ahead to 2022
TikTok reports a 56% reduction in acquisition costs for e.l.f.’s add-to-cart strategy, using an ongoing mix of ad objectives. The named partner was Tinuiti.
Patrick O’Keefe, VP Integrated Marketing Communications in the case
Platform self-report. Add-to-cart cost is not customer acquisition cost; campaign dates, attribution detail and profit are not fully specified. Historical ad product names are not setup instructions.
Read the source notes and related topicsUS audience researchAthena Chapekis and Anna Lieb on AI-summary clicks
900 US adults · March 2025 browsing; search results reconstructed in April
Pew observed traditional-result clicks on 8% of visits with an AI summary versus 15% without one. The analysis was published July 22, 2025.
Athena Chapekis and Anna Lieb, Pew Research Center
Observational Google-only research across query types. Results were reconstructed later, and shopping intent was not isolated. This is not a causal forecast of lost ecommerce revenue.
Read the source notes and related topicsHow to apply it
When a report says conversion, ask which action it means: a product view, cart addition, paid order or repeat purchase. The distinction matters in e.l.f.’s published campaign result. Each action can teach you something, but only the order and its costs show whether that sale paid for itself.
Reconcile campaign reporting with the store’s order record over an agreed period. Check refunds, cancellations, customer status and attribution windows. Keep a separate note for discovery that is visible in customer feedback but cannot be assigned confidently to an ad. You do not need perfect attribution to make a decision; you do need consistent definitions and a clear account of uncertainty.
A measurement sheet listing event definition, data source, attribution window, verified net orders and the decision each metric can legitimately support.