Compare definitions before numbers
One tool may report orders, another customers and another conversion events. Some include taxes or shipping; others apply different adjustments. These differences can explain discrepancies without any technical failure.
Also check time zones, currencies and whether the current period is complete. A midnight boundary or delayed import can move records between days.
Reconcile a sample
- Write each report's metric definition and date basis.
- Compare the same eligible population and period where possible.
- Inspect a sample of records or aggregate checks without exposing unnecessary personal data.
- Identify duplicates, missing events and known attribution differences.
- Document unresolved gaps and the decision they affect.
Do not add attributed revenue from several advertising platforms as if each were a separate set of orders. Multiple systems can credit the same purchase.
Difference table
| Difference | Check |
|---|---|
| Event totals | Meaning, duplicate firing and eligibility. |
| Revenue | Currency, discounts, tax, shipping and adjustments. |
| Daily timing | Time zone, import delay and date basis. |
| Channel credit | Attribution window and model. |
Worked example: overlapping credit
Illustrative example: two platforms each credit part of the same customer's purchase under their rules. Summing the credited amounts may exceed the actual order value. Keep the operational total separate and report platform credit as platform credit.
The discrepancy does not prove one channel is useless. It means the reports cannot be treated as independent additive business records.
Turn reconciliation into an operating habit
Assign an owner and keep a short change log. When tracking or reporting definitions change, annotate the date and affected metrics. A recurring small audit can be more useful than a large emergency exercise after the team has already scaled spending.
Must every report exactly match the store?
Not necessarily. Their purposes and definitions may differ. The important issue is whether differences are understood and material decisions use suitable evidence.
What if the gap cannot be fully explained?
State the uncertainty, use the most reliable business record for the immediate question and prioritize the investigation according to its decision impact.
Put this into practice
Take one material discrepancy, compare definitions and reconcile a small sample. Write the known explanation and remaining uncertainty beside the dashboard rather than hiding it.
Primary-source reading for platform details: Shopify: analytics field definitions; Google Analytics: attribution.
Related foundation: CAC vs ROAS: which metric should guide your marketing budget?. How these guides are prepared.
Related portfolio work: Etsy & Shopify. The worked examples in this guide are illustrative and are separate from the portfolio’s project evidence.
