Conversion optimization / Measurement guide
CRO for small stores: measure contribution per session
Connect conversion work to the value of each order.
Download worksheetEditable CSV · Open in your spreadsheet app
For a small store, conversion work should improve the value of the orders you can fulfill. Track contribution alongside checkout completion, then decide which problems need a direct fix and which need an experiment.
Choose a consistent denominator
Use one defined source of eligible online-store sessions. A person can return in several sessions, so label a session-based measure accurately. Keep device, channel, market, and reporting period consistent when comparing results.
This matters now: Shopify changed session measurement in September 2026, including bot filtering and how continued activity is counted. A conversion-rate change across that update can reflect measurement rather than changed buying behavior. Review the official measurement update before combining older and newer baselines.
Put contribution beside conversion
For an operating view, define contribution as net merchandise revenue after discounts and returns, less product cost, payment fees, variable fulfillment costs, shipping subsidy, and any media spend included in the analysis. Write the exclusions down. This is a planning measure before fixed overhead, not a complete profit statement.
Contribution per session = contribution for the measured orders ÷ eligible sessions in the same scope.
| Invented comparison | Period A | Period B |
|---|---|---|
| Eligible sessions | 10,000 | 10,000 |
| Completed orders | 200 | 240 |
| Orders per session | 2.0% | 2.4% |
| Average contribution per order | $25 | $18 |
| Total contribution | $5,000 | $4,320 |
| Contribution per session | $0.50 | $0.432 |
These numbers illustrate the arithmetic, not a client result. More orders accompany lower contribution in Period B. A heavier discount or more expensive fulfillment could create that pattern. The table alone cannot tell you what caused it.
Match the method to the problem
Fix reproducible defects such as a broken checkout control or an incorrect delivery message. Record the change and verify that the task works afterward. You do not need to keep a known defect live to create a control group.
For a choice between credible alternatives, plan an experiment with the baseline rate, the smallest effect worth acting on, and an appropriate sample-size calculation. If the store cannot reach that sample within a useful period, use customer task observation, support evidence, and a measured rollout. Report the result as directional when causality is unresolved.
Keep a weekly decision sheet
| Field | What to record |
|---|---|
| Customer task | The step a customer is trying to complete |
| Evidence | Error report, support pattern, or observed difficulty |
| Change | One specific improvement and its release date |
| Main measure | The outcome and exact calculation |
| Guardrail | Refunds, margin, errors, or support workload |
| Context | Promotions, stock changes, traffic mix, tracking changes |
| Decision | Keep, revise, investigate, or roll back |
Devekko's Shopify growth work can start with this sheet and a prioritized review. The first deliverable should make the next decision easier to justify.
Your next step
Review your store’s growth priorities
Share the task, your current systems, and the decision you need to make. We’ll discuss a useful first scope.
Review your store’s growth priorities