Campaign Data Refresh & Analysis Checklist

How to use this

Run this before every later CRO campaign, after Evaluation & Decision has closed the previous campaign. Progress is saved in this browser on this device and is not shared. Reset only when the current campaign is closed and the team is ready to gather evidence for the next one.

Use the latest four complete weeks against the preceding four complete weeks, plus the historical baseline. Four complete weeks equals 28 complete days and keeps the weekday mix aligned. Use another clean comparison only when a promotion, stockout, outage, tracking failure, theme launch, acquisition change, dated event, or seasonal question makes the default misleading. Record the exception before analysis.

Reset for a new campaign

1. 1. Scope the refresh

   Confirm Client Readiness and the Initial Baseline are still usable Define the business question, priority surfaces, audience, and intended decision Lock the current four-week window, preceding four-week comparison, and historical baseline Record timezone, currency, promotions, stock, outages, traffic changes, and other known context

   Required output: A bounded refresh request with comparable periods and known context.

2. 2. Collect quantitative evidence

   Refresh the Shopify store and funnel metrics Refresh channel and attribution data Refresh landing-page or page-level data for the priority scope Refresh product and offer performance Refresh mobile and desktop performance Refresh LCP, INP, CLS, and relevant technical-health data Pull relevant GA4 data to cross-check or fill known measurement gaps Pull relevant Klaviyo and Search Console evidence for the question

   Required output: A current quantitative dataset covering the question, comparison window, and major segments.

3. 3. Collect behavioral and operating evidence

   Review heatmaps for the priority surfaces Review representative session recordings for the same scope Refresh customer language from reviews, support, surveys, interviews, or ad comments Review experiment and rollout history that affects the current control Record promotion, campaign, acquisition, and audience changes Record stock, merchandising, outage, theme, app, and tracking changes

   Required output: Current behavioral, customer, and operating context aligned to the quantitative window.

4. 4. Validate the refresh

   Confirm date coverage, daily grain, timezone, currency, and complete periods Merge and deduplicate overlapping source rows Reconcile totals and metric definitions across sources that should agree Flag unknown attribution, null mappings, impossible rates, and other anomalies Flag partial periods, bots, staff or QA traffic, consent gaps, tracking failures, and duplicate events Record missing inputs, cleaning decisions, and their effect on confidence

   Required output: A reproducible refresh with a complete quality and caveat log.

5. 5. Analyze the evidence

   Compare funnel movement with the preceding four weeks and historical baseline Compare channel and device mix and performance Compare landing-page, page-level, product, and offer performance Review weekly and daily movement for spikes, drift, and reversals Separate traffic-quality changes from on-site conversion changes Synthesize behavioral and customer themes, labelling thin or biased evidence appropriately Triangulate the strongest findings with at least two evidence types and separate observation from interpretation

   Required output: A ranked analysis showing what changed, where the strongest leak appears, and how confident the team should be.

6. 6. Build the campaign evidence pack

   Name the strongest supported leak and the decision it creates Record the current baseline, comparison windows, and source links Record observed evidence separately from interpretations and proposed explanations Record caveats, unresolved gaps, and the evidence still needed Confirm Data & Strategy can frame the campaign without recollecting or guessing at the evidence

   Handoff gate: Data & Strategy receives one current evidence pack with a ranked leak, source links, measurement window, relevant segments, operating context, and explicit caveats.

Continue to Data & Strategy