Data Quality
Part of Ecommerce customer value analysis
Handling incomplete histories in a customer-value report
Separate short follow-up, missing earlier orders, uncertain identity and unsettled adjustments in customer-value reports.
An incomplete purchase history needs a visible status in a customer-value report. Short follow-up, missing earlier orders, uncertain identity and unsettled adjustments each affect a different part of the result. Classify the gap before calculating or comparing value.
| Gap | Effect on the report | Treatment |
|---|---|---|
| Follow-up still open | Later purchases have not had the full opportunity to occur | Show a partial-age result outside the completed-window comparison. |
| Earlier orders unavailable | The apparent first order may not be the actual first | Label entry and new-customer status unverified. |
| Orders cannot be linked confidently | A person may appear under several customer keys | Show unmatched orders and identity coverage separately. |
| Adjustments still arriving | Retained value may change | State the adjustment cut-off and mark value provisional. |
A customer with 30 days of follow-up can have a verified first order. A migrated record may have 90 days of recent follow-up, but its true first order remains unknown. A single missing-data percentage would conceal that difference.
Impact of Data Gaps on Customer-Value Reporting
- Follow-up still open
- Partial-age result; not comparable to completed window
- Earlier orders unavailable
- First order status unverified; avoid labelling as confirmed new customer
- Orders cannot be linked confidently
- Identity coverage shown separately; unmatched orders flagged
- Adjustments still arriving
- Value marked provisional; adjustment cut-off stated
Protect the completed-window population
For a completed 90-day view, include entrants whose entry precedes the observation cut-off by the full 90 days. Count and report entrants omitted because their follow-up is still open. Do not fill their remaining days with zero purchases and label the result complete.
If earlier orders are unavailable, restrict the analysis to customers with verifiable entry histories, or create a separately labelled first-observed-order group. The restricted result describes that covered population, which may differ from all customers. Do not describe first observed orders as confirmed new customers.
Shopify lists Customer cohort analysis among its Customers reports, but the description does not establish whether imported or missing earlier history is complete. Treat uncertain identity as unresolved unless verified. An apparent match alone does not confirm that records belong to the same person or that their histories are complete.
Steps to Ensure Integrity in Customer-Value Reports
- Flag incomplete follow-upDo not fill remaining days with zero purchases; label results as partial-age
- Restrict analysis where history is uncertainUse verifiable entry data or create a separately labelled group
- Treat uncertain identity as unresolvedDo not assume imported or migrated records are complete
Show adjustments and coverage
An adjustment can change an order's retained value after its purchase window has closed. Shopify sales reports define orders by the date they were placed; they describe sales reversals as order adjustments that produce negative monetary value. To follow a customer group, associate later adjustments with the selected order IDs through a stated cut-off, not just a recent trading-period total.
Report eligible entrants, complete-window entrants, unverified entry histories, unresolved identities, unmatched order value and the adjustment cut-off. If a gap concentrates in one product or acquisition group, avoid ranking groups until you understand its effect. Remedies may include waiting for follow-up, recovering older records, improving permitted identity linkage or refreshing adjustments. Explain any resulting revision to an earlier figure.
Key Metrics to Report in Customer-Value Analysis
- Eligible entrants
- Total customers meeting entry criteria
- Unverified entry histories
- Customers with missing earlier order data
- Unresolved identities
- Records without confirmed matching across systems
- Unmatched order value
- Total value from orders not linked to a single identity
- Adjustment cut-off date
- Latest date used for adjusting retained value
