Define the status model
Write what counts as dispatched, delivered, refused, returned, cancelled and pending. Courier and store status labels may not have identical meanings. Reconcile material differences before using the rate to judge campaigns or regions.
Also specify whether collection is part of the completed business outcome. Delivery and cash reconciliation can occur at different times.
Calculate the rate
- Select a cohort using a clear dispatch or order-start date.
- Allow the relevant observation window to complete.
- Count eligible delivered outcomes.
- Divide by the stated denominator.
- Report pending and other statuses separately where needed.
Do not silently change the denominator between reports. A rate based on resolved attempts can look higher than one based on all dispatched orders still awaiting outcomes.
Worked example: two distinct questions
Illustrative example: 120 orders are created, 80 are dispatched and 60 are delivered after the defined window. Delivered share of dispatched orders is 60 / 80 = 75%. End-to-end delivered share of created orders is 60 / 120 = 50%.
Both can be useful. The first inspects fulfillment progression; the second includes losses before dispatch. Neither should be presented without its definition.
Interpretation table
| Pattern | Investigation |
|---|---|
| Low confirmation before dispatch | Intent, contactability and expectation clarity. |
| Low delivery after dispatch | Address, delivery execution and customer expectations. |
| Many pending orders | Outcome timing and operational follow-up. |
| Delivery strong, collection unclear | Reconciliation and cash timing. |
Link the rate to economics
A high delivery rate does not by itself establish profitability. Product margin, acquisition cost and courier charges still matter. Conversely, the rate can improve while volume falls because the business filters unsuitable demand more effectively.
Should pending orders count as failures?
Not automatically. Use a defined observation window and show unresolved outcomes. Treating fresh orders as completed failures can distort comparisons.
Can campaigns be compared by delivery rate?
Yes, where the operational link is reliable and cohorts are comparable. Account for product, region and timing differences.
Put this into practice
Add the denominator, cohort date and pending count to the delivery report. Recheck any campaign decision previously made from a rate whose definition was unclear.
Related foundation: Meta Ads for e-commerce in Morocco: measure delivered orders. 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.
