Fulfillment service levels: define cutoffs, accuracy and escalation

When does the dispatch commitment start?

Use an agreed eligible-order definition and record the events that start and stop the clock.

An order can enter the store before it is ready for the warehouse. Payment may still be pending, an address might need correction or an item may be unavailable. List the conditions for release and ask which timestamps the provider uses. Compare proposals only after the starting conditions are understood.

Identify the final event too: picking completed, a label created, a parcel ready for collection or handover to the carrier. These describe different stages. Record which one the provider commits to and how it is evidenced. Your customer-facing promise should be based on the full route rather than a warehouse event taken out of context.

How should cutoff times and peak periods be described?

Specify timezone, working days, eligible orders and the treatment of late or blocked work.

Ask what happens to an order released just before the cutoff and one released immediately after it. Include weekends and the calendar actually used by the warehouse. If a promise depends on pre-booked capacity or forecast volumes, record the conditions and how the merchant is told that capacity is no longer available.

For a peak, agree the forecast information required, the point at which it is reviewed and who approves a change. Do not invent a universal peak multiplier or assume that a standard commitment scales automatically. Ask the provider to explain how your proposed peak scenario would be staffed, prioritised and reported.

How can accuracy and timeliness be measured meaningfully?

Define the numerator, denominator, observation period and exclusions before comparing a percentage.

For order accuracy, decide what counts as an error and whether the measure concerns orders, lines or units. For dispatch timeliness, define the eligible population and the committed event. Keep exclusions visible. A headline figure without these definitions cannot tell you how your own exceptions would be treated.

Add case-level records behind each summary. A merchant needs to locate the affected order, identify the failure and agree a next action. Review small samples of reported successes and failures during onboarding to check that both teams interpret the definitions the same way.

Who owns blocked orders and cancellation requests?

Assign each exception a reason, an owner, a next action and a visible current state.

An order held for missing data should not disappear from the working queue because it is excluded from a time measure. Agree how the warehouse reports it, who can supply the correction and how the clock behaves after release. The merchant and provider should see the same unresolved case rather than separate unexplained totals.

For Shopify fulfillment services, cancellation can be requested after the service has accepted the work, and the request can be accepted or rejected. This is a useful product example of why a request and a confirmed stop are different states. Define the equivalent response path for your own warehouse connection.

Reference: Shopify: fulfillment service apps

What belongs in the review and escalation process?

Review recurring causes and unresolved work alongside summary measures, and document decisions.

Set the review frequency to the operation’s needs. During onboarding or a peak, a short issue review may be more useful than a monthly score alone. Keep the discussion concrete: affected orders, evidence, cause, corrective action and an owner. Decide which repeated failures trigger a change to the workflow.

Agree an escalation route for urgent and recurring problems, including the authority to pause releases if needed. Ask the provider to explain its proposed contractual remedies and have those terms reviewed where appropriate. This worksheet helps frame the operating discussion; it does not prescribe legal wording or guarantee a particular remedy.