Guide · · 12 min read
E-commerce Fulfillment SLAs in Canada: The KPIs Brands Should Put in a 3PL Contract
A useful fulfillment SLA does more than promise fast shipping. It defines each metric, the clock, exclusions, evidence, reporting cadence and what happens when performance misses the agreed standard.
An e-commerce fulfilment agreement can look precise while leaving the decisions that matter unresolved. A promise to ship quickly means little unless the parties agree when an order becomes eligible, which time zone controls the cutoff, how weekends and statutory holidays are treated, and which system provides the evidence. The same problem appears in accuracy, receiving, inventory and claims: a percentage without a definition is difficult to manage and even harder to improve. A strong ecommerce fulfillment SLA Canada framework turns the brand's customer promise into a small set of observable warehouse commitments. It also recognizes that the 3PL does not control every event. A warehouse can control correct picking, timely label creation and tender to a carrier; it cannot guarantee every carrier scan or weather-affected delivery. This guide explains how to define contract-ready KPIs, assign ownership, establish reporting and create a fair review process. It is an operational planning resource, not legal advice; qualified counsel should review the final agreement and any remedy language.
Start with the anatomy of a measurable SLA
Every metric should contain six elements: scope, starting event, stopping event, calculation, exclusions and evidence. Scope identifies eligible orders or receipts. The start and stop events define the clock using system timestamps rather than recollection. The calculation specifies numerator, denominator and rounding. Exclusions should be narrow, visible and coded at order level; examples may include a brand-requested hold, missing product data or a documented platform outage. Evidence identifies the authoritative WMS, commerce platform or carrier event. Finally, establish the measurement period and minimum sample size so a small number of orders does not create misleading percentages. Keep service targets separate from broader objectives. An objective can guide continuous improvement, while an SLA carries the review or remedy process written into the contract.
A percentage is not yet an SLA. Pair every target with an eligible population, start and stop timestamps, exclusions, data source, reporting period and owner.
Define order and fulfilment accuracy without ambiguity
Order accuracy should measure whether the customer received the correct SKU, quantity and, where applicable, lot, personalization or insert. Decide whether damage caused by warehouse handling is included and whether carrier damage is tracked separately. A defensible formula is accurate eligible orders divided by all eligible shipped orders, but the parties must agree how an error is substantiated. Customer complaints alone can overstate or understate errors; supporting photographs, return inspection and WMS scan history can help adjudicate exceptions. Track error reasons beneath the headline metric—wrong item, shortage, overage, wrong lot, packaging error and address transcription—because each requires a different correction. Also define whether an order with two errors counts once at order level or twice at line level. Reporting both order and line accuracy often gives the clearest operating picture.
- Name the required dimensions: SKU, quantity, lot, customization, insert and packaging.
- Separate warehouse errors from customer-data and carrier events.
- Set an evidence and dispute window for reported inaccuracies.
- Preserve reason codes so corrective action addresses the real failure mode.
- State whether the metric is calculated by order, line or unit.
Connect order cutoff to an observable carrier handoff
A same-day dispatch SLA needs an eligibility rule. Define the exact local time at the fulfilment facility, days of operation, payment and fraud status, inventory availability, address validity and any orders that require special handling. The start event should normally be the time an eligible order is released to the WMS, not the time the shopper clicks buy. Label creation is not proof that a parcel left the building. Use a documented manifest close, trailer departure, carrier acceptance or first physical scan as the stop event, and report late first scans separately when custody transfer occurred on time. In Canada, pickup schedules and service availability vary by location, carrier and day, so calendars and cutoff changes should be maintained as controlled reference data rather than assumed.
Measure dock-to-stock from an appointment-ready receipt
Dock-to-stock measures how quickly usable inbound goods become available for allocation. Start the clock only when the shipment arrives within its confirmed appointment window with the required advance shipment notice, labels and documentation. Stop it when the accepted units are counted, inspected as agreed, put away and available in the WMS. Quarantined, damaged, unidentified and over-delivered goods should move to separate exception clocks rather than disappear from the denominator. Brands should also define priority rules for launches and stockouts, plus appointment capacity during peak periods. Reporting median and percentile elapsed time alongside the SLA pass rate exposes a long tail that an average can hide. The receiving metric should be paired with receipt accuracy so rushing stock into availability does not undermine count or lot integrity.
| Metric | Suggested start | Suggested stop | Key segmentation |
|---|
| Order accuracy | Eligible order released | Verified shipment or claim decision | Order, line, error reason |
| Cutoff performance | Eligible release before cutoff | Documented carrier handoff | Facility, channel, service |
| Dock-to-stock | Compliant inbound checked in | Accepted stock available in WMS | Supplier, receipt type, exception |
| Inventory accuracy | Count population established | Count reconciled | SKU, location, variance reason |
| Claims cycle | Complete claim received | Decision communicated | Claim type, carrier, ageing |
Treat inventory accuracy as a controlled reconciliation process
Inventory accuracy is not simply the portal matching itself. Define whether the comparison is book-to-physical by unit, location, SKU or value, and how unavailable statuses such as quarantine, damage and pending returns are represented. A sound programme combines continuous cycle counts, targeted counts after anomalies and periodic reconciliation. The SLA should specify count frequency by velocity or risk class, tolerances where appropriate, recount procedures, adjustment authority and root-cause coding. Virtual bundles deserve special attention because their availability depends on component stock. Inventory movements should retain user, time, source document and reason code. When a variance affects open orders, the escalation clock should be distinct from the monthly accuracy calculation so the immediate customer impact is addressed promptly.
Create separate clocks for claims, exceptions and reporting
Claims work crosses several parties, so one broad resolution promise is rarely fair or useful. Define a 3PL acknowledgement time, evidence-completion check, carrier submission time where relevant, customer update cadence and closure event. Identify who provides invoices, photographs and product-cost evidence, and who owns communication with the shopper. Track warehouse-caused loss or damage separately from carrier claims and concealed product defects. For routine reporting, agree on dashboard availability, monthly close timing, data correction rules and a named owner for anomalies. A monthly service review should examine trends and corrective actions, while urgent incidents need an immediate escalation path. No service-credit formula can compensate for weak event data, so access to order-level records should be part of the reporting design.
Do not use one clock for an entire carrier claim. Separate warehouse acknowledgement, document readiness, carrier submission, status updates and final disposition.
Build a balanced, contract-ready scorecard
Limit the contractual scorecard to outcomes material to customers and inventory, then retain diagnostic measures for operations. Weighting everything equally can encourage the wrong behaviour—for example, protecting dispatch speed by bypassing quality controls. Specify a reporting cadence, review date, cure process and governance path before discussing financial remedies. Consider whether remedies apply only after a repeated miss or material threshold, whether they are capped, and whether chronic failure triggers a transition discussion. Those choices require commercial and legal review. Operationally, the scorecard should show target, actual, eligible volume, excluded volume, prior trend, variance explanation, action, owner and due date. Both parties should be able to reproduce the calculation from exported data.
- Map each customer promise to the party that can control it.
- Test formulas against historical orders before contract signature.
- Require order-level exclusion codes and an audit trail.
- Review targets after material channel, SKU or volume changes.
- Pair each miss with corrective action, owner and completion date.
Validate the SLA during implementation and improve it in operation
Before launch, run sample data through every formula, including split shipments, cancellations, backorders, bundles, weekend orders and manual holds. Agree which clock wins if commerce, integration and WMS timestamps differ. During an initial stabilization period, report the full scorecard even if formal remedies have not begun; this reveals faulty definitions before they become disputes. After go-live, use a regular operating review to distinguish one-off incidents from systemic issues and to approve controlled definition changes. A mature 3PL relationship does not hide misses behind exclusions, nor does it penalize events outside a party's control. It uses reliable data to protect the customer promise, resolve exceptions and invest in the process constraints that matter most.
Frequently Asked Questions
What should an ecommerce fulfilment SLA in Canada include?
At minimum, define order and line accuracy, order cutoff performance, carrier handoff, dock-to-stock, inventory accuracy, claims handling and reporting. Each needs eligibility rules, timestamps, a formula, exclusions, evidence, cadence and ownership.
Is label creation the same as carrier handoff?
No. A label only proves that shipment data was created. Handoff should rely on a manifest, physical acceptance, trailer departure or carrier scan agreed by the parties, with late or missing scans reported separately.
How should dock-to-stock be calculated?
Measure from check-in of a compliant, appointment-ready inbound to the time accepted inventory is put away and available in the WMS. Track damaged, unidentified or quarantined units through explicit exception workflows.
Which inventory accuracy formula is best?
It depends on the operation. Unit-, location-, SKU- and value-based measures answer different questions. Define the comparison, statuses, tolerance, count programme and reconciliation method rather than relying on an unlabeled percentage.
Should a contract include service credits?
That is a commercial and legal decision. If used, remedies should be tied to reproducible measures, proportional to impact, subject to clear exclusions and paired with cure and governance provisions reviewed by qualified counsel.
How often should SLA performance be reviewed?
Operational dashboards may update daily, with urgent exceptions escalated immediately. A monthly scorecard and service review is common for trend analysis, while targets and definitions should also be revisited after major changes in channels, products or volume.
Related ByExpress resources