On-time delivery: the number that flatters you

What on-time delivery measures

On-time delivery (OTD) is the percentage of orders delivered by their promised date. It is calculated against the date you committed to, not the date the carrier estimated, which is an important distinction most brands blur. If you promise three days at checkout and the carrier takes five, that is a miss regardless of what the carrier's own performance dashboard says.

OTD is the single cleanest measure of whether your operation does what your storefront claims.

Every order is a promise. Keeyu keeps the promise. On-time delivery is that promise expressed as a number.

Why the number usually flatters you

Most OTD reporting is calculated on delivered orders. That quietly excludes the worst cases: the order that never shipped, the one still sitting in a warehouse queue, the parcel that stopped scanning and was never resolved. Those do not appear as late deliveries because they are not deliveries at all.

It once took a retailer two and a half weeks to locate an order of mine. Not to deliver it, to locate it. That order was invisible to any on-time calculation because it had not moved into a state that gets measured.

If you want an OTD number that means something, measure against every order placed, and count unresolved orders as misses.

Where on-time delivery actually gets lost

• Before the warehouse: orders that never synced and sat unfulfilled
• In the warehouse: fulfilment holds, stock discrepancies, oversells
• At handover: shipments created but never collected
• In transit: stalled parcels, failed attempts, exceptions nobody actioned

Only the last of those is a carrier problem, and it is the one most brands spend their time on. At Clutch Glue, 70 US Shopify orders never reached the warehouse and went three days without shipping. Not one of those was a carrier failure, and all 70 were about to be late.

The difference early detection makes

One of our customers had a consumer place a roughly $500 express order. We spotted the delay early, notified the fulfilment team, and the customer received the order on time. The promise held, and the customer never knew it had been at risk.

That is the whole mechanism. Almost every late delivery was preventable at some earlier point where the order stopped moving and nobody was watching. The window between the break and the customer noticing is where on-time delivery is won.

Peak trade and the promise you make at checkout

During peak, brands reliably overinvest in acquisition and underprepare fulfilment for the volume it creates. At P.E Nation and SurfStitch I made sure the warehouses had staff to process what we were about to sell, because the orders were coming whether or not anyone had planned for them.

The other half is the promise itself. A delivery estimate that assumes everything goes right is not a promise, it is a hope. Set the date you can hold under normal failure rates, then build the detection that catches the exceptions.

How to improve it

Measure against orders placed rather than orders delivered. Instrument time-in-state at every stage so a stalled order raises its hand on its own. Define a decision rule per failure type, reship, refund, redirect or claim, and automate the routine cases. Tell the customer before they ask when a date is going to move.

That is proactive e-commerce operations. Detect. Decide. Act. The customer gets what they want, on time, as promised.

See how the workflows run or book a demo.

Related reading

For stalled shipments, read delivery exceptions. For the fulfilment layer, see fulfillment. For the tickets lateness generates, read WISMO.