How to Raise Your COD Confirmation Rate
Confirmation is the cheapest place to stop a bad order, and the most expensive place to be slow. Here is what actually moves the confirmation rate — speed, script, channel mix and the checkout fields upstream of the call.
Tassyir Team
7 min read
Key takeaways
Confirmation rate is time‑sensitive above all else. Orders called within an hour confirm far more often than orders called the next day.
Most unconfirmed orders are not fake — they are unreachable. Fix phone number capture at checkout before you hire another agent.
Lead with WhatsApp, fall back to a call. A message costs nothing to send, arrives instantly and gets answered by people who ignore unknown numbers.
State the delivery fee and the total on the call, every time. Surprise at the door is a return, and a return costs you both legs of shipping.
Three attempts across two days is the practical ceiling. Past that, cancel and move on rather than dispatching a parcel with a weak confirmation behind it.
Confirmation is the cheapest place in a cash‑on‑delivery operation to stop a bad order. Every order you cancel on the phone costs you one call. Every order you fail to cancel and dispatch anyway costs you outbound shipping, return shipping and a parcel out of stock for a week.
Which makes the confirmation rate a strange metric: you do not actually want it at 100%. You want it as high as it can honestly go, with the orders that were never real filtered out early rather than at the customer’s door.
Why orders go unconfirmed
Before changing anything, split your unconfirmed orders into three buckets. They have completely different fixes and lumping them together is why most teams plateau.
Unreachable. The number does not ring, is switched off, or is wrong. Usually the largest bucket, and it is a checkout problem, not a call‑centre problem.
Reachable but unanswered. The phone rings and nobody picks up. This is a channel and timing problem.
Reached and declined. The customer answers and says no. This is the only bucket that is genuinely about the offer, the price or the delivery fee.
Teams instinctively try to fix the third bucket, because it is the one they hear. In most stores it is the smallest of the three.
Speed is the single biggest lever
Intent decays fast. Someone who saw an ad, tapped through and typed their number in under two minutes is at peak intent right then. By the following morning they have forgotten the ad, and by the second day they no longer recognise your number.
If you change one thing, change how fast the first attempt goes out. Calling within the hour rather than the next working day routinely moves confirmation rates by double digits, and it costs nothing but scheduling.
Practically that means confirmation has to run whenever orders come in — including evenings, when most COD traffic converts. A queue that only gets worked 9 to 5 will always underperform on orders placed at 21:00.
Fix phone capture before you hire another agent
A large share of unconfirmable orders were doomed at checkout. Some straightforward changes upstream:
Validate the number format on the form. Reject anything that is not a plausible mobile number before the order is even created. A typo caught here is free; caught later it costs three call attempts.
Use a numeric keypad field. Obvious, and still missing on plenty of stores. Every keystroke of friction on a phone field produces mistyped digits.
Ask for one number, not two. An optional second number sounds helpful and mostly collects the same digits again. Better to make one field mandatory and validated.
Show why you need it. A single line under the field explaining that you will call to confirm before shipping visibly improves the quality of what people type.
Lead with WhatsApp, fall back to a call
Plenty of people who will not answer an unknown call will reply to a message within minutes. A message also does something a call cannot: it leaves the address and the total in writing, where the customer can check it and correct it.
A sequence that works well in practice:
Immediately: a message confirming the item, the full address, the delivery fee and the total, asking for a yes.
After 60–90 minutes with no reply: the first call.
Same evening: a second call at a different hour of the day.
Next day: one final message before cancelling.
Varying the hour matters more than varying the day. Two calls at 10:00 on consecutive mornings reach a narrower slice of people than one at 10:00 and one at 19:00.
What the confirmation call should actually say
A confirmation call has one job beyond getting a yes: removing every reason the parcel could be refused at the door. That means saying the uncomfortable part out loud.
Re‑read the address back. Wilaya, commune and the landmark. Most failed deliveries are addressing failures, not refusals.
State the delivery fee separately and the total explicitly. Surprise at the door is a return. If the fee is going to annoy them, it is far cheaper to find out now.
Confirm the variant. Size, colour, quantity. Cheap to check, expensive to get wrong.
Give a delivery window. People who know a parcel is coming this week answer the courier’s call. People who have forgotten do not.
Know when to stop
Three attempts across two days is the practical ceiling for most stores. Past that, the marginal order you rescue tends to be a weak one that comes back as a return anyway — and each additional attempt costs agent time that would confirm two fresh orders.
Cancelling cleanly is a legitimate outcome. An order that never ships costs you a few minutes. An order that ships on a shaky confirmation costs you two shipping legs and a week of stock.
Measure it properly
Report confirmation rate alongside the two numbers it exists to protect — the delivery rate and the return rate. Confirmation rate on its own can be gamed by confirming anything that answers the phone, which is exactly how stores end up with a great confirmation number and a terrible return number.
The pairing to watch is confirmation rate up, return rate flat or down. If confirmation climbs and returns climb with it, the call is not doing its job.
Once confirmation is running well, the next number to attack is your return rate.
Frequently asked questions
What is a good COD confirmation rate?
80–90% of placed orders confirmed is a healthy target for a store with clean traffic. If you are below 70%, the problem is almost always reachability — wrong or mistyped phone numbers — rather than customers changing their minds.
How quickly should you call a new COD order?
Within the hour, and ideally within fifteen minutes. Intent decays fast on impulse purchases. The same order called at 30 minutes and at 24 hours are effectively two different orders with two very different confirmation rates.
Should you confirm by phone or by WhatsApp?
Both, in that order of fallback: WhatsApp first, call second. A message reaches people who will not pick up an unknown number, it leaves a written record of the address and total, and it costs nothing per attempt. Reserve calls for orders that go unanswered.
How many times should you try to confirm an order?
Three attempts spread across two days. Beyond that the marginal confirmation is not worth the agent time, and an order confirmed reluctantly on the fourth attempt has a much higher chance of being refused at the door anyway.
Does confirming orders actually reduce returns?
Substantially, provided the call does real work. A confirmation that only asks “did you order this?” barely helps. One that re‑reads the address, states the delivery fee and the total, and gets a clear yes removes most of the reasons a parcel comes back.
Tassyir Team
E‑commerce operations, Tassyir
We build the platform Algerian merchants use to run cash-on-delivery stores — orders, stock, couriers and finances in one place. Everything here comes out of the operational data and merchant conversations behind that product.
