-
Repeat rate within the windowShare of first-time customers who ordered again within, say, 180 days.
-
Median days to reorderHow quickly the second order comes, not just whether it does.
-
Grouped by acquisition serviceThe service on the customer's first order, not the carrier and not a blended average.
-
Matured customers onlyFirst order at least a full window old, so everyone in the cohort had the chance to return.
There is a belief a lot of DTC founders carry without ever testing it: the customer you won on fast shipping is worth more than the one you won on the cheap, slow option. The story is intuitive, a faster, cleaner first delivery makes a better impression, so that buyer comes back and the ground buyer drifts. It might be true for your brand. It might not, or the effect might be too small to pay for. The trouble is that the assumption usually gets treated as settled when it was never measured, and an untested belief like this quietly sets your shipping budget. If it is wrong, you are either overpaying for speed or starving the lever that actually drives repeat orders.
It is a hypothesis. And it is a measurable one.
Group retention by the service that shipped the first order
The mistake is looking at repeat rate across all customers, or splitting it by carrier. Neither answers the question. The question is about the experience a customer had when they were acquired, so the unit of analysis is the service on their first order: FedEx 2Day versus FedEx SmartPost, Priority versus Ground Advantage, whatever tiers you actually ship.
For each first-time customer, tag the service of that first shipment. Then, within each service, measure two things:
- The share of those customers who placed a second order within a fixed window, say 180 days.
- The median time to that reorder.
One number tells you whether the tier retains, the other tells you how quickly. A tier that produces a higher repeat rate and a faster second order is doing real work. A tier that costs three times as much and produces the same repeat rate is a cost you are absorbing for no return.
Only measure customers who have had the chance to come back
A cohort is only fair if every customer in it has had the full window to reorder. A buyer acquired three weeks ago cannot have a 180-day repeat rate yet, and if you leave them in the pool they drag the number down and make your newest, fastest tier look worse than it is. Measure only "matured" customers whose first order is at least a window old. Recent tiers you rolled out last quarter will not have enough matured customers to judge yet, and that is fine. Show the tiers that can be judged and be honest that the rest need more time.
The gap is a signal, not a verdict
If the fast-acquired cohort does repeat more, resist reading it as pure cause. The customer who paid for or chose faster shipping is often a different customer to begin with: higher intent, higher average order value, a product they wanted sooner. Some of the retention gap is the delivery experience, and some of it is who selected that experience. You are looking at correlation with a plausible mechanism, not a clean experiment.
That is still useful. It tells you which direction to lean and roughly how much is at stake. If the gap is large and consistent across months, faster shipping is behaving like a retention investment, and you can weigh its cost against the repeat revenue it appears to protect. If the gap is small, the speed you are paying for is not buying loyalty, and the money is better spent elsewhere. Either way you are now deciding with a number instead of a belief.
This is the same idea as treating lifetime value as a carrier metric: the delivery experience does not stop at the front door, it shows up later in whether the customer comes back. Acquisition service is just the sharpest place to look for it, because it is the first experience the customer ever had with you.
What to do with the answer
Measure it, then act on it in one of three ways. If a faster tier clearly retains better and the math holds, protect it, and consider pushing more first orders onto it even at higher cost. If a slower tier retains just as well, stop paying for speed you cannot see in the reorder data. And if you cannot tell yet because the fast tier is too new, keep the cohort running and check again next quarter. The worst option is the one most brands default to, which is to keep believing the hypothesis and never look.