Ask most brands what their retention problem is and they'll talk about winbacks. Customers who bought a year ago and drifted off. Big dormant segment, sat there, doing nowt.
That's not where your money is.
Your money is in the four weeks after somebody's first order, and most email programmes are asleep for exactly that period.
A first time buyer has roughly a 27% chance of ever buying again. Get them to that second purchase and the odds of a third are roughly double, somewhere near 54% on the most commonly cited figure.
One decision point. Everything after it gets easier. And the average ecommerce brand converts about 28% of customers into repeat buyers, which means seven in ten people who buy from you once never come back.
That is how the industry works. It is also the largest addressable gap sat in most accounts, and it gets a fraction of the attention dormant customers get.
The window is much shorter than you think
Most brands get this part badly wrong, and they get it wrong in a way that's invisible from the reports.
Analysis of over 40,000 repeat buyers across a DTC portfolio found the second purchase curve is steep and front loaded. Half of all repeat purchases happen within the first 30 days. Three quarters happen within 90. Only about 13% happen between months three and twelve, and under 4% take longer than a year.
The median time to second purchase in that dataset sat between 15 and 35 days. The average sat between 50 and 100 plus.
Use the median. The average is lying to you.
A handful of people who come back after fourteen months drag the mean way out to the right and make the window look far wider than it is. If you've built your post purchase timing off an average, you're arriving after most of your potential repeat buyers have already decided.
There's a smaller detail in there worth noticing too. Over 6% of repeat purchases in that dataset happened the same day. People mid decision, adding something they'd forgotten or splitting an order. If your order confirmation page and your confirmation email have no offer on them, those people are just gone.
What most brands do instead
The standard playbook runs like this, and it is the thing to stop doing.
Order confirmation. Shipping notification. Review request at day 14. Then suppress the customer from campaigns for 30 to 60 days, because you don't want to bother somebody who's just bought.
Line that up against the numbers above. You've built a deliberate silence into the exact period when half your repeat purchases happen.
It feels considerate. Nobody wants the "I've only just ordered and you're already at me again" complaint. But the suppression window is usually set by nervousness rather than data, and it's set once, at account build, and never revisited.
This is where the agency you're paying is meant to earn its money, and where a lot of them don't. The eight standard flows get built, the suppression rule gets copied from the last account, and nobody ever goes back and asks whether the timing matches how this particular brand's customers actually behave.
So the fix starts with knowing your own timings rather than inheriting someone else's.
Work out your own number first
Don't take 15 to 35 days as gospel. It's a portfolio median across mixed categories. Yours will differ, and the difference matters.
Pull your order data and calculate the median gap between first and second order. Do it per product category if your catalogue is broad, because a consumable and a durable behave nothing alike.
If you sell coffee, that gap might be 21 days. If you sell furniture, it might be 400, and your entire retention strategy should be built around referral and complementary categories instead.
You want three figures: the median days to second order, the percentage of first time buyers who ever place a second, and how both of those look split by acquisition channel. Paid social buyers repeat differently to organic ones. Discount acquired buyers repeat differently again.
Two hours in a spreadsheet. It'll tell you more about your business than a month of open rate reporting.
Building for the window
Once you know the number, the structure follows.
Days 0 to 3. The confirmation and shipping emails are the highest opened messages you will ever send. Stop wasting them on logistics alone. Add a complementary product, a small add on window, a reason to come back.
Days 3 to 10. Education, not selling. How to use it, how to get the most from it, what people usually get wrong. This is the period where somebody either becomes a user of your product or sticks it in a cupboard. Cupboard means no second order, no matter what you send later.
Days 10 to 21. The review or feedback ask, and a soft cross sell based on what they bought. If your median is around 21 days, this is where your strongest second purchase push belongs.
Around your median. The direct ask. Whatever your data says the typical gap is, be present a few days before it, not after.
Brands running a structured post purchase sequence see better second order rates than brands sending transactional emails only. You'll find specific uplift percentages quoted all over the industry. Treat them with suspicion, because they're vendor numbers with no methodology attached. The direction is not in dispute. The size of it is anyone's guess until you test it on your own list.
The suppression rule, revisited
You do still need one. The answer isn't to hammer new customers with every promotional campaign.
The workable version is narrower than most: suppress recent purchasers from discount campaigns specifically, not from all sending. They keep receiving the post purchase sequence, product education, and anything non promotional. What they don't get is the 20% off email three days after paying full price, which is the actual thing that generates the complaint you're worried about.
That single distinction resolves it, and it's an afternoon's work in the segment builder.
What this does to your payback period
If you're spending on paid acquisition, the first order is often break even at best. Plenty of DTC brands lose money on order one and have done for years. That's a known and accepted position, provided something comes after it.
The second purchase is where you stop being a charity.
Put numbers on it. CAC of £45, contribution margin of £30 on a first order, £38 on a repeat because there's no acquisition cost attached to it. Order one leaves you £15 down. Order two puts you £23 up. Order three and you're properly in profit.
Now change one variable. Move your first to second conversion from 27% to 35% and nothing else. Across a thousand new customers that's eighty extra second orders, which is roughly £3,000 of contribution you didn't have, with no additional acquisition spend. Then those eighty customers carry the higher third order probability with them into next quarter.
This is why the 1 to 2 window is a finance conversation rather than an email one. It's the cheapest lever on payback period you own, and it's usually being managed by whoever built the flows eighteen months ago.
One warning that follows directly from the maths. If your second purchase push leans on a discount, you're eating the margin that was meant to repair the first order. A 20% discount on order two can wipe out most of the contribution advantage repeat orders are supposed to give you.
Time it right and you often don't need the incentive at all. That's the entire argument for calculating your median rather than guessing.
Why this beats a winback every time
Winback flows are satisfying to build. Big segment, dramatic subject lines, feels like you're reclaiming something.
But you're going after people whose odds of returning have already collapsed. Under 4% of repeat purchases come from people over a year out.
The person who ordered eleven days ago has a 27% chance of coming back, and something you send could move it. The person who ordered fourteen months ago has already told you what they think.
Spend your effort where the probability still moves.
There's a reporting angle to this too. Winback flows show attributed revenue, which is why they survive. Some of that revenue is people who were coming back regardless and happened to open an email on the way. Test one with a holdout and you'll often find the incremental number is a fraction of the reported one.
The post purchase window rarely gets that scrutiny, because it's less glamorous and the emails are less fun to write.
Work out your median gap. Build the sequence around it. Then go and look at what your suppression rule has quietly been doing for the last two years.
If nobody can tell you what your median is, that's the more useful finding.