Post-purchase flow: the complete guide

Write down the goal before writing any messages, because the sequence looks completely different depending on the answer. In most stores the goal is the second order, and everything in the flow either moves towards it or protects it.

That means the sequence has four jobs, and they come in a fixed order. Reduce anxiety while the customer waits. Make sure the product succeeds when it arrives. Collect the proof and feedback that helps the next customer. Then ask for the next purchase, once the first one has been a good experience.

Skipping to the fourth job is the common mistake. A discount on the next order sent two days after somebody bought, before the parcel has even arrived, asks them to buy again from a store they have not yet had a result from. It converts a few people and wastes the window on everyone else.

The shape of the flow depends on what you sell. A consumable bought every six weeks makes replenishment the centre of the sequence, and the timing of that one message matters more than everything else combined. A considered purchase bought once every few years makes accessories, care, referral and the ongoing relationship the point, and the reorder ask barely features. A gift purchase means the buyer and the user are different people, which changes what usage advice can even say.

Work out your own reorder gap from your order history before designing anything. Take the customers who bought twice, measure the median gap between their first and second order, and do it per product category rather than for the store as a whole. That single number sets the timing for most of what follows, and guessing it is how stores end up sending replenishment reminders a month after people already reordered somewhere else.

The order and delivery notifications are the most opened messages your store will ever send, and in many stores they are the default template nobody has looked at since launch. Fix them first, because improving a message everyone reads beats adding one most people ignore.

The order confirmation needs to answer the questions a customer has at that exact moment. What did I buy, what did it cost including delivery, when will it arrive, where is it going, and what do I do if that address is wrong. Put the delivery estimate on it. That single line prevents a large share of support messages, and its absence is why customers write to ask.

The dispatch notification needs the tracking link working at the moment you send it. A tracking number that returns nothing for six hours creates exactly the worry the message was meant to prevent, so delay the send until the carrier has the parcel scanned rather than sending it at the moment the label is printed.

Add the messages most stores skip. Out for delivery, delivered, and a note when something is delayed. The delay message in particular is worth more than any marketing you could send that week. A customer told about a delay before they notice it stays calm; one who finds out by waiting does not.

Make them look like your store rather than like your platform's default, and keep the transaction information first. It is reasonable to include a small relevant addition at the bottom, and it is not reasonable to push the tracking link below a promotion.

Finally, check what these messages say on a phone and check them in the language and currency the customer bought in. These are the emails your customers will judge your competence by, and they are read more carefully than anything you write on purpose.

The waiting period is dead time for the customer and free attention for you, so use it for something that improves the arrival rather than for a sales message. Anticipation is the mood here, and it is easy to work with.

What belongs in this window depends on your product. If it takes preparation, say so and show it, because the customer knowing a maker is working on their order changes how they feel about waiting. If there is something to prepare at their end, measurements to take, space to clear, an account to set up, a surface to check, tell them now while it is useful rather than after the parcel arrives.

A short message about the brand works well here and almost nowhere else. Who makes this, why you sell it, what you care about. A first time customer is more receptive to that in the days after paying than at any other point, because they have just chosen you and want to feel it was a good decision.

Keep the sending light. One or two messages across the wait, on top of the transactional notifications, is plenty. If your delivery takes a day, this whole section collapses into the dispatch notice. If it takes three weeks because you make to order, the wait needs more support and a clear expectation set at the start.

Avoid two things. Do not sell in this window, because a customer who bought yesterday and receives a promotion today feels like a target rather than a customer. And do not send anything that would make somebody worry, such as a policy heavy message about returns before they have even received the item.

When delivery slips, say so first and say what you are doing about it. Being told by the store rather than by an empty doorstep is the difference between a shrug and a complaint.

Time a message to arrival, because the first hour decides whether the purchase was a success and everything afterwards depends on it. Trigger it from the delivery event where your platform provides one, or from a realistic estimate where it does not.

What the message contains depends entirely on what you sell, and this is the point in the sequence where a generic template is most obviously wasted. Something that needs assembling gets the instructions and the common mistake people make on step three. Something worn gets fit and care advice. Something applied gets the amount to use and how long before it does anything. Something technical gets the setup route and the one setting people always miss.

Write it from your support inbox. The questions people ask in the first week after receiving your product are already recorded, and answering the top three before they are asked removes the contacts and improves the result at the same time.

This is also the moment to prevent the avoidable return. Most returns in a store fall into a few repeated causes, and a good share of them are misunderstanding rather than fault: the wrong size chosen, the product used incorrectly, an expectation about colour or scale that the photographs set badly. A message that catches the misunderstanding while the item is still in its box turns some of those returns into satisfied customers.

Keep it useful and short, and put the human route in it. A line saying how to reach a person, with a real answer time, costs nothing and it is what converts a problem into a fixed problem rather than into a review.

Where your product suits it, this is where a short video or a set of photographs earns its place, because showing is faster than describing for anything physical.

A customer who gets a good result from the first purchase buys again. One who never quite worked out how to use it does not, and they rarely tell you why. So spend the couple of weeks after arrival making the product succeed.

The content here is specific to what you sell and it is usually already written down somewhere in your business. How to get the most from it. What to do differently after the first few uses. How to look after it so it lasts. What people commonly combine it with. The mistake that stops it working properly.

Space it out rather than sending everything at once. Two or three messages over a fortnight, each covering one thing, will be read. A single long guide will not. And tie each one to where the customer is likely to be in their experience of the product rather than to a fixed schedule, if you know enough to do that.

This is the natural place for the accessory or the companion product, because now it is advice rather than a sale. The thing that makes the first product better, the refill that will be needed, the item that most customers buy second. Introduce it as part of getting a better result and let the customer decide, and it will outperform the same product presented as an offer.

User content works well in this section. What other customers do with the product, how they use it, the setups and combinations they have found. It shows the range of what is possible, and it also quietly signals that a community exists around what you sell.

One rule. Everything in this part of the sequence should be useful even if the customer never buys anything again. If a message only makes sense as a sales message, it belongs later.

Time the review request to when the customer has actually formed an opinion, which is later than most stores ask. A request sent on the day of delivery collects impressions of the packaging. A request sent after enough use collects a review about the product, and that is the review that helps the next buyer.

How long depends on the product. Something used immediately can be reviewed within days. Something whose value shows over weeks, a skincare product, a mattress, a tool, needs the corresponding wait. Set the delay from how long your product takes to prove itself, and check it against the reviews you already have: if existing reviews mostly describe delivery rather than the item, you are asking too early.

Ask for one thing and make it easy. A single link, a rating that can be given in one tap, and the option to write more for those who want to. Every extra step loses reviewers, and the people you lose first are the moderately happy ones whose reviews are the most useful and most credible.

Ask about the specific product they bought, by name and with its image. A generic request to review your store collects fewer and vaguer reviews than a request about the item in their hand.

Do not filter for positive reviews. Routing unhappy customers to a private form while sending happy ones to the public one is a practice that damages trust and can breach the rules of review platforms. Ask everybody the same question. If the answer is bad, you have found a problem worth knowing about.

Do not pay for reviews. Incentives change what people write and, in many places, must be disclosed. If you want more reviews, ask more people and make it easier rather than buying them.

Follow up once if there is no response, then stop.

Invite the complaint, because a problem you hear about is a problem you can fix, and a customer whose problem was fixed well is more loyal than one who never had a problem at all. A customer who could not find a way to tell you writes a review instead.

Build an explicit check in a few days after arrival. A short message asking whether everything arrived as expected, with a direct reply route, catches the damaged item, the wrong size, the missing part and the misunderstanding while it is all still easy to resolve.

Make the returns route clear rather than hidden. Stores sometimes bury this in the hope that fewer people will use it, which produces more frustration, more support work and worse reviews. A customer who can see an easy return is more likely to buy in the first place, and a portion of them will exchange rather than refund if the exchange is the easier option.

Watch for the signals that something went wrong without anybody saying so. Tracking that stalls, a delivery marked complete that the customer says never arrived, an order with several attempted deliveries. Where you can see these, a proactive message beats waiting.

When something has gone wrong, fix it before you sell anything else to that person. Any promotional message sitting in a sequence should pause for a customer with an open support case or an active return, because a discount arriving in the middle of a complaint reads as though nobody is paying attention.

Finally, read what comes back in aggregate. The reasons customers give for returns and complaints, sorted by frequency, are the most direct list of product page fixes you will ever have. Most of them point at a photograph, a size chart or a sentence that promised something the product does not do.

The ask goes where the customer is ready, which is a date derived from your own data rather than a round number of days. For consumables, that is shortly before they run out. For everything else, it is when the next relevant need appears.

For replenishment, work out how long the product lasts in normal use and send the reminder with enough time to arrive before the customer runs out. Sending on the day they finish is too late, because they have already bought something else. If your product has variable usage, ask at purchase or use the customer's own reorder history once you have one, and get more accurate with every order.

Make reordering trivial. One tap to buy the same thing again, with the size, quantity and variant already chosen. Every additional decision costs you orders on a purchase the customer had already decided to make.

Offer a subscription where the product genuinely repeats, and present it at the moment the second order is due rather than at the first purchase. A customer who has used the product and liked it understands the offer. A first time buyer does not yet know if they will want it again.

For products that do not repeat, the second order comes from a different direction. The accessory, the companion item, the next size up, the same product for someone else, or the seasonal version. Base the recommendation on what other customers who bought that item actually bought next, which your order data can tell you, rather than on what you would like to sell.

And choose the moment carefully. A single well timed message beats a monthly nudge, and it costs you far less in unsubscribes. If the reorder point is four months away, be quiet for four months and stay useful in between.

Split the flow by customer history, because the same messages are wrong for both groups. A first time buyer needs reassurance and an introduction. A fourth time buyer being introduced to your brand story will notice that nobody is paying attention.

For a first order, spend the sequence on trust and on making the product succeed. Who you are, how to reach you, how to use the thing, and a light introduction to the rest of the range. The goal is a good first experience, and the second order follows from that rather than from being asked harder.

For a repeat customer, cut most of it. They know who you are. Keep the transactional messages, keep anything genuinely specific to the product they just bought, and move faster to the useful part. Consider thanking them for coming back in a way that is not a discount, since recognition costs nothing and is rarer than money off.

Split by product too. Somebody who bought a consumable enters the replenishment path. Somebody who bought a durable item enters the accessory and care path. If your catalogue has genuinely different categories, this split matters more than the customer history one, because the useful content is entirely different.

Handle the gift case if it applies to you. When the delivery address differs from the billing address, or the order carried a gift message, the buyer may never use the product. Usage instructions aimed at them will be irrelevant, and a review request may be impossible to answer. Softening the sequence for these orders takes one condition and avoids a lot of awkward messages.

Keep the number of variants manageable. Two or three paths that are genuinely different beat eight that differ by a sentence, and the maintenance burden of the eight is what causes flows to go stale.

The outcome is repeat purchase, so measure repeat purchase. Open rates on post purchase messages are high in every store because people are waiting for a parcel, which makes them useless as a signal of whether the flow is working.

Track three business numbers. The share of customers who place a second order, measured on cohorts by the month they first bought. The median gap between first and second order, which should shorten if the flow is doing its job. And the value of a customer over their first year, which is the number that justifies what you can afford to spend acquiring them.

Use cohorts rather than an overall figure. Repeat rate for the whole store moves whenever acquisition volume changes, which makes it almost impossible to read. Comparing the customers who first bought in one month against those who first bought in another, at the same number of days after their first order, is the only clean comparison.

Alongside those, watch the operational effects. Support contacts per order should fall as the transactional and arrival messages improve. Return rate should fall where the flow is preventing misunderstandings. Review volume should rise. Each of these has a value that never appears in a marketing report.

Attribute carefully. Almost every second order will be credited to something, and the post purchase flow, your advertising retargeting and your campaigns will all claim the same order. The cleanest read is a holdout: hold a portion of new customers out of the flow and compare their repeat rate against the rest at the same age.

Give it time. This is the slowest flow in the account to prove itself, because the outcome you care about happens one reorder cycle later. Set the review date accordingly, and resist judging it on the first month.