Follow Up With Existing Customers
Your best buyers are the ones who already paid you, and yet you never hear from them again after delivery. This sequence first asks how things went with what they bought, and only that answer decides between a reorder, an add-on offer or help. Here is why the question comes before the offer, why an unhappy customer gets nothing to buy, and how to pick the right gap after the first purchase.
Why your buyers disappear after delivery
The sale is done, the parcel is out, the invoice is paid. After that nothing happens, except that your customer gets the same monthly newsletter as every prospect who never spent a cent. An analysis of more than 1,500 Shopify stores shows how reliably that leads nowhere: the average repeat purchase rate sits at 25 percent, and the average customer orders 1.8 times. Three out of four people who once trusted you with their money never come back.
Replacing those three gets more expensive every year. Contentsquare analysed 90 billion sessions across 6,000 websites: the cost per website visit rose 9 percent in 2024 and 19 percent over two years, while the conversion rate fell by 6.1 percent in the same period. So you pay more for visitors who buy less, and meanwhile you leave alone exactly the people who already know you, already trusted you and already made it through your checkout once.
The usual answer is a blanket email to everyone with a discount code. It arrives on a Tuesday because it is Tuesday, not because anything happened on the recipient’s side. The difference is measurable: across 23 billion emails analysed, automated messages triggered by an event convert at 1.9 percent, classic campaigns to the list at 0.07 percent. The lever is not the discount, it is the occasion.
And you already have an occasion: the purchase itself. Your shop or your CRM knows who received what and when. That signal becomes a starting point, behind it a waiting period based on how long the product lasts, then a single message with one question and three answer options. Depending on what your customer taps, it continues into a reorder, into a smaller add-on offer, or into help.
Putting the question before the offer is the whole trick. An offer without a question quietly assumes the first purchase went well. For one in four it did not, and that person now reads a sales message from someone who clearly has no idea how things went for them. The question costs your customer one tap and gives you the cleanest split you can get: on one side someone who is happy and therefore open to buying at all, on the other someone with a problem. People with a problem do not buy. They complain, or they go quiet and are gone.
How the sequence is built
The signal from your shop or CRM is the trigger
The sequence does not start on a date, it starts on an event: order delivered, membership active for four weeks, course unlocked. The blueprint marks where that signal comes in and defines which details have to travel with it for the later message to make any sense at all: which product, which date, which name. Without that connection every follow-up stays a mass mailing with a different sender, because only the event turns “sometime soon” into a specific moment for this specific customer.
Derive the gap from the product, not from the calendar
The waiting node behind it holds the most important number in the whole sequence, and you calculate it from consumption rather than from gut feel. For a consumable you work backwards from the pack: 60 capsules at two a day last 30 days, so you ask around day 20 to 24. For a membership you ask at the point where the first routine has either taken hold or quietly stopped. For a course, just past the point where most people drop off. The rule behind it is always the same: early enough that your customer has not reordered elsewhere, late enough that they actually have an opinion. Ask too early and you get “no idea yet” and have spent your one good occasion. Ask too late and the decision was made somewhere else long ago.
The first message asks, it does not sell
Days or weeks have passed since the purchase, so the free reply window is closed. That makes the first message an approved template, and it asks exactly one thing: how it went with what they bought. That question can be answered without committing to anything, which is precisely why you get a reply, and the reply opens the free window. The channel works in your favour here: in a Kantar survey of 11,056 adults across 22 markets, 72.4 percent said they are more likely to buy from a brand they can reach by message. A message that can be answered is already part of the offer.
One question, three answers, each going somewhere else
The decision node offers three options: it is working well, I am still figuring it out, it did not work for me. More than three will not fit on a phone screen, and more is not needed, because these three are mutually exclusive and cover the field between them. The middle option is the one everybody forgets, and it is what keeps the sequence honest. Without it, anyone who is unsure taps “working well” because that is the politer answer, lands in the sales branch and still buys nothing. Afterwards you cannot tell whether your offer was weak or your sorting was.
The help branch gets no offer
Anyone who taps “still figuring it out” or “did not work for me” gets no call to buy, but the one document or short video that removes the most common cause. Behind it sits a two-option question on whether that helped, and the result sets a tag. This is not politeness, it is arithmetic. A sales message to someone with an unsolved problem sells nothing and costs you the channel on top, because in a messenger your customer does not reply by ignoring you, they reply by blocking you. Someone who gets help here is, two weeks later, someone who likes the product, and then the offer carries. The blueprint draws in that delayed re-entry.
The happy branch asks exactly one follow-up question
Whoever is happy gets three named routes: the same again, the smaller thing that goes with it, not right now. You decide those three offers in advance and put them on the buttons, and your customer picks the one that interests them. That is deliberately a self-report rather than something derived from order data, because the customer knows better than your database whether they need a refill or want something else entirely. “Not right now” is not a lost contact, it is a tag and a later attempt. The blueprint places your order or checkout link directly behind the choice, because that is where the decision has already been made and every further step only reopens it.
Silence gets exactly one second attempt
A branch of its own catches everyone who does not react. A few days later a second message goes out, and it does not repeat the first one, it changes the angle: instead of asking about their experience it offers something concretely useful even without a reply. After that it stops, and the contact gets a tag instead of a third message. Two attempts is the point where persistence turns into nuisance, and in a messenger that point arrives sooner than in an inbox, because your customer sees your message on the same screen as the ones from their family.
Example scenario: a shop selling a consumable
A shop sells a product where one pack lasts around four weeks. After delivery nothing happens, except that the buyer drops into the same monthly mailing list as every prospect.
Once a month a blanket email with a discount code goes to the whole list. Whoever happens to need a refill right then might order, everyone else skims past it. Whoever did not get on with the first pack receives the same code as everyone else, stops replying altogether and sends a return three weeks later. Nobody finds out why it did not work, because nobody ever asked. In hindsight the team only sees that someone did not come back, and cannot tell whether that was the product, the timing, or simply the customer forgetting the brand existed.
The shop reports the delivered order, and the sequence waits until roughly three quarters of the pack should be used up. Then one message goes out with one question and three answers. Whoever taps “working well” gets the follow-up question and the order link at exactly the moment they were already thinking about a refill. Whoever is still unsure gets the usage guide, and the buying question only comes two weeks later. Whoever says it did not work lands with a reason and a tag in front of a human being instead of in the next discount email. A mailing list turns into a process that runs at a different moment for every customer. The team rebuilds the core of it in under six hours in its own messenger tool.
What you get
The blueprint with all three branches
The full sequence laid out visually: trigger, waiting period, the question about their experience, the three branches, the delayed re-entry out of the help branch and the purchase goal. Each node says what it is there for and which information it takes in or hands out. The wording is yours, the structure behind it is settled.
The timing as a rule you can recalculate
Where the gap after the first purchase comes from for your product: pack size and consumption, the run time of a membership, or the point in a course where most people stall. Plus the gap before the second attempt when someone stays silent. It is written as a rule you can adjust when your range changes or you notice you are asking too early.
The branching logic including the help branch
Which answer leads into which branch, what happens in the help branch instead of an offer, and where the offer does reappear there later on. Plus the three named follow-up routes for the happy branch and the tags that later tell you which answer a contact came from.
A buyer persona with over 50 data points
Buying motives, objections and reasons for hesitation in your market. That is what tells you which add-on offer genuinely makes sense after a first purchase instead of merely costing more, and what dissatisfaction actually hangs on in your field. It decides whether the document in the help branch solves a problem or just keeps someone busy.
The starting list for the build
The event from your shop or CRM with the fields that need to come with it, the three follow-up offers spelled out, the help asset and the place for your order or checkout link. You know what needs to be on the table before the first click, not once half the flow already stands.
More funnel blueprints
Bring Old Contacts Back
One question about the reason revives stalled conversations and splits them into three paths: too early, slipped through, no longer needed.
View blueprintWin Back Abandoned Carts
A question about the reason instead of a reminder: someone who abandoned needs no convincing, just an answer to their specific objection.
View blueprintAsk your buyers before you sell to them
The follow-up sequence arrives fully drawn: the structure, the branching logic for all three answers, the timing as a rule you can recalculate, your buyer persona and the asset list. You write the messages yourself, and the core takes under six hours to rebuild in your messenger tool. One-time payment, no subscription.
Build my funnel blueprintOne-time payment, no subscription. You get the complete strategy and overview first.