What the flow is for
The second order is the hardest one to earn. The repeat purchase and loyalty guide explains why: the first order is an experiment, and the second is a decision made once the customer knows whether delivery arrived, whether the product matched the page and whether it was easy to use. A post-purchase flow works on the last two. It helps the customer succeed with what they bought, asks for an opinion once they have one, and points to the next purchase when it becomes useful. The email channel page for retention puts it as onboarding the order, not just the account: help that arrives after delivery does more for the second order than a discount sent the day after checkout. When the product needs a visit, an installation or preparation by the customer before it delivers anything, the tutorial on onboarding new customers beyond software designs the steps the emails support.
Every message has a job, a trigger and a stream. The table shows a typical shape, with the order emails the shop already sends kept in their place and the flow's messages fitted around them.
| Message | Stream | Trigger | Job |
|---|---|---|---|
| Order confirmation | Service, existing | Order placed | Confirm what was bought, the amount and what happens next |
| Dispatch notice | Service, existing | Order dispatched | Tracking and delivery expectation |
| Delivery notice | Service, existing | Order delivered | Confirm arrival; how to get help |
| Getting started | Service or marketing, decided in step three | A wait after delivery, per product branch | Set-up, care and first-use help for what was bought |
| Review request | Decided in step three; marketing where unclear | A longer wait after delivery, per branch | Ask for an opinion once the customer has one |
| Next purchase | Marketing | When the complementary item becomes useful | Suggest the item that follows from what was bought |
The flow's measure is the second-order rate, the repeat purchase rateShare of customers who placed more than one order in the period. of first-time buyers within a window matched to the category's natural cycle, read against a holdout. Clicks and review counts are diagnostics.
Before you start
- A working welcome series. The welcome series tutorial establishes the pattern of trigger, spacing, exits and tests reused here, and its FAQ already says a post-purchase sequence should take over from people who buy during the series.
- Order events in the sending platform, as they happen: order placed, dispatched, delivered, returned and refunded, each with the products, categories, quantities and values involved, and the customer's identifier. Delivery is the event most shops lack; find out whether the shop or carrier integration can send it.
- A list of every service email the shop already sends, from which system, with its timing and content.
- A consent position, decided with whoever owns compliance, for the messages that will be treated as marketing.
- Help content for the main products or categories: set-up, care, common mistakes and how to get help. Support tickets and return reasons show where customers struggle.
- An order-data view of what customers buy next, and how long after, for the main first-purchase products.
- A named owner and a review date, as the automation hub asks of every flow.
The steps
Set the goal and who enters
Write the goal in one sentence: a successful first use and a second order within a window that fits the category's cycle. The goal decides the metric, the second-order rate of entrants within that window, and the metric decides the holdout in the last step. Every order enters, but not every customer receives the same sequence. Branch first-time buyers from repeat customers: a repeat customer does not need the introduction to the brand, and a customer who has bought the same product before does not need its set-up guide again. Decide what happens to orders sent to another address as a gift, since the buyer will never use the product and the recipient never agreed to hear from the shop. Read the customer's recorded basis for marketing at entry, so that the marketing messages later in the flow know whether they may send.
Map the order emails the customer already receives
List every message the customer receives between placing the order and using the product: the confirmation, any payment or account messages, dispatch, delay notices, delivery, and return or refund messages, with the system that sends each one and when. Lay them on a short journey map from order to first use; the customer journey mapping framework adds lanes for what the customer does, asks and feels at each point, and the questions in support tickets fill those lanes quickly. The rule for the new flow follows from the map: it never duplicates, delays or competes with an order email. No flow message arrives before the order confirmation, none repeats the tracking details the dispatch notice carries, and none lands on the same day as a delay notice.
Decide which messages are service and which are marketing
The promotional versus transactional comparison describes the UK regulator's direct marketing guidance as separating service messages a customer needs from marketing a business wants to send, with the content and purpose of each message deciding which it is, not the label on it. Its test is practical: remove the promotional content and ask whether the customer would still need what is left. Order, dispatch and delivery emails are service messages and stay free of marketing. A suggestion for the next purchase is marketing. Help emails and the review request depend on what they contain; decide each with whoever owns compliance, and where the answer is unclear, treat the message as marketing. Marketing messages need the customer's recorded basis, which the list building and consent guide explains, including the existing-customer conditions the UK guidance describes, and an unsubscribe in every message. Rules differ by market and nothing here is legal advice.
Trigger each message from the order's real progress
Trigger from what has happened to the order, not from a fixed number of days after it was placed. A help email timed from the order date arrives before the parcel when delivery is slow, and a review request timed the same way asks about a product still in a depot. Use the delivery event to start the clock for help and the review request, and the dispatch event for anything that belongs before arrival. Decide what happens when no delivery event comes, as it will not for some carriers: wait for an estimated delivery date from the shop, or hold the message, rather than guessing. For split shipments, trigger from the delivery of the item the message is about. Set sending windows in the customer's local time, as the abandoned cart flows tutorial does, so that nothing arrives overnight.
Write help that matches what was bought
Branch the help email by product or category, because a customer who bought one thing and a customer who bought another need different help, as the retention email page says. For each branch, write the few things that decide whether the first use goes well: how to set it up, how to care for it, the mistake new owners make, and where to get help, with a reply address that reaches a person. Take the content from support tickets and return reasons rather than from the product page, which was written to sell. Keep a help email that was classified as a service message free of offers, or it changes category. Start with the branches that cover most first orders and a general version for everything else, and add branches where the return reasons show a product that customers get wrong. Customers who cannot be emailed still receive the parcel, and the guide to packaging and order inserts shows how a first-use card can carry the same help to every buyer.
Time the review request after use
The reviews and social proof guide sets the rules this step follows. Ask after the customer has used the product long enough to have something to say, which is later than most stores send it and varies by product, so set the wait per branch from the delivery event. Keep the request short and specific to the order: name the product, show its image and ask one question that invites detail. Send one reminder and then stop. Ask every customer, not only those expected to be pleased, and never put the request in the same message as a discount or a loyalty prompt. Any incentive for reviewing must not depend on what the customer writes, as that guide sets out. Skip the request when the customer has already reviewed, returned the item or has an open complaint. Referral invitations are a separate ask with their own timing, and the guide to loyalty and advocacy email covers when to send them and how to measure them against a holdout.
Suggest the next purchase from your own order data
The suggestion should be the item customers who bought this product actually bought next, at the point it becomes useful, which the shop's own order data shows. The repeat purchase guide describes the common paths, such as a consumable for a durable, completing a set, or a cross-category step, and the cross-selling entry explains how pairs are chosen from what customers buy together and judged on margin rather than on clicks. Cross-sellingOffering a customer a related product alongside the one they are buying; upselling offers a higher-value version instead. of this kind is the main lever the flow has on average order valueAverage revenue per order. and on lifetime valueTotal value a customer brings over time., so keep one suggestion per branch rather than a block of bestsellers. For consumables, hand the customer to the replenishment flowAutomated messages that remind a customer to re-order a consumable, timed to the product's re-order cycle from the store's own data, not a fixed delay. timed from the re-order cycle instead of rebuilding that timing here. Any discount for the second order is a policy read from the customer's record, not a line in the template.
Set exits for re-orders, returns and complaints
Set the exits and make sure the platform checks them before each send, not only at entry, as the automation hub explains. A new order ends the next-purchase message, whether or not it contained the suggested item, and starts a new run of the flow for the new order. A return or refund ends the review request and the suggestion; a customer sending something back does not want either. An open support ticket or a complaint pauses the flow until it is resolved. The tutorial on pausing marketing to customers with an open issue turns that pause into a hold on the contact that every flow, campaign and ad audience checks. An unsubscribe stops the marketing messages and never the service ones. Then set priorities beside the other flows: a customer who buys during the welcome series moves to this flow, the cart flow exits on the order, and a cap on marketing messages per person per week applies across flows and campaigns.
Test the flow end to end
Run the email automation flow launch checklist against the flow, then place test orders that exercise every branch: a first-time and a repeat customer, each main product branch, a slow delivery, a split shipment, an order with no delivery event, a return before the review request, a support ticket mid-sequence and a re-order before the suggestion. For each, check that every message arrives after the order email it follows, with the right product content, personalisation that falls back cleanly, tagged links and a working unsubscribe that leaves the service messages untouched. Confirm that a gift order follows the rule set in the first step.
Hold out a share and review
Hold out a random share of first-time buyers at entry: they receive every service message and none of the flow's marketing messages. Compare the second-order rate per entrant in the two groups over the same window. The difference, per entrant, is what the flow adds, which its own report cannot show because many customers would have ordered again anyway; the email metrics and testing guide explains how to size a holdout groupA randomly chosen part of an audience kept from exposure to measure what exposure caused. and read it. Read the result by monthly cohort of first-time buyers, as the cohort analysis tutorial builds it, and use the conversion rate calculator for the rates. Review each branch's clicks, unsubscribes and the reviews collected monthly, and the holdout quarterly.
A pet bed maker's post-purchase flow (illustrative)
Snugwhistle Pet Beds measures its flow on second orders
Snugwhistle Pet Beds, a maker of washable pet beds, used to send one review request two days after dispatch, which often arrived before the bed did. It rebuilt the flow in four messages, triggered from the carrier's delivery event and branched by bed type: a care email three days after delivery on washing and drying that bed's cover, a "settling in" email at day ten on helping a pet take to a new bed, a review request at day twenty-one, and a suggestion of a spare washable cover at day forty-five, when the first cover has usually been washed several times. Over six months, 4,000 first-time buyers ordered; 400, chosen at random, were held out of the four messages and received only the order, dispatch and delivery emails.
| Group | First-time buyers | Second orders within 120 days | Second-order rate | Of which spare covers |
|---|---|---|---|---|
| Flow | 3,600 | 353 | 9.8% | 191 |
| Holdout | 400 | 28 | 7.0% | 9 |
The flow added about 28 second orders per thousand first-time buyers, most of them spare covers. Reviews per hundred orders roughly doubled once the request waited for use. The settling-in email drew the most replies, which the team now reads for new product questions.
The conversion rate calculator for email flows enters entrants and second orders for the flow and its holdout, then reads the gap between the two rates as the flow's contribution.
Verify it worked
- Read the journey map beside the flow. No flow message should arrive before the order confirmation, repeat the dispatch tracking details or land on the same day as a delay notice.
- Follow a test order through the platform. Help and review messages should start from the delivery event, not the order date.
- Open each help branch. The content should match what was bought, and any help email classed as service should carry no offer.
- Trigger a return, a complaint and a re-order on test orders. Each should end or pause the right messages before the next send, and an unsubscribe should leave the service messages untouched.
- Compare the holdout with the flow. Held-out first-time buyers should receive the service messages only, and the second-order rate per entrant should be readable for both groups.
If it goes wrong
- The help email arrives before the parcel. The likely cause is timing from the order date, or a missing delivery event. Start the clock from delivery. Where no event arrives, wait for the shop's estimated delivery date or hold the message.
- A customer who returned an item gets a review request. The likely cause is that exits were checked only at entry. Make the platform check them before each send.
- Opted-out customers stop receiving delivery notices. The likely cause is service and marketing messages sharing one unsubscribe. Separate the streams so that opting out stops marketing only.
- Help emails get ignored. The likely cause is one sequence for every product, or content taken from the product page. Branch by product and rewrite from support tickets and return reasons.
- The flow's report looks strong but the holdout shows little difference. The likely cause is credit taken for second orders that would have happened anyway. Judge the flow on the gap between the two groups, and check any second-order discount against the holdout.
Next steps
- Read the guide to e-commerce email and SMS flows to see where this flow sits among the shop's other flows.
- Read the abandoned cart flows tutorial to see how its incentive policy and local-time sending windows work beside this flow.
- Read the email channel page for retention to see why help after delivery is framed as onboarding the order.
Common mistakes
- One review request sent on dispatch. It asks for an opinion about a product still in transit.
- Timing from the order date. Slow or split deliveries put the help email ahead of the parcel.
- Marketing inside the order emails. A prominent offer turns a service message into a promotional one.
- The same sequence for every product. Help that does not match what was bought is ignored.
- Bestseller blocks instead of the next purchase. Use the item customers actually buy next, from order data.
- No exit on returns or complaints. A customer sending a bed back receives a request to review it.
- Unsubscribing people from service messages. Opting out of marketing must never stop a delivery notice.
- Reading the flow's own report as its effect. Without a holdout, the flow takes credit for second orders that would have happened anyway.
When the flow is live, the automation hub sets out the governance it needs: an owner, a review date and change control for a flow that runs while nobody watches.
