How to Automate Review Requests After Purchase

The reliable way to get more customer reviews is to ask every customer, automatically, a few days after the product is in their hands or the job is finished, with a short message from a real person's name that asks how things are going before it asks for anything. Timing does more work than wording: the request has to land while your business is still the most recent thing in the customer's memory, and no human remembers to send it at scale, which is why the ask belongs to software. Send it once, follow up once, route the replies to a person, and never filter who gets asked by how happy you think they are. Do that for a quarter and the review page starts to reflect the customers you actually have instead of the loud minority who write unprompted.

The short version

Satisfaction is a state; a review is an act. The gap between them is a prompt, and almost nobody crosses it without one.
Timing is mechanical. Trigger on delivery or job completion, never on the order, and wait long enough for the thing to be used.
Ask a question before you ask for a review. "How is it working out?" earns a reply, and the reply is the natural moment for the request.
One platform, one link, one tap. Every extra choice between the message and the form loses people.
Ask everyone, incentivize nobody. US rules prohibit steering only happy customers toward public reviews and paying for positive ones.
Buy the app when your trigger lives in your store platform; build when it lives anywhere else.

Why don't happy customers leave reviews on their own?

In 1962, Bennet Murdock read lists of words to people and asked them to recall as many as they could. The result, published in the Journal of Experimental Psychology, became one of the most replicated findings in memory research: the last items on the list are recalled best, the first items hold on reasonably well, and the middle collapses. Psychologists call it the serial position effect. Your customer's month is a list of that kind, and your delivery is one item on it. For a day or two it is the most recent thing that happened to them. Then a school pickup, a dentist appointment and forty emails push it into the middle, where it is still true but no longer reachable.

That is the mechanical reason the review page stays quiet on top of a pile of genuinely happy customers. Being pleased with a purchase is a feeling that fades on the recall curve like anything else. Writing a review is an action that needs a trigger, a device in hand and about ninety seconds of attention, all at the same moment. The customers who supply their own trigger are the ones with a grievance, because irritation renews itself every time they look at the thing that annoyed them. The happy majority needs the trigger supplied from outside, and "we'll ask when things calm down" is a plan to supply it after the memory has already slid down the list.

There is a second reason worth naming. Owners feel that asking is pushy, so they wait for reviews to happen organically, then read the sparse, skewed result as a verdict on the business rather than on the process. It is the same misunderstanding that keeps most referral programs from working: people are glad to help, they just do not think of it unprompted.

When is the right time to ask for a review?

The right moment is a small window that opens after the product has been used or the job has been experienced, and closes when the memory ages out. Hitting it consistently is a matter of picking the right trigger event and the right delay for each thing you sell.

Trigger on delivery or completion, never on the order. The order confirmation, the shipping notice and the payment are all events on your side. The customer's experience begins when the box is opened or the technician leaves. A carrier's delivered scan, a fulfillment status change or a job closed in your scheduling system is the event to build on.
Delay by product type. Consumables and simple goods: two or three days after delivery. Furniture, equipment and anything that needs assembly or setup: about a week. Services performed in person, such as repairs, cleaning, installs and treatments: that evening or the next morning, while the result is visible. Software and subscriptions: after the first moment of real use, not after signup.
Pick a humane send window. A weekday morning or early evening in the customer's time zone. Nothing after 9pm, nothing on Monday before coffee.
Follow up once, then stop. If there is no click and no reply after five to seven days, one gentle second message is fine. A third is spam.
Suppress when the timing would be cruel. Never ask before the product has arrived, during an open support conversation, after a refund or return, or from a customer who was asked within the last few months.

What should the review request say?

Less than you think, and in a different order than most templates suggest. The message that performs is a short note from a named person that asks a question first.

The two-step works because a question invites a reply and a reply creates a conversation, and inside a conversation the request for a review is a natural next line rather than a cold ask. The first message: "Hi Dana, this is Marcus from Northline. The standing desk should have arrived Tuesday. How is it holding up so far?" A customer who replies "Love it, the motor is quieter than expected" has just written half a review, and "Would you mind putting that on our Google page? Here is the direct link" is an easy follow-through that a human can send in ten seconds. A customer who replies with a problem has just been routed to support before they went public with it, which is the second-best outcome and sometimes the best.

A few rules for the wording itself:

A person's name, not the brand. "Marcus from Northline" gets replies; "The Northline Team" gets archived. The reply address must be monitored by someone who answers within a business day.
One platform, one link. Decide where reviews matter most for you and send everyone there. A message offering Google, Facebook and Yelp asks the customer to make a decision, and decisions are where people stop.
Deep-link to the form. The Google link that opens the write-a-review box, the product review form on your own store, not your homepage.
Under eighty words. The request is not the place to describe your values.
Say what it does for you. "Reviews are how small businesses like ours get found" is honest and it works.

Two constraints are legal rather than stylistic. The Federal Trade Commission's consumer review rule, in force since late 2024, bans buying reviews whose sentiment is conditioned on the payment, suppressing negative reviews through threats, and presenting a filtered set of reviews as though it were the whole picture. Google's own contribution policy goes further and prohibits offering incentives for reviews at all and selectively soliciting only satisfied customers, a practice known as review gating. The safe design is simple: every eligible customer gets the same ask, nobody is paid, and the "are you happy? yes / no" screen that routes unhappy people away from the public form does not exist in your flow. Separately, texting customers in the US requires their prior consent, so the SMS version of the ask belongs only where they opted in.

How do you automate review requests, step by step?

Choose the one platform that matters. For a local service business that is almost always Google. For an online store it is usually the product reviews on your own product pages, because that is where the next buyer is standing. Add a second destination only after the first is producing.
Find where the trigger event lives. For shipped goods it is the carrier's delivered status, which your store platform or shipping app can expose as a webhook. For services it is a job marked complete in your scheduling or field-service tool. For subscriptions it is a usage milestone. If the word is unfamiliar, webhooks for business owners explains it in five minutes.
Set the delay as a rule per product or service type. Store it as data the store manager can edit, so "desks: 7 days" changes without a developer.
Write the suppression rules. Open support tickets, refunds or returns in progress, unsubscribed contacts, wholesale accounts, customers asked in the last ninety days, and orders flagged as problem deliveries. These rules are what make automatic asking safe.
Write two messages and name the sender. The question-first note, and the single follow-up. Decide who the human sender is and make sure their inbox receives the replies.
Choose the channel by where the customer already talks to you. Email is the default and needs no consent beyond an unsubscribe link. SMS gets read faster but only where the customer opted in. Where customers already message you on WhatsApp, the ask can ride that thread, using the approach in our WhatsApp automation guide.
Route replies to a person and give them a script. The reply is where the actual review request happens, so it needs a human with the deep link ready and a same-day standard.
Measure four numbers weekly. Asks sent, reply rate, reviews per hundred asks, and the average rating over time. A falling reply rate usually means a deliverability problem; a rising share of complaints usually means the delay is too short.
Schedule the job and give it a heartbeat. The flow is a scheduled task that looks for eligible orders and sends; the pattern for running one dependably is in scheduled jobs on Vercel. Have it post a daily count somewhere visible so a silent failure gets noticed in a day rather than a quarter.

Before switching it on, check the flow against this list:

Trigger is delivery or completion, not order placed or shipped.
Delay matches the product, and is editable without a developer.
Suppression rules exist for tickets, refunds, unsubscribes and recent asks.
Sender is a named person with a monitored reply address.
One deep link to one platform.
Consent respected per channel, with an unsubscribe on every email and opt-in on every text.
Replies land with a human who has the link ready.
Every public review gets a response, positive and negative alike.
No gating and no incentives anywhere in the path.

Should you buy a review app or build the flow?

If your trigger is an order in Shopify or WooCommerce and your destination is the product page on that same store, buy an app. Judge.me, Yotpo, Okendo, Stamped and their peers do exactly this, with delivery-based timing, templates and the widget that displays the result, and building it yourself would be paying to reinvent a solved problem. We tell owners that plainly, and it ends a good share of these conversations in ten minutes.

Building becomes the right answer when the trigger or the destination falls outside what those apps see:

The trigger lives outside the store. A field-service job, a B2B delivery confirmed by a driver, a booking system, a custom CRM. No storefront app can watch those events.
The destination is Google or an industry site, per location. A three-location service business needs each customer sent to the right place's link, which is routing logic the apps rarely handle.
You want the two-step conversation. Question first, human reply routing, then the ask. Apps send a template with a link; the conversational version needs a small inbox and some state.
The review flow should feed your other sequences. A customer who replied warmly is a referral candidate; one who replied with a problem should pause the marketing emails. That coordination belongs in the same place as your customer onboarding automation, not in a widget.

When we build it, the first phase is small on purpose: one trigger, one channel, one platform, the suppression rules, the human reply inbox and a page showing the four weekly numbers. It ships in two to three weeks at a fixed price, and it is designed to be judged on those numbers before anyone talks about a second channel or per-location routing. Owners who want the honest version of "does this fit us" can start from our services overview, and the honest version is sometimes "install the app."

Common pitfalls

Most broken review flows are broken by timing, not by wording, and the mistakes repeat across industries.

A case, details changed. A home-services company with a strong reputation and a thin review page had automated the ask years earlier, and technically it worked: every customer received a request. The trigger was "invoice paid," which felt natural to the office, but the company invoiced monthly, so the request arrived four to six weeks after the technician had left. By then the job had slid into the middle of the customer's list. The people who still remembered it vividly were the ones with a complaint, and the review page reflected that. Moving the trigger to "job closed" with a one-day delay changed nothing about the work and everything about the response: more replies, more reviews, and a rating that finally matched what customers said on the phone.

Other patterns to check against your own flow:

Asking on "shipped." The customer in Montana gets the request three days before the box, and the request reads as either careless or dishonest.
Same-day asks for things that need a week. Nobody has an opinion about a mattress on the day it arrives.
Asking during an open ticket. The fastest way to convert a fixable complaint into a public one.
Asking on every order. A weekly reorder customer asked weekly stops reading your email altogether.
Sending from noreply@. You have asked a question and blocked the answer.
Gating. Beyond the legal exposure, platforms detect the skewed pattern and reviews get removed.
Ignoring the replies you generated. A customer who answers "how is it going?" and hears nothing back is worse off than one who was never asked.
Silent death. The job stopped in March and nobody noticed until someone wondered in August why the count had flatlined.

FAQ

How long after a purchase should you ask for a review?

Count from delivery or job completion rather than from the order, then wait long enough for real use: two to three days for simple physical goods, about a week for anything that needs assembly or settling in, and the same evening or next morning for services performed in person. Send one follow-up five to seven days later if there was no response, and stop after that.

Is it legal to ask customers for reviews in the US?

Yes, as long as you ask honestly. The FTC's consumer review rule targets fake reviews, payment conditioned on a positive rating, suppression of negative reviews and misrepresenting a filtered set as complete; a plain, uniform request to every customer breaks none of those. Platform policies add their own limits: Google forbids incentives and selective solicitation, and Yelp discourages asking at all. Text messages additionally need the customer's prior consent.

Can I offer a discount in exchange for a review?

Avoid it. A reward conditioned on a positive review is prohibited by the FTC rule, and Google and Amazon prohibit incentivized reviews regardless of sentiment, so the practical risk is removed reviews or a suspended listing on top of the legal exposure. If you want to thank customers, thank them after the fact without having promised anything, and keep the thank-you disconnected from what they wrote.

Should I send review requests by email or text?

Email by default, because it needs no special consent beyond an unsubscribe link and carries a clickable deep link well. Text messages are read faster and work well for local services, but in the US they require prior express consent, so use them only for customers who opted in at booking or checkout. Where customers already message you on WhatsApp, continuing that thread is the most natural channel of all.

Which review platform should I send customers to?

The one where your next customer is deciding. For local and service businesses that is Google, because the rating shows up in maps and search before anyone visits your site. For online stores it is usually the product reviews on your own product pages, which influence the purchase at the moment it happens. Pick a single destination per business, put one link in the message, and only add a second platform once the first is producing steadily.

What happens if the automated request reaches an unhappy customer?

That is the flow working, provided the ask was a question rather than a bare link. An unhappy customer who replies to "how is it going?" has given you a private chance to fix the problem before it becomes a public rating, which is why replies must land with a person who can act. The suppression rules keep the ask away from customers with open tickets or refunds, and the rest is the ordinary discipline of responding quickly and fixing what went wrong.

Automate Review Requests After Purchase: The Timed Ask That Turns Happy Customers Into Public Proof