Imagine a customer who walks out of your shop at 2pm delighted, tells a friend she just found a great place, then gets a review request three days later while she is at work on a deadline and can barely remember what service she had.
The hypothesis behind this guide is that the delay, not the customer, is the variable to test first. A request sent the same afternoon, on the phone she just paid on, is a starting point to compare against your current setup, measured against your own response numbers.
The mechanics below cover timing, channel, and effort. None of them require a longer message. A shorter one is worth testing.
Timing as the first variable
- 0hService ends
- 30mPayment complete
- 2hBest ask window
- 24hAttention drops off
- 3d+Skip if not sent
- 0hService ends
- 30mPayment complete
- 2hBest ask window
- 24hAttention drops off
- 3d+Skip if not sent
Timing is one factor in whether a customer leaves a review, alongside the message and the platform. As a starting hypothesis to test against your own results, try the window between two and six hours after they walk out, then compare response counts with whatever timing you use today.
Not the next morning. Not that evening on the couch. That afternoon, while the visit is recent and the phone is already in hand.
Set the request to fire on the appointment-completed event, delayed by the shortest window your booking tool allows. Automate this so the request fires consistently rather than only when someone remembers.
One channel, the one they already use
Send the request on the channel the customer used to book. If they booked by text, send by text. If they booked by email, send by email. Do not send both, and do not switch on them at the finish line.
Two requests can read as pressure. One request on the channel they already use is the design to test first.
Channel behavior is a hypothesis, not a rule: compare text against email using your own response and opt-out data before deciding. Match the channel to how the customer chose to interact with you, not to how you prefer to communicate.
One tap to complete
Where the platform's current features and terms permit it, point the link at the review composer rather than the profile page. Google, Facebook, and Yelp each document their own link behavior. Confirm the current terms and features for each platform before configuring anything.
If the customer has to search for your business, tap the star ring, then tap write review, each added step is another point where attention can move elsewhere. Reducing the number of taps is the variable to test against your own response data.
Do not prefill any part of the review. Confirm the current terms of each review platform before configuring the request.
The message is short and specific
Name the service and the person. 'Thanks for coming in today for your [service]. If you have 30 seconds, a review helps us more than you would think: [link].' That is the whole message.
No branding block. No 'our small business' preamble. No emoji. Treat message length as a variable: measure response counts against your own recorded data rather than assuming shorter or longer performs better.
The eligibility rule
Eligibility is based on the completed appointment, never on predicted or expressed satisfaction. No-shows and cancellations are not completed appointments, so no request is sent for them.
Every otherwise eligible completed customer receives the same public request, including anyone who complained during the visit. Private service recovery runs in parallel: it never delays, filters, or suppresses that public request.
Frequently asked questions
start here
See how your review flow reads to a new customer.
The scan grades trust signals on your homepage and tells you the top five fixes in order.
No card, no sales call required. Report emailed to you.