Review request timing: when in the customer journey to ask

Most businesses ask at whatever moment is convenient for their software. The invoice goes out, so the request rides along with it. The problem is that an invoice is the least emotionally positive moment in the entire relationship, and you attached your ask to it.

Timing is not a small optimization here. It changes both how many people respond and what they write.

When is the moment of relief for my business?

Find the point where the customer's worry ends. That is the ask.

Restaurants. Not at the check drop, which is a payment moment. At the door on the way out, or in a followup message that evening. The guest is full, the evening went well, and nobody is asking them for money.

Home services. At the completion walkthrough, standing next to the finished work. The customer just confirmed the problem is solved. That is the highest emotional point in the entire job and it is over in ninety seconds.

Healthcare and dental. At checkout, after the appointment, before the statement arrives. A request that lands with a bill reads as a request from a bill.

Professional services. At the delivery of the outcome, not at the end of the engagement. If a case closes or a filing is accepted, that day is the day.

Anything sold for later use. After first use, not after delivery. A customer who has the box but has not opened it has nothing to review.

The rule underneath all of these: ask when the customer knows the answer, not when your system generates a document.

Does the evidence actually support asking sooner?

It supports the direction. It does not support a number, and you should know which is which.

There is no published study of review request timing with a disclosed sample and method. Every completion rate by hour that circulates comes from a company selling review software, using its own platform data, without an independent methodology attached. Treat those figures as marketing, including the ones that flatter the advice in this article.

What the evidence base does support, and what the vendors' own operating defaults quietly agree on, is a short list: ask within hours rather than days, lead with SMS rather than email, send exactly one reminder, and suppress the sequence the moment a review arrives. Standard vendor cadence is a message at about two hours after the visit and a single reminder two days later. That convergence is not proof, but when competing products with competing claims all ship the same timing, the timing is doing something.

The mechanism underneath it is not mysterious. Response rate falls as memory fades. A person asked the same day is being asked about something they can still picture. A person asked three weeks later is being asked to reconstruct an experience, and most decide it is not worth the effort.

Detail falls faster than response rate, and this is the part that costs you twice. The late review is short and generic. "Good service." The same customer asked the same day writes about the technician's name, what was fixed, and how long it took. Specific reviews are the ones the next customer actually reads.

What does the timing decision look like in numbers?

Build the model with your own volume. The completion rates below are placeholders, not research. Use them to see the shape, then replace them after thirty days of your own measurement.

Say you serve 400 customers a month.

The gap between the first row and the third is 23 reviews a month, or 276 a year. That is not a marketing improvement. On the competitor benchmarking method, where you count your rivals' ninety day totals and set your target one above the leader, 23 a month is usually the difference between chasing and leading.

Now add the reminder. One reminder at two days, with suppression on receipt so nobody who already wrote a review gets nagged, typically lifts the first touch yield. If it adds a third, the same day row goes from 33 to 44.

Then check the compliance cost of the fastest option. The temptation at the moment of relief is a tablet in the customer's hands before they leave. Google's policy says merchants should not require or pressure users to leave reviews while on the premises. A verbal ask plus a card they take with them is an invitation. A staff member standing there while they type is pressure. Same moment, same conversion logic, completely different exposure.

How to make these numbers yours. Run one ask point for thirty days and count reviews. Switch and run the next for thirty days. Two months of your own data beats any table, including this one.

How do I attach the ask to a moment that already happens?

Pick an existing step and add one sentence to it. Never create a new step.

Better still, trigger it off your system of record rather than a calendar. The point of sale, the practice management system, the field service software. Toast, Square, Dentrix, Open Dental, ServiceTitan, Jobber and Housecall Pro all know the exact moment a job closed or a ticket paid, and a request fired from that event happens whether or not anyone remembers it.

This is the operating principle behind every review system that survives contact with a busy week. A separate task called "ask for reviews" dies by week three. A line added to the closeout checklist or the confirmation message runs indefinitely because it rides on something nobody can skip.

That matters more than it sounds, because the moment of relief usually sits inside your busiest hour. Maple's analysis of 1.2 million calls across more than a thousand restaurant and local business locations found 68% of calls arriving during the lunch and dinner rushes, 28% at lunch and 40% at dinner. That is call data rather than review data, but it describes the same constraint: the moment your customer is most ready to be asked is the moment your staff have the least capacity to ask. A trigger fires anyway. A person does not.

Name the step. Not "after service." The specific thing: the walkthrough, the check presenter, the discharge instructions, the text that confirms the job is done.

Then remove every step between the ask and the review. A short link or a code that opens the review form directly, one tap, no login screens if you can help it. QR codes and kiosks can do this and stay compliant, and the compliance part matters every bit as much.

Make it one sentence a human would say. If the script sounds like it was written by a marketing department, staff will not say it out loud, and a script nobody uses is not a system.

And ask everyone who reaches that step. Sorting by who seems happy is gating, which Google bans outright and which cuts your volume at the same time. A negative review beats no new reviews, and that is the argument that usually unblocks an owner who has been afraid to ask.

What can go wrong even with perfect timing?

Three things, and they are all upstream of the ask.

The customer lands on the wrong profile. Multiple listings, an old location, or a duplicate. The review goes somewhere that does not help you. Before you scale the asking, confirm the link points where you think it does, and that the profile itself is set up correctly, starting with the categories that are the single strongest lever on a Google Business Profile.

The business changed and the profile did not. If you renamed, rebranded, or moved recently, hold the push until the dust settles. A name change or rebrand creates a period of instability, and reviews collected during it can land on a listing you are about to abandon.

The experience was not good. No timing fixes this. If you ask at the perfect moment and the customer is unhappy, you will get an accurate review. That is working as intended, and the fix is operational.

Where does the timing rule break down?

Two places, and the second one catches people who did everything else right.

When the outcome is not knowable at completion. A dental implant, a landscaping install, a legal filing awaiting a decision. Asking at the moment the work stops is asking before the customer knows whether it worked. In those cases the moment of relief is the first followup visit or the first confirmation of the result, which may be weeks out, and a same day ask will underperform a patient one.

When you have a backlog and you try to fix timing retroactively. Realising the ask should have gone out sooner does not create permission to email two years of customers today. Google's policy names content exhibiting unusual volumes or patterns of review contributions as a violation in its own right, and a silent profile producing forty reviews in a week is that pattern. The recovery is a drip at your competitor's rate plus one, and getting reviews from customers who loved you two years ago is a separate exercise with its own pacing rules.

What to do this week

Write down your customer journey in six steps or fewer. Booking, arrival, service, completion, payment, followup. Circle the moment the customer knows the outcome was good.

Add one sentence to that step, and wire the trigger into whatever system already records that the step happened.

Test your review link on your own phone from a fresh browser. Count the taps between the link and the typing box. If it is more than two, fix that before anything else.

Then check where your current requests are being sent from in your journey. If it is the invoice, move it. That one change is usually the largest single improvement available.

Want a read on where your ask should sit?

Email me at eric@seod.com with your six journey steps and tell me where the request currently goes out. I will tell you which step is your moment of relief, write the one sentence to attach to it, and flag anything in the current flow that could get your reviews filtered.

It takes me about ten minutes and it is usually one change, not five. If your timing is already right, I will say so and tell you where I would look instead.

Otherwise keep reading the reviews and reputation library.