If you added three channels at once and reviews went up, you learned nothing you can act on. Here is how to run the test properly and what attribution is genuinely possible.
You changed three things last quarter: added a QR code at the counter, started sending texts after service, and trained the team to ask in person. Reviews went up. You have no idea which of the three did it, so next quarter you keep all three running, including the two that are doing nothing.
Review attribution is unglamorous and it is the difference between a review programme you can improve and one you simply maintain.
This guide covers how to work out which channel actually produces your reviews, what you can and cannot measure, and how to run a test that gives a usable answer.
Review platforms do not tell you where a review came from, so attribution has to be inferred from timing, volume changes and deliberately separated channels.
Set expectations before building anything elaborate.
Not available:
Available through inference:
The honest position is that attribution here is directional. You can tell which channel is working, not precisely how many reviews each produced.
🛠️ Action Step: Before adding any new channel, record your monthly review count for the previous three months. Without that baseline, nothing afterwards is measurable.
Introduce one review channel at a time and hold it for a month, because simultaneous changes make attribution impossible.
This is the whole method and it is where most programmes fail.
A workable sequence:
It is slower than launching everything at once and it produces knowledge rather than a mystery. If you introduce three channels simultaneously and reviews double, you have learned nothing you can act on.
Use distinct short links or QR codes per channel, because click data on your own link is the closest thing to real attribution available.
The platform will not tell you the source, but your own redirect will.
How to set it up:
You measure clicks, not reviews, which is a proxy rather than an answer. But a channel producing no clicks is certainly producing no reviews, and that alone is worth knowing.
💡 Pro Tip: The gap between clicks and reviews is itself informative. A channel producing many clicks and few reviews has a friction problem after the click, usually meaning the link goes somewhere requiring extra steps.
Run different methods at different locations, because a parallel comparison produces an answer in one month rather than four.
If you have more than one location, you have a natural experiment available.
How to structure it:
Normalise by customer volume. A busier location will produce more reviews regardless of method, so the comparison has to be per transaction.
With one location, the same logic applies across teams or shifts, though the sample is smaller and the result less reliable.
Track reviews as a percentage of customers served, because raw counts move with business volume and hide what your channels are doing.
A busy month producing more reviews tells you nothing about whether your asking improved.
What to calculate:
A typical local business converts a low single-digit percentage of customers into reviewers. Knowing your own number tells you whether a change helped, which a raw count never does.
📊 Flento Data: Across local profiles, review rate per customer is a more stable indicator of programme health than monthly volume, because volume tracks business activity rather than asking effectiveness.
Expect in-person asking with a direct link to outperform every automated channel, though automation wins on consistency.
Most businesses discover the same pattern.
Typical findings:
The practical answer for most businesses is both. Automate so the ask always happens, and train the in-person ask because it converts best when someone remembers.
🔥 Quick Win: If you run only one channel, make it the automated post-service message. It is the one that survives busy weeks, which is when manual asking stops.
Structure for that is in how to build a review management system.
Flento provides the request tracking that makes directional attribution possible.
Google Review Management Software tracks requests sent and reviews received over time, which is the comparison the whole method rests on.
Smart QR Code Generator creates distinct codes per placement so you can tell which physical channel works.
Google Business Profile Optimizer keeps the profile complete so friction after the click is not the limiting factor.
✅ Done? Track requests and reviews over time so the comparison is possible. Get started free
Q: Can you tell which channel produced a specific review? A: No. Platforms do not expose a source. Attribution is inferred from timing, volume changes and separated channels, which gives a directional answer rather than a precise one.
Q: How do you test a new review channel properly? A: Introduce one at a time and hold it for a month against a recorded baseline. Launching several simultaneously makes the result uninterpretable.
Q: Do separate links actually help? A: Yes, for click data. You measure clicks rather than reviews, but a channel producing no clicks is certainly producing no reviews, and the click-to-review gap reveals friction.
Q: Why track rate rather than review count? A: Because raw counts move with business volume. Reviews as a percentage of customers served tells you whether your asking improved; a monthly total does not.
Q: What channel usually performs best? A: In-person asking with a direct link converts best per request, while automated messaging wins on consistency. Most businesses benefit from running both.
Q: What if I only have one location? A: Compare across teams or shifts, or run sequential monthly tests. The sample is smaller and slower but the method is the same.
If you added three things at once and reviews went up, you know something worked and nothing about what. That is a comfortable place to stay and an expensive one, because you are maintaining channels that may be doing nothing.
Record the baseline. Add one thing. Wait a month.
It is slower than the alternative and it is the only version that tells you anything. Try Flento free.
