Most companies' review scores are lower than they deserve, for a simple reason: unhappy customers write reviews on their own, and happy ones almost never do. The angry ones have a reason to type. Everyone else has already moved on.
The result is a rating that measures who felt strongly enough to post, not how good you are. The good news is that this is one of the most fixable problems in marketing.
Why the score is skewed
Think about what it takes for a customer to leave a review unprompted. They have to remember, find the site, sign in and write something. Frustration provides that energy. Quiet satisfaction doesn't.
So the reviews you collect without trying come mostly from people with a problem. Even a business with very happy customers can end up with a mediocre average built from a few bad weeks.
AI models read that average too. Ask one whether a company is any good and it will often reach for the rating and a sentence about what people complain about. Both come from whoever reviewed you last.
The fix: ask, at the right moment
The silent majority will review you if you ask them, and if you ask at the right time. Satisfaction isn't constant across a customer journey. It peaks at predictable moments, and those are the moments to ask.
Finding the peak
For most businesses, the peak isn't the moment of purchase. It's the moment the customer feels the benefit. A few examples:
- Software: just after a customer finishes setup and sees their first real result, or after a support ticket is resolved quickly.
- Physical products: a few days after the item has arrived and been used, not on delivery day, when it might still be in the box.
- Services: right after a milestone the customer cares about, like a project delivered or a problem solved.
- Repeat purchases: the second order is a strong signal. Someone who comes back is someone who's happy.
Look at your own journey and ask: when do customers say thank you? That's usually the moment.
Asking once, and simply
One request, at that moment, sent to every customer who reaches it. Make it take under a minute: a direct link to the review page, no long survey first. Don't chase people with three reminders. A single well-timed ask works better than a campaign.
The rules: no incentives, no filtering
There are two shortcuts that look tempting and cause real damage.
Paying for reviews, or rewarding positive ones. Offering a discount or gift in exchange for a good review is against the rules of every major review platform and, in many places, against consumer protection law. In the US, the Federal Trade Commission has specifically banned buying fake or sentiment-conditioned reviews. Platforms remove reviews they suspect were incentivised, sometimes taking a company's whole history with them.
Gating the ask. Some tools send a quick "how did we do?" survey first, then only ask happy customers for a public review. This is often called review gating, and platforms such as Google prohibit it. It's also a bad bet: the resulting score looks manufactured, and it can collapse in one sweep.
Ask everyone. Your honest average is almost certainly higher than your current one.
What happens to the distribution
When you start asking everyone at the right moment, the new reviews come mostly from people who are satisfied, because most of your customers are. Over a quarter or so, the recent reviews start to outweigh the old angry ones, and the average moves toward your real level of quality.
The critical reviews don't disappear, and they shouldn't. Answer every one of them in public: what went wrong, what you changed, how to get it sorted. A thoughtful reply to a one-star review often persuades more readers than the review put off. And a profile with some criticism, well handled, is more believable than a wall of five stars.
What this won't fix
If your product or service is genuinely poor, asking more people will confirm it faster. Review programmes amplify an accurate reputation in both directions. If you're worried about what more reviews would say, that's worth knowing, and worth fixing first.
But if you suspect your score is worse than your customers' real opinion, it probably is. The moment the happy ones will post is already somewhere in your customer journey. You just have to find it and ask.