The default is silence
Almost everybody reads reviews. Almost nobody writes them. Consumer surveys put readership near universal 97%, and the same body of work finds that a large majority of businesses never reply to the reviews they do get 75%. Those two facts describe the shape of the problem: a very large audience, a very small number of contributors, and almost no conversation between them.
This matters because of how people read a missing review. A customer who had an ordinary, fine experience has no story to tell, so they tell nobody. The people who do write are the two tails: something went unusually well, or something went wrong. Left alone, a rating drifts towards whichever tail is louder in your category, which is usually the unhappy one, because a disappointment carries an obligation to warn other people and a good stay carries none.
So the first thing to fix is not the average. It is the sample. A rating built from the two tails is not a measurement of your business, it is a measurement of your extremes.
Volume is the number you are short of
When a business worries about reviews, it usually worries about the star average. The evidence points somewhere else. Analysis of small business revenue against review counts found that businesses carrying more reviews than the survey average earned materially more, and that the businesses with the largest review counts earned roughly twice the average 54% / 2x.
Read that carefully, because it is a correlation and it runs in both directions: busy businesses collect more reviews, and businesses with more reviews get chosen more often. You cannot separate the two from that data alone. What you can take from it is the weaker, sturdier claim. A rating carried by nine hundred ratings is read as a fact about the business. A rating carried by nine is read as an accident, and most people discount it accordingly, which is exactly how the GoCredible Score treats it.
A practical consequence: if you have a 4.8 from thirty ratings, the highest value work available to you is not defending the 4.8. It is getting to three hundred ratings, even if the average settles lower on the way.
What actually changes the odds
Three things move the number of people who write, and none of them is a discount code.
Timing. Ask while the value is still in the room. For a hotel that is at checkout, not four days later; for a clinic it is after the follow up call that told the patient they were fine. An ask that arrives once the feeling has faded is asking for a favour with nothing attached to it.
The shape of the ask. A request lands differently when something was given first, and the effect is strongest when what was given was unrequested and specific to that person Robert Cialdini. In practice that is not a gift. It is the receptionist who remembered the room preference, or the manager who fixed the billing error the same afternoon and then, separately, said that a review would help the team. The favour has to be real and it has to come first.
Friction. Every extra tap loses people. One link, one destination, no app install, no login the customer does not already have. If your ask sends someone to a page that then asks them to choose a platform, you have converted a willing reviewer into a decision they did not want to make.
What does not work, and carries real downside, is paying for it. Incentives change what people write rather than how many people write, every major platform removes incentivised reviews when it finds them, and a business that has been caught tends to lose the honest reviews alongside the bought ones. The same applies to asking only the customers you expect to be happy. That is review gating, it is against platform policy, and it produces an average that tells you nothing you did not already assume.
What to do with what arrives
Collecting is half of it. The other half is answering, and it is the half almost nobody does, given that roughly three quarters of businesses never reply at all. A reply is not written for the person who complained. It is written for the next person reading, who is deciding what happens if something goes wrong for them. That reader is the whole audience.
Keep replies short, name the specific thing, say what changed, and move the rest off the platform. Never argue the facts in public, even when you are right, because a reader cannot verify your version and will remember only that the business argued.