Online Reviews: Collecting, Responding and Using Them in Sales

Why do reviews often show a skewed picture of your business?

A business has served 400 customers in ten years and has 23 reviews online. Four of them are scathing, the rest enthusiastic. The many customers who were satisfied but not enthusiastic do not appear. A prospect reading the profile nevertheless sees a seemingly complete picture and decides accordingly.

The question this article sets out to answer: How does a business collect reviews so that they give an honest picture, how does it respond to them, and how does it use them in sales without breaking rules? To answer it, I first explain why reviews are biased. Then we weigh up what asking for a review achieves, good and bad. After that come the rules, the process and use in sales.

How to set up the Google Business Profile and why replies to criticism work is described in the article Google Business Profile. This article goes one step further: from a single profile to handling reviews on all platforms.

If you do not ask, you let the extremes speak. If you ask everyone, you give the middle a voice.

The thinking model: extremity bias

An online review is a public assessment of a provider by a customer, usually with stars and text, on a platform the provider does not control. Besides the Google Business Profile, review platforms include industry portals, employer review sites and comparison sites.

The most important thinking model for handling reviews is extremity bias. It means that whether a customer writes a review on their own depends on how strong their experience was. Those who are very satisfied or very annoyed have a reason to write. Those who received a decent, unremarkable service almost never write. The visible reviews therefore mainly reflect the extremes. Karaman (2021) was able to demonstrate this because her data also contained the private satisfaction scores of customers who had not written a review. Unsolicited reviews showed both this bias and a tendency to conform to the reviews already visible.

How big the difference between asking and not asking is shows a second study by the same author. Karaman (2026) analyzed data from an international hotel group that, by its own account, operates about five percent of all hotels in the United States. In June 2014, loyalty program members completed 218,041 satisfaction surveys. A randomly selected share was then asked to write a review on the website, without any incentive and in neutral wording.

Almost nobody wrote unprompted. 92 percent of the 11,436 reviews came from customers who had been asked. For a small business, this means the number and composition of your reviews depend less on your performance than on whether and whom you ask.

Ask or wait: two readings

The first reading clearly favors asking. Karaman (2021) found that asking activates all customers, but those with moderate experiences more than those with extreme ones. The bias does not disappear, but the reviews become considerably more representative. A prospect then sees a picture closer to actual satisfaction.

The second reading names the side effects. Gao, Wang, Ding and colleagues (2025) used a natural experiment: TripAdvisor introduced a program that allowed hotels to ask guests for reviews after their stay. Compared with the same hotels on Expedia, the picture was mixed.

The authors explain the decline in spontaneous reviews with a crowding-out effect: people who see that reviews are being collected in an organized way feel their own voluntary contribution is less necessary. Karaman (2026) found another side effect. Those who wrote a review spent on average 17 percent less with the hotel group afterwards; across everyone asked, revenue fell by 7 percent. The effect was unevenly distributed. Very satisfied customers with the top score of 10 out of 10 did not change their behavior. Clearly dissatisfied customers who wrote a critical review spent considerably less afterwards.

How should the two readings be weighed? I consider the first the more important, for a simple reason: a business with few, biased reviews loses prospects it never gets to know. The side effects should nevertheless be taken seriously. The conclusion is not to ask only satisfied customers, because that is not allowed. It is: solve the problem first, then ask. Whoever settles a customer’s frustration in a direct conversation and only makes the request after a clean acceptance asks all customers at the same point and in the same way, but not in the middle of an unresolved conflict.

Which rules apply to reviews?

Reviews are not a free-for-all, either legally or under platform rules. The following overview summarizes the most important points as of November 2026, from a German perspective. It is a simplified orientation and not legal advice.

When policies are violated, Google can remove fake reviews and restrict the business profile, for example by temporarily blocking new reviews. The German UWG obligations on consumer reviews have applied since European rules were implemented in May 2022. How courts interpret the requirements for reasonable verification in detail is still evolving. If in doubt, get professional advice.

How do you collect reviews fairly?

From extremity bias and the rules follows a process I call the ask-alike rule: every customer is asked at the same point in the process, through the same channel and with the same words. This makes reviews more representative and protects against the accusation of selection.

1. Fix the timing

   A fixed step in the process, for example after acceptance or after the final invoice. Not when someone happens to be thanking you, because that would be selection.

2. Settle open items beforehand

   At acceptance, ask whether everything is done. An open defect is fixed before the request goes out. This applies equally to all customers.

3. Send a direct link

   By email or as a QR code on the invoice. One click less brings more reviews.

4. Word it neutrally

   Ask for an honest review, not for five stars. No incentive, no prize draw.

5. Assign responsibility

   One person sends the request after every job. Without clear responsibility, only those who stand out get asked.

With business customers, there is an additional feature. Many contact persons are not allowed or do not want to publicly review on behalf of their company. Accept a no without pressing and instead offer to agree on a more detailed testimonial or a case study together. How that works is described in the articles Using testimonials and references and Reference projects as case studies.

How do you respond to reviews?

Replies to reviews are aimed mainly at later readers. Wang and Chaudhry (2018) found in hotel review data that replies to negative reviews improved the ratings of later guests, while lengthy replies to positive ones tended to hurt. The effects were small on average, 0.07 to 0.13 stars depending on the time window, but the direction is clear. The article Google Business Profile describes a four-part structure for replies to criticism: acknowledge, explain, change, offer.

With business customers, particular restraint applies to details. Contract details, prices or names of the customer’s employees do not belong in a public reply, even if they would support your own view. How companies communicate in crises and why openness works better than justification is discussed in the article Deny, justify, distract.

How do you use reviews in sales and in operations?

Reviews have two uses that are often considered separately. Externally they are proof; internally they are feedback. Both can be combined with little effort.

• Show them on the website. A few recent reviews with a link to the platform, next to the offer. Anyone displaying reviews on their own site states whether and how their authenticity is checked.
• Refer to them in proposals. One sentence with average, number and a link to the platform. That is verifiable and therefore effective.
• Use them in the first conversation. If a prospect raises a critical review, explain openly what happened and what has changed.
• Evaluate them as an early warning system. Once a quarter, sort all reviews by topic. If references to the same problem pile up, that is a task for operations, not for marketing.

For businesses with a service area, reviews are also part of the local prominence by which search engines rank their results. How this fits together is described in the article Local SEO.

What follows in general

Reviews are an example of how transparency has to be designed in order to be honest. A profile nobody maintains is not neutral; it shows the extremes. A profile that shows only enthusiasts is embellished and breaks rules. The picture only becomes honest when everyone is asked equally and criticism is visibly answered.

The limits of the findings are clear. All studies cited come from the hotel industry and mostly from the United States. The decline in spending after writing a review was measured for a hotel group with a loyalty program, not for trade businesses or B2B service providers. What transfers is the mechanism: whoever writes commits themselves, and whoever is dissatisfied and commits publicly comes back less often.

Conclusion and a first step

The opening question was
a business collects reviews, responds to them and uses them in sales without breaking rules. The answer: it asks all customers at the same point, in the same way and without incentives, settles open problems before asking instead of selecting, answers criticism specifically and praise briefly, and evaluates reviews as feedback on its own processes. This creates a picture that is more honest than the extremes and more credible than any glossy page.

• List all platforms on which your business is reviewed and secure the access.
• Set a fixed point in the process for requesting a review.
• Write a neutral text with a direct link, without any incentive.
• Check whether anywhere only selected customers are asked or benefits are offered, and stop it.
• Answer unanswered critical reviews from the last twelve months.
• State on the website whether and how the displayed reviews are checked.
• Count the topics of critical reviews once a quarter.

Reviews are not an advertising tool you control but a mirror you set up honestly.