Review Sites for Startups: How to Earn Honest G2 and Capterra Reviews

1. Blog
2. Demand & Channels

By Saroj Jha · October 5, 2026 · 11 min read

Review sites help a B2B startup when its buyers already check them, and they reward a steady flow of honest reviews from a fair cross-section of customers. Ask every eligible customer, never pay for praise or hide criticism, and put what reviewers say to work in sales. The five steps below show how, with one example company.

When a CFO builds a shortlist, a review site is one of the few places to hear from peers. An empty profile can quietly cost you a place on that list; a padded one can cost you more. Since October 21, 2024, a federal rule has let courts impose civil penalties for knowing violations, such as fake reviews and review suppression^[3].

The market has also consolidated. In January 2026, G2 announced it would acquire Capterra, Software Advice and GetApp from Gartner^[4]. Gartner's annual report says: "On February 5, 2026, we completed the sale of Digital Markets for approximately $110.0 million, prior to customary purchase price adjustments"^[5]. In my view, that makes one well-run review program matter more, not less.

Running case: Example Co. Fictional

Example Co. is a seed-stage B2B SaaS startup whose software automates the month-end close for finance teams at companies with 50 to 500 employees. The controller champions it and uses it daily, the CFO or VP Finance signs, and IT reviews security. It costs about $1,500 a month (about $18,000 a year), a typical sale takes about 45 days, and the goal is 30 new customers in the next 12 months. At this point in its story, later than in the other posts, it has 40 customers and three reviews on one site, and two CFOs recently asked where they could read what other finance teams think. Example Co. and all its numbers are illustrative.

Step 1: Check whether your buyers use review sites, then pick one

Start with your buyers, not the sites. Ask the last ten people who bought or evaluated your product where they looked, then count your two closest competitors' reviews on those sites.

The number of reviews matters more than a perfect score. A meta-analysis of 96 studies (1,532 effect sizes across 40 platforms) found that online word of mouth "is positively correlated with sales (.091)", and that "eWOM volume has a stronger impact on sales than eWOM valence"^[6]. Most studies were of consumer products, so read it as direction, not a B2B forecast.

Sites also have thresholds. G2's methodology says a product needs "at least 10 reviews in the corresponding category" to be included in its Grid reports^[1]. Clear that bar on the one site your buyers mention most before spreading out.

Example Co., step 1 Fictional

Six of ten recent buyers checked a review site, and both CFOs named the same one. There, Competitor A has about 85 reviews and Competitor B about 40; Example Co. has three. The team picks that site and targets 15 new reviews in two quarters, clearing a 10-review bar with room to spare. All numbers are fictional.

Do this

Ask your last ten buyers which review site, if any, they checked, and count your top two competitors' reviews there.

Step 2: Learn the rules before you ask anyone

Two sets of rules apply: US federal rules and each site's own policy. This is not legal advice, but the main lines are clear.

The FTC's Consumer Reviews and Testimonials Rule (16 CFR Part 465) bans fake reviews (§ 465.2) and offering "compensation or other incentives in exchange for, or conditioned expressly or by implication on, the writing or creation of consumer reviews expressing a particular sentiment, whether positive or negative" (§ 465.4). It restricts undisclosed insider reviews (§ 465.5) and bans review suppression, such as using "an unfounded or groundless legal threat, a physical threat, intimidation, or a public false accusation" to stop or remove a review (§ 465.7)^[7].

Incentives themselves are not banned: "The rule does not prohibit giving incentives for reviews, as long as there isn't an express or implied requirement that the reviews have to express a particular sentiment." On asking only happy customers, the FTC's Q&A says the rule has no specific ban, "But this practice could violate the FTC Act"^[3].

The FTC's Endorsement Guides (16 CFR Part 255) require a connection that "might materially affect the weight or credibility of the endorsement", such as an incentive, to be "disclosed clearly and conspicuously" when the audience would not expect it (§ 255.5). And when advertisers procure, suppress or edit reviews, they "should not take actions that have the effect of distorting or otherwise misrepresenting what consumers think of their products" (§ 255.2(d))^[8].

The rule speaks of "consumer reviews", and the FTC's Q&A does not address business buyers. I would not build a plan on that gap, because the sites draw the same lines:

• G2 says "eligibility to receive an incentive is never based on the opinions, positive or negative, expressed in the review", labels incentivized reviews, bars reviews by the vendor's and direct competitors' employees, and bans "collecting reviews by any method that segments out potentially negative reviews or exclusively targets positive reviews"^[9].
• Capterra, Software Advice and GetApp share one set of guidelines (updated May 4, 2026): "Vendors must solicit reviews in a fair and unbiased manner", which includes "not selectively targeting users based on expected sentiment or prior feedback". Vendor incentives go through the site's own program and must be "offered equally to all eligible participants, regardless of the rating provided"^[10].

What the FTC's cases show

The FTC alleged that Fashion Nova "used a third-party online product review management interface to automatically post four- and five-star reviews to its website", while "from late-2015 until November 2019, Fashion Nova never approved or posted the hundreds of thousands of lower-starred, more negative reviews." Fashion Nova agreed to pay "$4.2 million to settle Federal Trade Commission allegations", and the FTC called it "the FTC's first involving a company's efforts to conceal negative customer reviews"^[11]. It is a consumer case, but publishing the good and holding back the bad is easy to copy in any review widget. In December 2025, the FTC sent warning letters to 10 companies about possible violations of the new rule and noted penalties of up to "$53,088 per violation"^[2]. These are the FTC's allegations and statements, not an account of how those companies ran things internally.

AI is named in the rule's own announcement. The FTC said the rule covers reviews that "misrepresent that they are by someone who does not exist, such as AI-generated fake reviews"^[12]. And AI branding is no shield. The FTC alleged that Sitejabber, "a company offering an AI-enabled consumer review platform", "collected ratings and reviews from consumers for its online business clients at the time of purchase, before they received or had the chance to experience the products or services they bought", and "used these ratings and reviews to deceptively inflate the average ratings and review counts of its clients on the company's review platform." The final order, approved 5-0 in January 2025, "prohibits Sitejabber from making, or assisting anyone else in making, misrepresentations about any ratings, average ratings, or reviews it collects, moderates, or displays"^[13]. These are the FTC's allegations and order terms, not an account of how Sitejabber ran things internally. In my view, the lesson for any startup is simple: a rating collected before the customer has used the product is not a review of the product.

Example Co., step 2 Fictional

Before any request goes out, the team writes a one-page policy: ask every active customer; never filter by NPS or support history; use only the site's own link and incentive; no reviews from staff, investors or their families; never pressure anyone. The policy is fictional; the rules it follows are not.

Do this

Write your one-page review policy and have someone senior sign it before the first request goes out.

Step 3: Ask every eligible customer, at a moment of value

Asking is legitimate, and it makes your reviews more representative. A study that could see customers' private satisfaction ratings found that "solicitations significantly increase the representativeness of rating distributions", because customers with moderate experiences respond more when asked than those at the extremes^[14].

Incentives need care. Across a set of studies, "Incentivized (vs. unincentivized) reviews contained a greater proportion of positive relative to negative emotion", even when the reward did not depend on the rating^[15]. That is why sites label them. If you offer anything, use the site's own program and let the label show.

Ask right after a result the customer can describe: for most B2B products, after the first real outcome, not at signing.

If you sell an AI product, two things change. First, in my view the moment of value comes later. Ask after the customer has used the product on real work long enough to see where it is right and where it needs checking, not after the first impressive output. Second, never draft reviews for customers, with an AI writing tool or without; the review has to be their experience in their own words. AAJ's judgement: in the request, invite reviewers to name the task they use the product for. A buyer of an AI tool wants to know whether it worked on a job like theirs, and a neutral question about the task does not steer the rating.

Example Co., step 3 Fictional

The trigger is a customer's second completed close, when the controller has seen it work twice. The founder emails all 40 controllers with the site's own review link, plus one reminder. At an assumed 40% response, that is about 16 reviews over two quarters. After that, every new customer gets the same email after its second close. All numbers are fictional.

Keep the flow steady. G2 says a review's weight decays "gradually for the first 90 days and maintains a stronger weight for the first 18 months, after which the rate of decay accelerates. After about three years, the review reaches approximately 3% of its original weight"^[1].

Do this

Pick your trigger moment, then send the request to your whole customer list, not a hand-picked few.

Step 4: Answer every review, especially the critical ones

A critical review is evidence that your reviews are real. Trying to make it disappear is what hurts: the rule bans intimidation and groundless legal threats, G2 bans "discouraging negative reviews of your product or service"^[9], and Capterra's guidelines bar using non-disparagement clauses to influence reviews^[10].

Faking it rarely lasts either. A study of Yelp found that "roughly 16% of restaurant reviews on Yelp are filtered" as suspicious, and that restaurants with weak reputations were more likely to commit review fraud^[16]. A study of products that bought fake reviews on Amazon found that buying fake reviews brought only a short-term rise in average rating, and that after firms stopped buying them, their average ratings fell and the share of one-star reviews rose^[17].

Reply in public: thank the reviewer, name the issue, say what changed, offer a direct contact. Then fix it.

Example Co., step 4 Fictional

Review 9 gives three stars: reconciliations work, but connecting bank feeds took two weeks. The founder replies within two days, says a guided setup call now happens in week one, and offers a direct email. The review stays up, and the account executive now raises setup time in demos. All details are fictional.

Do this

Reply to every review within a week, and log each criticism as a product or onboarding task.

Step 5: Put reviews to work in sales and marketing

Reviews earn their keep after they are published. Quote them in sales decks, accurately and in context; use them to find case-study customers; and read them for the words buyers use.

What Jamf's filing shows

Jamf, which sells software to manage and secure Apple devices at work, told investors in its annual report for 2022: "In addition, hundreds of independent customer ratings on popular software review websites, including Gartner Peer Insights, G2Crowd, and Capterra, have earned Jamf recognition as the 'Customers' Choice.'" Its marketing section adds: "we capitalize on the voices of our highly satisfied and loyal customers using a variety of customer advocacy tactics including case studies and videos, software reviews, social amplification, references, and referrals"^[18]. This is what the filing says, not an account of how Jamf made its choices internally.

Example Co., step 5 Fictional

Once a month, sales, product and the founder read new reviews for 30 minutes. One line goes into the demo deck with the reviewer's job title; reviewers who describe a measurable result, such as days saved at close, are invited to a case study interview; setup complaints go to the onboarding backlog. The routine is fictional.

Do this

Set a monthly 30-minute review of new reviews, with sales and product in the room.

Free template

Turn a strong review into a case study

The free Case Study Kit has a permission request, the eight things a sceptical buyer checks and a structure for the write-up.

What changes by stage?

• Pre-seed — Claim your listing and keep its facts right; spend the time on customer conversations.
• Seed — One site, one written policy, and a request to every customer at a moment of value. Aim past the site's minimum (on G2, 10 reviews in a category).
• Series A–B — A steady program on each site buyers use, with a named owner, connected to case studies, partners and any marketplace listings.

What mistakes should I avoid?

• Asking only your happiest customers. Both G2 and Capterra ban it, and the FTC says it could violate the FTC Act.
• Rewards for positive reviews. Any reward must be the same whatever the rating, and disclosed.
• Pressuring or hiding critical reviews. Reply and fix instead.
• One big push, then silence. On G2, older reviews count for less.

Example Co. is a fictional company. Example Co. and all its numbers are illustrative. This post is not legal advice.