TestGorilla reviews, sorted by who was paid to write them

Searching this usually means one of two things: someone has quoted you a price and you want to know whether the product is any good, or you already pay for it and you are checking whether other people had the same problem you did.

Both are answerable from the public record, but not by reading the star rating. TestGorilla carries five different scores across five sites, and the spread between them is larger than the difference between most competing products.

The five scores

Read from each site on 30 July 2026.

The third and fourth rows are the first thing to know. GetApp and Software Advice publish not just the same total as Capterra but the same star distribution (135, 78, 24, 12 and 17 reviews from five stars down to one) and the same secondary ratings down to the decimal. All three are Gartner properties drawing on one pool. A comparison page that averages “Capterra, GetApp and Software Advice” is counting 266 reviews three times and calling it corroboration.

That leaves three genuinely independent corpora, totalling 3,424 reviews. They disagree, and the disagreement is the useful part.

Who wrote the 266

Capterra attaches a source label to every review and explains each one in a tooltip. We read all eleven pages of its TestGorilla reviews on 30 July 2026 and paired each label with that review’s rating. Of 266 reviews, 265 resolved.

The weighted mean of those three rows is 4.13, against Capterra’s published 4.1, so this is the whole corpus rather than a convenient slice of it.

In Capterra’s own words, an incentivised review is one where “software users are invited to submit an honest review and offered a nominal incentive for their time and effort”, and a vendor-referred one is where “this reviewer was invited by the software vendor to submit an honest review and offered a nominal incentive as a thank you”. A non-incentivised review is anyone walking in off the street.

None of this is against the rules and none of it means a review is false. Disclosure is exactly what these sites are supposed to do, and TestGorilla is doing what most vendors in this category do. What it means is narrower and more useful: 83% of the corpus behind the 4.1 was written by someone who was asked, and the 17% who were not asked rate it a full point lower. Not one of the 47 reviews the vendor itself referred came in below three stars.

The same pattern is visible on G2, though we could only sample it: G2 paginates client-side and a page parameter returns the first page whatever you set it to. Of the ten most recent reviews on its reviews tab, dated 6 to 28 July 2026, eight carry G2’s own “Incentivized” flag and six came from a seller invite. Ten reviews arrived in 22 days there. Capterra has received one review in all of 2026.

The 4.1 is largely a review of a 2022 product

The same crawl gives the age of every Capterra review.

Seventy-one percent of the corpus was written in 2021 and 2022. TestGorilla changed how credits are consumed in January 2026, and the test library has roughly doubled since the earliest of those reviews. A 2021 reviewer praising or damning the pricing is describing a product you cannot buy today.

This cuts both ways, and it is worth being fair about it. The old reviews understate the library, which has grown. They also predate the contract terms that generate most of the recent complaints. Weight the recent ones and you get a harsher read, not a kinder one.

What the scores agree on

Capterra breaks its rating into four dimensions, and the ranking is consistent across every site that publishes one.

G2’s tag counts say the same thing from the other end. Its five most-cited positives are assessments (52), ease of use (51), candidate management (27), features (23) and hiring efficiency (21). Its five most-cited negatives are assessment issues (23), expensive (16), pricing issues (15), missing features (13) and poor customer support (12). G2’s own summary of its corpus puts it plainly: “some users note that the pricing structure can be rigid, particularly for smaller teams.”

So the corpus is not really split on whether the software works. It is split on what it costs and how it is sold.

The complaint that decides the purchase

Read the low-rated reviews in order and the same situation recurs: a small team buys for one hiring round, discovers the plan is a twelve-month commitment, and cannot get out of it.

An HRM at a company of under twenty people, writing on Capterra on 2 October 2025 and rating it two stars, praises the product first: “the tests are good and features are well developed. Easy to use. Good cheating detection.” Then: “we thought it was monthly, but only afterward learned it was a locked 12-month contract”, and the team “emailed support almost 20 times begging for cancellation or a fair solution, but were only offered a rigid 3-month pause.”

A CTO in human resources, 4 November 2025, one star: “they only offer contracts that are quite pricey. If you’re hiring your first employee, and you actually don’t need this service for a year, customer support is quite disappointed too.”

An HR manager at an IT services firm, 15 January 2025, one star: the annual package auto-renewed, a refund was requested within one to two hours, and was refused.

The interesting reviews are the positive ones that say it anyway. A COO rating it four stars on 23 January 2025: “we probably would have paid for a plan for a month or two - but can’t justify the cost for an entire year for one hire.” A customer success manager, also four stars, 3 February 2025: “the one year subscription only did not allow us to buy the subscription at this time, we would have definitely move forward if the monthly subscription was available.” These are people who liked the product and did not buy it.

On Trustpilot, 8 July 2026: “Useful product but the pay annually plans are extortionately priced for the occasional user.”

Three further Trustpilot reviewers, on 5 March, 24 March and 20 April 2026, allege they were billed after cancelling. Those are allegations by named individuals, we cannot verify them from outside, and none of the three carried a company reply at the time we read the page. They are noted here because a buyer should know they exist, not because we can tell you they are accurate. Trustpilot’s own profile note says the company has replied to 31% of negative reviews.

If you buy, the practical takeaway is unglamorous and specific: get the renewal terms, the cancellation window and the refund policy in writing before the card goes in, and diarise the cancellation date the day you sign.

What the reviewers do not complain about

A fair reading has to include this, because it is most of the corpus.

Ease of use is the single most-praised attribute on both sites, and Capterra’s own analysis finds 98% positive sentiment out of 46 reviews mentioning setup. The breadth of the library comes up constantly. A CCO on 3 February 2025, five stars: “love being able to build my own test with premade test templates, but also that I am able to create my own questions.” Her one complaint is that “so many features are in the premium”, which is the same objection as everyone else’s, pointed at the plan rather than the price.

The free plan is real, and it is the part most comparison pages skip. Ten credits a month, no card, and a skills assessment costs one credit per candidate. For a team assessing fewer than ten candidates a month, TestGorilla can genuinely cost nothing, and several reviewers say the free tier is where they should have stayed.

The price you are quoted depends on how you arrive

This started as a fact-check and turned into a finding worth passing on.

On 30 July 2026, testgorilla.com/pricing gave us different numbers for the same plan depending on how it was loaded.

Reproduced twice on a forced reload. Core is not one price: it is a credit-tier dropdown, and the tiers on offer change with the currency. The gap between the top and bottom rows is close to two to one on cost per candidate assessed, for the same plan name on the same day.

We are not calling that a trick, and there are ordinary reasons a pricing page behaves this way, from regional packaging to a live experiment. But it does mean that any page quoting a single TestGorilla price, including our own breakdown of TestGorilla pricing, is quoting one rung of a ladder. Open the page yourself, in your own currency, and read the credit allowance next to the price rather than the price alone.

The half of the audience the B2B sites do not hear from

Trustpilot’s distribution is the shape that tells you something: 62% five stars, 12% four, 4% three, 2% two, and 20% one star. That is not an average product. That is two populations.

Reading the thirteen most recent reviews there, from 2 March to 15 July 2026, we count five from buyers, seven from people who sat an assessment and one we could not place. The candidate complaints are about length, relevance and what happens when the platform fails mid-test. One reviewer on 15 April 2026 describes an assessment that would not advance past question five, and being told afterwards that the whole submission had been deleted.

Our own earlier research into recruiting forums, collected on 29 July 2026, found the same theme: “Testgorilla tests are lengthy as hell” in one recruiting subreddit, and “total time: 1 hour 30 minutes. Feels designed to filter people on stress tolerance, not skill” in another. That corpus is overwhelmingly candidates rather than buyers, and one of those forums is hostile to assessments in general, so treat it as directional rather than as evidence.

Why a buyer should care: assessment length is a setting you control, drop-off is a cost you pay, and nobody in this category publishes completion rates. The number to look at is your own, measured after your first ten invitations.

So should you buy it

If the conclusion you reach is that the product is fine but the contract is not, that is a different question from this one, and it is the one our TestGorilla alternatives comparison is for: what else exists, what each vendor actually charges, and the four cases where staying is the right call.

Where we fit, and what we are not claiming

SharpAssessment sells practical and job-knowledge tests priced by the candidates you assess, with no annual contract. That is a direct answer to the loudest complaint on this page, which is precisely why you should discount our saying so. We have an obvious interest in making the twelve-month commitment the centre of the story.

What we can do is not repeat the pattern this page documents. There are no customer counts, testimonials, logos or review scores anywhere on this site, and there is no review-site profile of ours to link to. Publishing one before it exists would be the same move as ranking ourselves first on our own comparison table, and this page would not survive it.

Every figure above can be checked in a browser in about twenty minutes. If one of them is wrong, it is wrong in a way you can prove, which is the only standard worth holding a page like this to.