Reading reviews as a churn report
Every prop firm can read its own reviews, and it is tempting to read them as a churn report: count the one-star complaints, sort them by cause, and treat the result as the reasons traders leave.
Some of those reasons are there. A trader whose payout stalled, whose account closed over a rule they dispute, or whose identity check never cleared can say so in public, and almost all the reviews we read were complaints of that kind. A firm cannot size the behavioural side of its churn from the same page: the trader who sized up after a loss, kept trading through a bad session, and lost the account.
What we read
In October 2026 we read the most recent one- and two-star Trustpilot reviews of 10 prop firms: 251 reviews in English, published in the previous 12 months, newest first and up to 70 per firm. How far back each read reaches differs by firm, from a full year to a single day.
The rule was strict. A review counted as behavioural only when its writer described their own behaviour driving the loss or the breach: trading on frustration, sizing up to win a loss back, breaking their own plan under pressure. A loss on its own did not count.
Under that rule, 1 of the 251 counted. All 251 were then re-read, and every behavioural count confirmed, in a second, separate read under the same rule. The count, its rule and its sources sit with the rest of our figures on the churn statistics page.
What the rest were about
Almost every other review was about the firm: disputed breaches, payouts, identity checks, platform problems. Some quoted the firm's own findings about their trading and disputed them. These complaints deserve a firm's attention, and most describe problems a firm can usually also see in its own tickets and logs.
Why the count moves with the rule
A review can give more than one cause, or none. A trader who breached in a volatile session might blame the firm's rule, the platform, or nobody at all, and whoever reads the review has to decide what it is about. So the count depends on the rule.
A looser rule, one that also counts a trader's emotional reaction to the firm's rules or a repeat purchase after a failure, finds more. The earlier version of this post rested on a looser count, and our own April records show how unstable such a count is. In April 2026 we counted one prop firm's negative reviews twice, 5 days apart: first under a rule we had not written down, then on a fresh sample under a written rule. The two counts disagreed on how many reviews were behavioural and on which pattern came first.
Who writes a review
There is a second limit, and no counting rule fixes it. A review is written by whoever chose to write one, about whatever they chose to put in it. It is not a record of every trader who left, or of why each one left.
Where the behaviour is recorded
The behaviour behind behavioural churn happens in the trades. Each pattern, such as a larger order seconds after a loss or five trades in 15 minutes, sits in the order history with a timestamp, before the account closes and before anyone writes a review.
That record belongs to the firm, and it is where the behavioural side of churn can be counted. Four of the five churn warning signs we set out are read from it. It is also where Discentra works: it watches the trade feed for these behavioural triggers and places a coaching call to the trader within 5 seconds, before the next order.
What a read of reviews is still good for
Reviews show which complaints recur, whether they are growing, and what a trader reads about the firm before buying a challenge. That is worth knowing on its own terms.
If you want it for your firm, we will read your most recent one- and two-star reviews for free, one by one, under the same rule, and send one page: what they show, and what they can't. For the side reviews can't show, the Behavioural Diagnostic runs on one cohort of your own trade data. Both are set out on what reviews can't show you.
Method and limits
- Sample: Trustpilot one- and two-star reviews in English, published in the 12 months to 5 October 2026, newest first, up to 70 per firm, for 10 prop firms: 251 reviews. The firms are not named.
- Coverage: differs by firm, from a full year to a single day.
- Rule: behavioural only where the writer describes their own behaviour driving the loss or the breach. A loss on its own does not count.
- Check: all 251 re-read, and every behavioural count confirmed, in a second, separate read under the same rule.
- What it is not: a share of churn, an industry rate, or a judgement on any one firm.
