Trustpilot Competitor Analysis: Read the Transparency Page

Most people arriving here want to read a rival’s reviews, which is the least interesting thing this platform offers. Reviews are opinions from a self-selected group, and their value in competitive work is real but limited and well understood. The distinctive asset is a second page attached to every claimed profile, which documents not what customers think but what the business itself has been doing: how it gathered its reviews, how it answered the bad ones, and how often it asked for them to be taken down.

That distinction is worth holding on to because it changes what kind of evidence you are handling. Review text is testimony. The transparency record is behaviour, observed and published by a third party with no stake in flattering the business. Behaviour is harder to manufacture and easier to compare across companies, which is why a competitor’s conduct record is more useful than their star rating in almost every situation where you might reach for either.

The limits are equally clear. This is a consumer platform. It is strong on self-serve products, on billing and cancellation, on delivery and support, and on anything where the buyer and the user are the same person. It is weak on considered business software, where purchases involve committees, deployments run for months, and almost nobody writes publicly afterwards. Judging a rival’s enterprise product by its consumer reviews will produce a confident answer to a question you did not ask.

1. 1Open the transparency page before the reviews. Every claimed profile has a companion page describing how the business has used the platform over the previous twelve months. Reading it first changes how you read everything else, because it tells you what kind of corpus you are about to look at rather than leaving you to guess from the text.
2. 2Split the review sources into solicited and unsolicited. The page counts organic, verified, invited and redirected reviews separately. A competitor whose reviews are overwhelmingly invited has a rating that reflects their collection programme as much as their service. A competitor with almost no invitations has a rating shaped by whoever felt strongly enough to arrive unprompted.
3. 3Read the star distribution by source and take the gap. This is the single most useful number on the platform. The same business is scored separately by the people it asked and the people it did not, and the difference between those two distributions is a measurement of invitation bias for that specific competitor rather than an assumption about the category.
4. 4Read the flagging record, including the outcomes. The page reports how many reviews the business reported to Trustpilot, at which ratings, and how those cases were resolved, including how many were found not to breach the guidelines. A pattern of flagging one star reviews that are then judged legitimate describes how a competitor responds to criticism.
5. 5Check the reply behaviour on negative reviews. The share of negative reviews answered, and the date of the most recent reply, are both published. This is one of the few places you can observe a competitor's support posture as a measured fact rather than infer it, and it is a fair thing to know about a rival you meet in deals.
6. 6Discount the corpus using the platform's own published contamination rate. Trustpilot publishes an annual measurement of how many fraudulent reviews it removed and what share of submissions that represented. Apply that rate as a rough discount on any conclusion drawn from raw volume, and take it as a reason to weight long specific reviews over short enthusiastic ones.
7. 7Record the four numbers with a date every time you check. Score, total reviews, the invited share and the flagging count. The transparency page covers a rolling twelve months, so evidence of past behaviour leaves it on a schedule. A dated row is the only way to keep a record of what a competitor was doing two years ago.

Every claimed business profile has a companion page describing how that business has used the platform over the previous twelve months. It is public, needs no account, and applies to your competitors exactly as it applies to you. Very little competitive writing mentions it, which is strange, because it is the most unusual thing this source does.

The last row is the one that repays the most thought. Flagging genuinely rule-breaking content is legitimate, common and expected, and a business with a handful of flags is doing normal housekeeping. What the page lets you see is the shape of the pattern: the volume, the star ratings targeted, and crucially the resolution, including how many reviews remained online because they were found not to breach the guidelines. A rival who repeatedly reports one star reviews that turn out to be valid is telling you something about how they respond to criticism, and it is a fact rather than an inference.

Every discussion of review data eventually arrives at the same objection: the sample is self-selected, so the rating is unreliable. That objection is usually correct and almost always unquantified. Here it can be quantified, on a competitor, for free, in about a minute.

The mechanism is simply that the star distribution is published per review source. Solicited reviewers and unsolicited reviewers are scored separately on the same profile, for the same business, over the same period. Subtract one from the other and you have that competitor’s invitation gap as a measurement rather than as a general worry about review platforms.

Trustpilot’s own framing supports reading it this way rather than as evidence of misconduct. It states that invited reviews are subject to the same fraud detection and moderation as any other review, and that consumers can always leave a review with or without an invitation. The rules prohibit influencing what a reviewer says, offering incentives for reviews, and inviting only the customers a business expects to be happy. So a large invitation share is a sampling fact and not an accusation, and it is the sampling that produces the gap.

One further number belongs in this calculation. Trustpilot publishes an annual measurement of fraudulent content in its own corpus, which is an unusual thing for a platform to do. In the Trust Report published on 29 May 2025 it reported removing 4.5 million fake reviews during 2024, amounting to 7.4 per cent of everything submitted that year, with 90 per cent of those removed automatically by its detection systems and 53 per cent more removed automatically than in the previous year. It gave 2023 as 6.1 per cent of submissions, and put the platform at over 300 million reviews with 61 million published during 2024. That is a contamination rate rather than a sampling bias, and the two stack: apply the published rate as a rough discount on raw volume, then apply the invitation gap to the rating itself.

Ordered by what a competitive researcher gets from each, which puts the review text considerably lower than most readers expect.

1. The review source split

The first thing to open and the fastest context you will get. Four counts covering twelve months tell you whether you are about to read a curated corpus or a spontaneous one, and every later judgement depends on knowing which.

2. The star distribution by source

The single most valuable figure on the platform, because it converts the standard complaint about review data into a measured quantity for one named competitor. Take both distributions, note the gap, and record it beside the headline score so the score is never quoted alone.

3. The flagging record and its outcomes

Volume, the ratings targeted, and how the cases resolved. Read the resolution column rather than the count, since a business flagging plenty of content that genuinely breached the rules is behaving properly and a business flagging valid criticism repeatedly is doing something else.

4. Reply behaviour on negative reviews

The proportion answered and the date of the most recent answer. This is one of the few places where a rival’s support posture is observable as a published figure rather than inferred from anecdotes, and it is a fair and checkable thing to know about a competitor you meet in deals.

5. The replies themselves

Once you know how many there are, read a dozen. Sort them into three piles as you go: faults the company owns outright, faults it hands back to the customer, and faults it says have since been resolved. The third pile is the valuable one, because each entry is a dated public undertaking that can be tested against what customers report six months later.

6. The claimed profile record

When they took control and what they verified to do it. A recently claimed profile beside a sudden rise in invited reviews is a reputation programme starting, which is a marketing event worth dating.

7. The one and two star reviews, read for specifics

Filter to the bottom two ratings and read for repeated concrete failures rather than for tone. Three reviewers independently describing the same billing behaviour is a finding. Thirty reviewers expressing general displeasure is a mood.

8. The recent five star reviews, read for what they praise

More useful than they look. What satisfied customers single out is a competitor’s real differentiator as experienced rather than as marketed, and it is frequently something their own positioning barely mentions.

9. The TrustScore beside the review count

The score is a computed figure rather than the arithmetic mean of the stars on the page, and the full formula is not published. Since the underlying counts are published, lean on those. Never compare two scores across businesses without the volumes beside them.

10. Platform notices on the profile

Rare and unmistakable. Where the platform has placed a consumer warning it hides the score and limits account functionality, which makes it the loudest signal available here. Record it, and think carefully before repeating it, since it describes conduct on a platform rather than the quality of a product.

11. The category and similar-company listing

A consumer-facing grouping rather than a software taxonomy, which occasionally places a rival beside businesses you would never have considered comparable. Worth a glance for exactly that reason, because it shows the frame a general audience puts them in.

The research route is free and needs no account

Profiles, review text and filters, the score, the review counts and the entire transparency record are all public. There is no gate on any of the techniques described on this page, and none of them require you to identify yourself or to have any relationship with the platform.

What the paid business account adds

A competitor view inside the business product lets you nominate rivals and chart their rating and review volume by month alongside your own, compare sentiment across a limited set of topics, and export the charts as an image or a spreadsheet. It is a convenience if you already pay for the platform, and it is the thing most articles on this subject describe. Worth saying plainly: a hand-picked set of rivals on a single platform covers less ground than the public route, which applies to any business and asks nobody for permission.

The one thing the free route cannot give you

History. The activity record covers a rolling twelve months and drops what falls outside it, and the paid product does not restore what the free one has already forgotten. There is no archive of a competitor’s past invitation behaviour or past flagging record anywhere.

What to record on each visit

Four numbers and a date: the score, the total review count, the invited share, and the flagging count. They take under two minutes and they are the only way you will ever be able to say what a competitor was doing eighteen months ago, because by then the page itself will not know.

Reading is straightforward. The whole platform exists so that people comparing businesses can read what other customers said. Using what you read is where the care is needed, and the risks here are shaped differently from the rest of this cluster because the material is testimony about a named company.

• Repeating a complaint turns it into your factual claim. A reviewer describing a bad experience is stating their experience. Your company reproducing it in sales material is asserting something about a competitor’s business, which brings comparative advertising and trade libel rules into play rather than anything about reviews.
• A few angry reviews will not support a general claim. Building a statement about a rival’s quality on three critical entries is weak evidence and real exposure at the same time. If you intend to say something about a competitor’s service, use the published aggregate figures, cite the volume behind them, and date them.
• Never write a review of a competitor or arrange for one to be written. Fake reviews are prohibited outright and are treated as illegal in the platform’s own terms, and enforcement escalates to account restrictions, public warnings that hide a score, termination, legal action and passing information to authorities. Nothing about it works out well, in any variation, for anyone who has tried it.
• Quote reviewers accurately or not at all. Trimming a quotation so it says more than the reviewer did is the most common way an honest comparison becomes a misleading one, and the original is public, dated and one click away for whoever you showed it to.

• Anything about a considered business purchase. Committee decisions, procurement, security review and multi-month deployments produce almost no consumer reviews. Proxy: the software review platforms and your own buyer conversations.
• What a rival was doing more than a year ago. The behavioural record covers a rolling twelve months and then forgets, with no archive anywhere. Proxy: your own dated log, started today.
• The size of their customer base. Review volume measures how hard a business asks, and the asking varies by an order of magnitude between companies. Proxy: filed accounts, named logos and disclosed counts.
• Whether the complaints are representative. You can measure the invitation gap, which is a real advance, and you still cannot see the silent majority who never reviewed at all. Proxy: your own win and loss conversations with buyers who evaluated them.
• How the score itself was produced. The formula behind the computed figure is not published, unlike the methodologies the software directories put out. Proxy: the underlying counts and distributions, which are published and are what you should be quoting anyway.

Quarterly is the right cadence for the numbers and it should never be skipped, because this is the one source in this cluster where missing a period destroys evidence rather than merely delaying it. Record the score, the total review count, the invited share and the flagging count, with the date, for each rival. Four fields.

Between those checks, read rather than count. Twenty minutes on the recent one and two star reviews for the competitor you meet most often in deals will surface operational failures months before they appear in any analyst commentary, and the failures are usually specific enough to use directly.

Reopen off-cycle on three triggers: a sudden change in review velocity, which almost always means an invitation programme started or stopped; a platform notice appearing on a profile; and a competitor changing how they bill, since billing changes reach this source faster than they reach anywhere else.

This source has a property none of the others in the cluster share: it deletes its own evidence on a schedule. The activity record covers a rolling twelve months, so a competitor’s invitation campaign, their flagging behaviour and their reply rate from eighteen months ago are not archived, not purchasable and not recoverable. The page is not a history that grows. It is a window that slides, and everything behind it is simply gone.

So the value of looking is almost entirely in having looked before. One visit tells you what a rival is doing now, which is rarely the thing that informs a decision. Whether their invitation share doubled, or their reply rate collapsed, is knowable only if somebody wrote down the earlier state while it was still on the page. Human diligence is badly suited to that, since the reward arrives a year after the effort and nothing marks its absence in the meantime. Accumulating that history without anyone having to remember is competitive intelligence software in its least glamorous form, and it is what Flares is doing while nobody watches, keeping each earlier state so a reputation programme starting or stopping appears as a trend instead of a page somebody happened to open twice. A reply rate falling can mean a support team is overwhelmed, indifferent, or halfway through being rebuilt, and no system will tell you which. Reading it correctly takes somebody who knows the company.