How TrustScore Works: Weights, Recency and Verified Reviews

TrustScore is the 0-to-5 score every reviewed business gets on TheUserOpinion. Here's how TrustScore works in one sentence: it blends five signals from published reviews, weighting the average star rating at 40%, review volume at 20%, recency at 20%, owner replies at 10% and verified reviews at 10%. That's the whole recipe. No hidden factors, no paid boosts and no editor nudging the number.

This post opens up the TrustScore section of our complete guide to trusting online reviews. We'll show the actual formula, work through real numbers and point out what the score can't tell you, because a rating site that hides its own math has no business asking you to trust anyone else's.

Who This Explainer Is For

It's for shoppers who see "Trust Score 3.4" beside a 5.0 star rating on a profile and wonder which number to believe. It's also for business owners who want to know what actually moves their score, and for anyone comparing our method with Trustpilot's, which shares the name but not the math.

It isn't for anyone looking for a shortcut. Every lever in the formula is tied to real reviews from real customers, and the shortcuts people try, like buying reviews or only inviting happy customers, break our review guidelines.

How TrustScore Works, Signal by Signal

Each signal is turned into a value between 0 and 1, multiplied by 5 so it sits on the same scale as stars, and then weighted. The parts add up to a number between 0 and 5. Here's what goes in:

For anyone who prefers code to tables, this is the calculation our database runs, written out in plain terms (from the calculate_trust_score() function, checked November 10, 2026):

TrustScore = (average rating × 0.40)
           + (min(ln(reviews + 1) ÷ ln(101), 1) × 5 × 0.20)
           + (min(reviews this month ÷ max(reviews ÷ 12, 1), 1) × 5 × 0.20)
           + (reply rate × 5 × 0.10)
           + (verified share × 5 × 0.10)

Every input counts published reviews only. A review that a moderator removes stops counting the moment it comes down, and so does one its author deletes.

Average Star Rating (40%)

This is the plain mean of every published star rating. A profile averaging 4.5 stars gets 1.8 points from this part (4.5 × 0.40), and a perfect 5.0 gets the maximum of 2.0. There's no weighting by reviewer or by age here. A one-star review from last week and a five-star review from last year count the same toward the average.

That's less of a problem than it sounds, because the profile also shows the raw star average and the star breakdown, so you can always see satisfaction on its own.

Review Volume (20%)

Volume rewards a bigger sample, but with diminishing returns. The formula uses the natural log of the review count, scaled so that 100 reviews earns full credit:

The log curve is the important part. Going from one review to 10 adds more than going from 50 to 100, because the first handful of reviews tells you far more than the next 50. After 100, extra reviews don't move this part at all, so a business with 5,000 reviews doesn't steamroll one with 150.

In plain English: under about 25 reviews, a big chunk of the score is still missing for every business, however good its stars are. As of November 10, 2026, no profile on TheUserOpinion has reached 100 published reviews yet, so nobody here gets full volume credit today.

Recency (20%)

Recency asks whether reviews are still coming in. It divides the number of reviews published this calendar month by the business's average monthly pace, which the formula estimates as its total reviews divided by 12. If this month matches or beats that pace, the business gets the full point.

Two details matter in practice:

• Small profiles get an easy bar. The pace is never set below one review a month, so any business with fewer than 12 reviews earns full recency credit from a single review this month.

• Big profiles need a steady flow. A business with 300 reviews has a pace of 25 a month. If only 10 arrive this month, it earns 40% of the recency point, or 0.40 instead of 1.00.

"This month" means the calendar month, so the count resets on the 1st. And the score is recalculated when something happens on the profile (a review is published, edited or removed, or a reply is added or deleted), not on a timer. A quiet profile's recency part reflects the last time anything changed. Our guidelines put it the same way: TrustScore is recalculated whenever a review is published, removed or answered.

Owner Replies (10%)

This is the share of published reviews that have a reply from the business. Reply to every review and you get the full 0.5 points. Reply to half and you get 0.25. The content of the reply doesn't change the math. Readers do see it, though, so a calm, specific answer is worth more than the 0.5 points suggest.

Verified Share (10%)

A review counts as verified on TheUserOpinion only when it arrives through a personal invitation. Per our guidelines, the business creates an invitation for the customer's email address, the customer opens a link that works once and expires after 14 days, and they write the review while signed in with that same confirmed email. The business owner and addresses on the business's own domain can't use an invitation.

Verified describes how a review arrived, not whether it's positive. A verified one-star counts toward the verified share exactly like a verified five-star. Ordinary reviews follow the same rules and count the same toward the star average.

Worked Examples

Formulas are easier to trust when you can check them. The first row below is a real profile, StelsVPN, as stored on November 10, 2026: six published reviews, all five stars, one this month, no replies and none verified. The formula gives 2.00 + 0.42 + 1.00 + 0 + 0 = 3.42, which is exactly the score in our database and the 3.4 shown on its profile. The other rows are example numbers we picked to show how the parts trade off.

Scores rounded to two decimals. Profiles show one decimal.

Look at the small, well-liked profile. Three five-star reviews and a quiet month leave it at 2.30, below the brand-new profile with a single review posted this month. That isn't a judgment that the business got worse. It's the formula saying "we haven't heard much lately".

What TrustScore Doesn't Tell You

This sounds good on paper, but every score has blind spots, and ours are worth knowing before you lean on it.

It isn't a satisfaction score. Stars carry 40% of the weight. The other 60% measures how much evidence there is, how fresh it is and how the business engages. Run the numbers on a made-up business with 200 reviews averaging one star, a steady flow, replies to everything and every review verified: it lands at 3.40, roughly level with the six-review profile above. Nobody should read those two the same way. That's why every reviewed profile shows the star average and the star breakdown right beside TrustScore.

It can't read. The formula counts and averages. It doesn't weigh what reviews say, so a 4.2 full of specific complaints about billing and a 4.2 full of praise score the same. Read the newest reviews yourself, and use our checklist on how to spot fake reviews when a page feels too tidy.

It's thin on small profiles. With fewer than about 25 reviews, a single new review moves the volume and recency parts a lot. Treat a TrustScore built on six reviews as a first impression, not a verdict.

It doesn't decay older stars. Unlike Trustpilot's method, the star average treats a three-year-old review the same as yesterday's. Recency only measures whether reviews are still arriving, not whether recent ones are better or worse. Sort by newest when the date matters to you.

How TrustScore Compares With Trustpilot and Amazon

The two methods people most often confuse with ours work differently on almost every point. Trustpilot is a separate review platform that also calls its score TrustScore. Amazon calculates star ratings for products.

Sources checked November 10, 2026: Trustpilot's help article TrustScore and star rating explained and Amazon's help page How Are Product Star Ratings Calculated?.

Trustpilot says its score gives the most weight to the newest reviews, uses a Bayesian average that adds the value of seven 3.5-star reviews to every calculation, and doesn't distinguish between verified, invited and unprompted reviews. Amazon says it uses machine-learned models instead of a simple average, weighing factors like recency and Verified Purchase status, and leaves ratings without Verified Purchase out of the overall star rating unless the customer adds text, an image or a video.

Our take: each approach has a real trade-off. Trustpilot's built-in reviews keep one early five-star from producing a perfect score. Amazon's model can weigh signals a simple formula can't. Ours gives up some of that cleverness in exchange for a formula anyone can check with a calculator, which is the point of a score called TrustScore. The catch is that ours rewards activity heavily, so the star average has to be read with it.

How Businesses Can Raise TrustScore the Honest Way

Every lever is tied to real customers, so the honest playbook is short:

1. Invite every customer, not just the happy ones. Invitations are the only route to verified reviews, and they're the most reliable way to add volume. Screening for happy customers breaks our guidelines and makes the page less believable.

2. Keep a steady pace. Recency rewards reviews that keep arriving, so a monthly invitation habit beats a single campaign.

3. Reply to reviews, including the bad ones. Replies are 10% of the score and far more of a reader's impression. Be specific, fix what you can and skip the copy-paste apology.

4. Fix the product. Stars are still the biggest single part, at 40%.

What doesn't work: paying for reviews, offering anything in return for a review or for changing one, reviewing your own business or a competitor's, or pressuring reviewers. Our guidelines ban all of these, and no business can pay us to change its score. If you run a business that isn't listed yet, you can list or claim your profile for free and start sending invitations.

The Bottom Line

How TrustScore works comes down to one sentence: 40% stars, 20% volume, 20% recency, 10% replies and 10% verified, all from published reviews, with no paid adjustments. It's useful for telling a well-documented, active business from one with a handful of old reviews.

Best use for shoppers: read TrustScore together with the star average and the newest reviews, especially on profiles with fewer than 25 reviews. Best use for businesses: invite every customer, keep a monthly rhythm and reply to everything. The main warning: a high TrustScore means plenty of fresh, engaged evidence, not automatically happy customers.

If you've used a tool listed here, find it and write a review. Every honest review makes the score more meaningful for the next person, and browsing a category like VPN user reviews is a good place to see the score side by side.