· 17 min read
How to Check If Shopify Store Reviews Are Real (Free)
Shopify has no native review system, so every rating is a third-party app's claim. Nine free checks that grade the evidence in about 20 minutes per store.
Most of a product page is the merchant describing themselves. The review block is the one section that claims to be other people talking, which is exactly why it is the part worth reading carefully — and the part most worth gaming.
There is no way to prove from the outside that a review was written by a real customer who received a real parcel. Anyone who tells you otherwise is selling something. What you can do is grade the evidence behind the claim: who renders the block, when the reviews arrived, how the ratings are distributed, whether the photos appear on other stores, and whether the numbers the page gives Google match the numbers it gives you.
That is what this guide does. Nine checks, roughly twenty minutes per store, no paid account. It is deliberately narrower than the legitimacy checklist, which is about whether a store will ship and refund, and narrower than the product page teardown, which treats the review block as one of twelve conversion slots. Here the review block is the entire subject.
Why Shopify Reviews Are Harder to Read Than They Look
Three structural facts explain most of the confusion.
Shopify no longer has a review system of its own. The native Product Reviews app was removed on 6 May 2024, and no first-party replacement was issued. That means every review you read on a Shopify store is rendered by a third-party app. The merchant can export the data, but the platform holds no canonical review record that you could compare against.
The reviews are not in the product data. Open /products/<handle>.js on any Shopify store and you get a JSON object with price, variants, images, tags, vendor, description, and timestamps — around two dozen fields. There is no rating field, no review count, no review array. I re-ran this check while writing: 26 keys, none of them about reviews. The number above the stars is produced by a script that runs after the HTML arrives.
Apps ship importers. Every mainstream review app can ingest a CSV, and several can pull from marketplace listings or another store’s export. So a store with 800 reviews and a three-week-old domain is not necessarily lying — it may have migrated. And a store that migrated is not necessarily honest — imported reviews describe products the store may never have sold. The point is that domain age and review volume are not comparable signals on their own.
The consequence runs through everything below. The count is a claim. The badge is app-defined. The only hard evidence is internal consistency: dates, distribution, text, photos, and structured data all pointing the same direction.
Five Reads That Take Five Minutes
1. Find the app that renders the block
View source on the product page and search for a review vendor’s domain. The ones you will meet most often are judge.me, loox.io, yotpo.com, okendo.io, and stamped.io. Two or three searches usually identify the one in use, and repeating the search on a collection page or the homepage tells you whether the app is store-wide or only wired to product templates.
Why it matters: the app sets the defaults. Some are photo-first, some text-first, some syndicate to shopping surfaces, and all of them have an import path. Knowing which app it is tells you what to check next — for instance, whether a “verified” label on this particular widget is likely to be tied to an order or is decorative.
It also tells you what changed. If a store’s reviews all stop in early 2024 and resume later under a different widget, you are probably looking at an app migration, not a suspicious gap.
2. Read the date histogram
Sort by newest and read the dates, not the stars. Three shapes are worth naming.
- A steady drip. Reviews arrive most weeks, roughly proportional to catalog size and traffic. This is what an operating store looks like.
- A wall. Dozens of reviews across two or three consecutive days, then silence. That is a campaign, an import, or a batch of incentivised requests — sometimes all three at once.
- A cliff. Hundreds of reviews through one date and nothing afterwards. Either the store stopped asking, or it changed apps and the old reviews did not come across.
None of these is proof on its own. A cliff on a store that is clearly still advertising is interesting, because it means the review programme was abandoned while the acquisition programme was not. A wall on a brand-new domain is nearly always an import.
3. Look at the shape of the rating curve, not the average
A 4.8 average over 300 reviews and a 4.8 average over 12 reviews are different objects. So are a curve with a fat four-star band and a curve that is a single vertical bar at five stars.
Physical products sold at volume produce some unhappy buyers: sizing, shipping damage, colour mismatch, expectations set by renders. A block of 200+ reviews with no one-star and no two-star entries is possible, but it usually means the store is filtering, gating, or importing. The legitimacy checklist treats the same signal as a buyer-protection question; here it is a data-quality question.
Read the negative reviews first, then look for merchant replies. A store that answers a one-star review with a specific remedy is telling you more about how it operates than any five-star paragraph will.
4. Work out what “verified” is verifying
Almost every review app has some notion of an order-linked review, and almost every one labels it differently. The label usually means “this review was submitted through a link tied to an order”, which is decent evidence that a transaction happened — and no evidence at all that the reviewer was not offered a discount in exchange for five stars.
Two things to check. First, whether the label appears on every single review, including ones dated before the store launched. Second, whether the store explains its method anywhere on the page or in a linked policy. A store that says “we ask every buyer and we do not filter by rating” is making a claim you can hold it to. A store that only prints a green checkmark is not.
5. Reverse-image-search one five-star photo
Pick a photo review that looks like a phone snap rather than a studio shot, crop the product, and run it through a reverse image search.
If the photo appears on other stores selling the same product under different review text, you have found an import pool and the review was not written for this store. If the photo appears nowhere else, including the brand’s own site, that is weak positive evidence — about as close to “a real person took this” as an outside observer gets.
Do not over-read the absence of a match. Plenty of genuine customer photos are indexed nowhere.
Four Reads That Take Ten Minutes
6. Search the text, not the sentiment
Take one distinctive sentence from a five-star review — something with a specific detail, not “great quality” — put it in quotes, and search. Then search for the same sentence inside the store’s own product description.
Repeated review text across stores is the clearest import fingerprint there is, because importers copy strings verbatim. Weaker but still useful: template phrasing with one variable swapped (“Amazing quality, fast shipping, will definitely order again” appearing eleven times under different names), and reviews that read like translated listing copy rather than a person describing an unboxing.
7. Compare review volume to catalog size
Count the SKUs, then count the reviewed SKUs. If a store has forty products and 90% of its reviews sit on two of them, the review programme is not running across the catalog. It is concentrated on the hero products, which is normal for a store testing traffic — and also exactly what a hand-curated review block looks like.
The more useful ratio is reviews per reviewed SKU, read over time. A hero product that gains twenty reviews a month is being sold. A hero product that has gained none in six months while the store keeps launching ads is a different story.
8. Check whether the same reviews live on a second store
If you research a niche you will meet the same product on several stores. Compare review text across two of them. Identical reviews with different dates, or the same reviewer name on both, tells you that at least one store did not collect them.
This is also how you catch the copycat pattern. The second store is not running its own review programme at all; it copied the first store’s widget data through an export and never asked anyone anything.
9. Compare what the page tells Google with what it tells you
Open the page source and look for aggregateRating in the JSON-LD. Three things can be true.
- Clean. The rating and count match the visible widget, and the block is nested inside the
Productobject. - Stale. The numbers do not match the widget. Usually a cached value the app refreshes on its own schedule — a symptom of neglect rather than fraud.
- Detached. The rating sits in a standalone block with
itemReviewed, separate from theProduct. I meet this pattern often in the wild. It is valid schema in isolation, but Google’s product-snippet guidance expects the rating inside the product entity, so the stars may never show in search results. It tells you the structured data was assembled by whoever installed the app and never audited afterwards.
None of these three prove anything about authenticity. They tell you how carefully the store maintains its own claims, which is a useful prior for everything else you read.
Read the Block’s Position Too
Where the block sits is part of the message. A review section with a star summary next to the buy button, a visible count, and an anchor in the page nav is being presented as evidence. The same reviews behind a collapsed tab at the bottom of a long page, with no star summary near the price, are being presented as a formality.
Neither arrangement is dishonest. But when a store’s reviews are excellent and buried, that is usually a theme or app configuration problem rather than a strategy — and it tells you the merchant is not treating the block as a conversion asset, which is itself information about how the store is run.
Three Profiles, Not a Score
Nine checks produce a lot of noise. In practice the reads cluster into three profiles.
| Profile | What you see | How to read it |
|---|---|---|
| Operating | Steady dates, mixed ratings with a few 1–2 star entries, merchant replies, photos that do not appear elsewhere, JSON-LD roughly matching | A real but biased sample. Bias runs toward buyers who open email. |
| Migrated | A wall of reviews at one date, consistent text style, labels that do not separate imported from earned, widget changed at some point | Content may be genuine but was not collected here. Weight post-migration reviews much more heavily. |
| Curated | All five stars, no negatives, template text, photos reused across stores, counts that never move | Treat the block as marketing copy. It tells you what the store wants you to believe. |
Most stores are a mix. Running an honest programme on two hero products while importing reviews for the long tail is the most common combination I see, and it is invisible unless you check per product instead of per store.
A Worked Example (Hypothetical)
Two stores, same category, same price band.
Store A: 4.9 average, 412 reviews, all dated within nine days of each other, forty with photos, and the same three sentences appearing in eleven reviews. The widget is one of the big apps, the JSON-LD count matches, and the merchant has replied to none of the four negative reviews.
Store B: 4.4 average, 168 reviews spread over fourteen months, a visible band of three-star reviews complaining about sizing, merchant replies on most negatives, and photo reviews that match nothing in reverse image search.
Both could be honest. Only one gives you data you can reason with. If I had to model a competitor’s offer from these two, I would take Store B’s complaints as the useful input — the sizing complaint is a product fact, and it is the kind of thing that decides whether a similar offer works. Store A’s block cannot tell me anything I could not have guessed from the product photos.
The numbers here are invented to make the pattern legible. The pattern is not.
What the Rules Require, and What That Does Not Prove
Two regulatory changes are worth knowing, because they explain why some blocks look the way they do.
In the United States, the FTC’s Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465, became fully effective on 21 October 2024. It prohibits fake or false reviews, buying reviews conditioned on sentiment, reviews written by company staff without disclosure, company-controlled “independent” review sites, review suppression through threats, and misrepresenting that a displayed set represents most or all reviews submitted when others were hidden by rating. It carries civil penalties.
In the EU, Directive (EU) 2019/2161, applied from 28 May 2022, added review practices to the list of commercial practices that are unfair in all circumstances: claiming that reviews come only from consumers who actually used or bought the product without taking reasonable steps to verify that, and publishing fake reviews.
The practical point for a researcher is that both regimes regulate conduct, not appearance. A compliant store can have a thin review base. A store with an immaculate-looking block can be violating the rule in ways you cannot see — the suppression provisions describe exactly the thing an outside observer cannot detect. Use the rules as context for why a store might disclose its method, not as a detection tool.
What This Cannot Tell You
- Whether a specific review is fake. You can find imports and templates. You cannot adjudicate a single paragraph.
- How satisfied buyers are. Reviews measure who was asked and who answered, not the population. A store that emails every buyer and a store that emails only happy ones can land on the same average.
- Whether reviews were suppressed. Hidden reviews leave no public trace. This is the largest blind spot, and no free method closes it.
- Anything comparable across categories. Review propensity varies with price, product type, and how the product is used. A $12 accessory and a $400 piece of furniture do not generate reviews at the same rate, so cross-category comparisons of reviews per SKU are meaningless.
- Sales volume. Review counts are not a sales estimate, and treating them as one is the most common error in this kind of research.
The 20-Minute Routine
- Two minutes: view source, identify the review app, note whether the widget appears to have changed.
- Three minutes: sort by newest, write down the date shape and the last review date.
- Three minutes: count one-star and two-star entries, and check whether the merchant replied.
- Three minutes: open three products, compare counts, note concentration.
- Three minutes: quote-search one distinctive sentence and reverse-image-search one photo.
- Three minutes: find
aggregateRatingin the JSON-LD and compare it with the visible count. - Three minutes: write one line — profile, confidence, and what would change your mind.
Then stop. The marginal value of a tenth check is small compared with the value of running the same nine on the next store.
If you are doing this across a list rather than a single URL, the free store analyzer supplies the catalog and app context to pair with these reads, and the app-detection guide goes deeper on the fingerprinting step. Stores that fail the review read often fail the dropshipping test for the same underlying reason: nobody is minding the detail.
Verified 2026 · 20% off Starter, Pro and Business for your first 3 months
Reading nine signals by hand works for one store and falls apart at ten, which is the honest argument for the paid layer: not “find the fake reviews”, but keeping a consistent read on many stores at once. The database covers 1.5M+ Shopify stores and 300M+ total ads, and TrendTrack promo code is WINNING - 20% off all plans for the first 3 months if the manual version has done its job.
If You Own the Store
The read is symmetric, so it is worth knowing what yours looks like.
Keep dates visible. Show the full distribution rather than a five-star carousel. Reply to negatives with a specific remedy instead of a generic apology. State your method in one sentence near the block — whether every buyer is asked, and whether reviews are filtered by rating. Label imported reviews as imported. Do not gate review requests by sentiment; beyond being a bad look, the US rule explicitly covers misrepresenting a displayed set as most or all reviews when others were withheld by rating.
What does not help: a wall of five-star reviews dated the week the store launched, stock photos in the photo-review carousel, or a verification badge applied to everything. Each of those is checkable in about two minutes by anyone who reads this page.
FAQ
How can I tell if Shopify reviews are fake?
You cannot prove a single review is fake from outside, but you can grade the block. The strongest signals are a wall of reviews on one date, template text repeated with one word changed, the same photo appearing on several stores, and a rating curve with no negative entries at all. Three or more of those together means treat the block as marketing copy rather than evidence.
Why do some Shopify stores have thousands of reviews on day one?
Almost always because the reviews were imported. Every mainstream review app can ingest a CSV, and several can pull from marketplace listings or another store’s export. A merchant migrating from another platform, or from a previous store, brings the history with it. That does not make the reviews false, but it does mean they were not collected by this store.
Does Shopify have its own review system?
No. The native Product Reviews app was removed on 6 May 2024 and no first-party replacement was issued, so every review block you see on a Shopify store is rendered by a third-party app. Two consequences matter here: the platform holds no canonical review record, and the review data is not present in the product JSON endpoint at all.
Is a verified buyer badge proof that a review is real?
No. The badge usually means the review was submitted through a link tied to an order, which is reasonable evidence that a transaction happened. It says nothing about whether the reviewer was offered a discount for a five-star rating, and imported reviews sometimes carry the same badge. Check whether the label appears on reviews dated before the store launched.
Can I see reviews a store has hidden?
No, and this is the largest blind spot in the whole method. Suppressed reviews leave no public trace, and the US rule on review suppression specifically describes the conduct an outside observer cannot detect. You can only note whether a store discloses its filtering method, and treat silence on that point as a missing signal rather than a negative one.
Do review counts tell me how many sales a store makes?
No. Review counts measure how many buyers were asked and answered, which depends on the app, the email sequence, and the product category far more than on volume. The relationship between reviews and orders is not stable enough to convert one into the other, which is why sales estimates built on review counts tend to be wrong by wide margins.