Shopify Product Reviews: The Complete Guide

A shopper is on your product page. They have read the description, looked at the photographs, and they still do not know the one thing they actually want to know: does this hold up.

Nothing you write can answer that. You are the seller. Every adjective on the page is one you chose, and the shopper knows it.

Shopify product reviews are the only content on the page written by someone with nothing to gain, which is precisely why they carry weight that your copy never will. This guide covers getting them written, getting photos into them, keeping them inside Google's rules, and handling the bad ones.

What do product reviews actually do?

Three separate jobs, and it helps to keep them apart, because a tactic that serves one can undermine another.

• They resolve doubt at the decision point. The specific doubt is usually not "is this good" but "is this right for me": sizing, durability, whether the colour matches the photo. Reviews answer those far better than a description can.
• They set expectations, which reduces returns. A review that says a jacket runs small does more to prevent a return than any size chart.
• They put text on the page that nobody at your company wrote. Real customer language often matches how people actually search, in ways your product copy does not. Your description says "moisture-wicking merino base layer"; the review says "didn't stink after three days hiking", which is closer to what someone actually types.

That first job is why placement matters as much as volume. A hundred reviews on a tab at the bottom of the page serve nobody. Ten reviews next to the add-to-cart button serve everyone.

Where reviews belong on the page

Roughly in this order of value:

• The star rating immediately under the product title. This is a summary, not a review, and its job is to stop the shopper wondering whether reviews exist at all.
• Two or three reviews above the fold on mobile, near the buy button. Not the newest, the most useful. A review that mentions sizing outperforms one that says "love it".
• Filters by rating and by attribute. Shoppers who want the critical reviews should be able to find them in one tap; making that hard is how a page loses the sceptics it was meant to convince.
• Photos as a strip. Customer images get looked at far more than customer text, and a strip of them communicates volume before anyone reads a word.
• The full list further down. Somewhere, but not as the only place reviews exist.

The pattern to avoid is the tabbed panel below the fold labelled "Reviews (0)", which manages to cost you a conversion and advertise the absence at the same time.

How many reviews do you need?

You will find a lot of confident numbers for this online. Almost all of them trace back to a vendor's own blog post citing another vendor's blog post, so I am not going to add another one.

What is reliably true is the shape of the curve, not the coordinates on it.

• Zero to one is the biggest single jump. A product with no reviews looks untested. A product with one looks like something a person has actually received.
• The first handful do most of the work. Going from three reviews to ten changes how a page feels far more than going from fifty to a hundred.
• Recency competes with volume. A product with six reviews from this year often reads better than one with sixty from three years ago, because the shopper is trying to work out whether the current version is any good.
• Coverage beats concentration. One product with 200 reviews and nine with none is a worse position than ten products with twenty each, because most of your catalogue is still unproven.

So the useful target is not a number per product. It is: get every product off zero, keep the reviews current, and stop optimising for a leaderboard.

Why don't your customers leave reviews?

Not because they did not care. Because of one of these, and each has a fix.

• You asked at the wrong moment. A request that arrives when the order ships is asking someone to review a parcel they have not opened.
• You asked for too much. A link to a login page, then a form with six fields, then a captcha. Every step loses most of the people who survived the previous one.
• You asked once. A single email that landed during a busy week is not a campaign.
• There was nothing in it for them. Not necessarily a reward, but "your review helps the next person choose the right size" is a reason, and "we value your feedback" is not.
• They cannot do it on a phone. Most of them are reading your email on a phone and most of the photos they would attach are on that phone. If the flow is awkward there, the flow does not exist.

There is a selection problem hiding underneath all of this, and it is worth naming. The customers most likely to push through a bad review flow are the ones with the strongest feelings, which skews toward the delighted and the furious, and away from the large middle who thought it was fine. That middle is the group whose opinion most resembles your next shopper's. Every bit of friction you remove pulls the corpus back toward them, which is a second reason to make the flow trivially easy: it does not just get you more reviews, it gets you more representative ones.

When should you ask?

After delivery, not after purchase. This sounds obvious and is the single most common mistake. The request fires on order confirmation because that is the easy trigger to build.

The right timing depends on the product, and the variable is how long it takes to form a real opinion.

Trigger from the delivery event where you have it, not from the fulfilment event. The gap between "shipped" and "arrived" is where badly timed requests live, and on international orders that gap can be three weeks, long enough that a request fired on fulfilment arrives before the parcel does, which is both useless and mildly insulting.

What makes a request get answered?

The request is a conversion problem like any other, and the same rules apply.

1. Send from a person, not a brand mailbox. A reply-to that reaches a human changes both the response rate and the tone of what comes back.
2. Name the product, and show it. "How is the Alder crew jacket?" beats "Review your recent order", because it removes the memory work.
3. Put the rating in the email itself. Star buttons that submit on click turn a multi-step task into one tap. The written part is then optional, and a surprising number of people continue anyway.
4. Ask one question, not a form. "What would you tell a friend who was thinking about buying it?" produces better copy than any rating scale.
5. Ask for the photo second. Once someone has committed to the rating, adding a photo is a small extra step. Asking for it up front makes the whole thing feel like work.
6. Send a second reminder, once. Three or four days later. Stop there.
7. Say thank you, and mean it. A short human reply to a review is the cheapest loyalty mechanic there is, and it makes the next request land better.

How do you get photos and video?

Customer photos are the highest-value review content you can collect, because they are the one thing that is unmistakably not a studio shot. They also answer the questions your photography deliberately avoids: what the colour looks like in a normal room, how it sits on a normal body.

• Ask on mobile, where the photos are. Any flow that requires a desktop upload will collect almost nothing, because nobody emails themselves a picture in order to leave a review.
• Ask after the star rating, not before. Momentum matters.
• Tell them what would help. "A photo of it in use is more useful than a photo of the box" gets you better images than an empty upload field.
• Show existing customer photos in the request. People match the format they can see.
• Get permission to reuse them properly. If you intend to use a customer photo in an ad or on a collection page, the permission for that has to be explicit, not assumed from the act of uploading.

What are Google's rules for review stars?

This is the part where a well-meaning setup can quietly cost you the rich result, so it is worth being precise. Google's structured data documentation sets out several requirements.

• The reviews must be visible on the page you mark up. Google asks that you make sure the review content you mark up are readily available to users from the marked-up page, and that it must be immediately obvious to users that the page has review content. Markup describing reviews that live behind a tab nobody opens, or on a different page, does not qualify.
• You cannot review yourself. Google states that if the entity that's being reviewed controls the reviews about itself, their pages are ineligible for star review feature. This is why store-level testimonial pages marked up as an Organization review generally do not earn stars, while product reviews on a product page can.
• Multiple reviews need an aggregate. If you include multiple individual reviews, also include an aggregate rating of the individual reviews, with the average rating value supplied.
• Do not aggregate reviews from other websites, and exclude fake or undisclosed incentivised reviews.

The practical version: mark up product reviews on product pages, show them without requiring a click, keep the aggregate honest, and do not import someone else's.

How should you handle negative reviews?

Leave them up, and answer them.

Shoppers actively seek out the critical reviews. It is often the first thing they filter to, because that is where the real trade-offs are described. A page with nothing but five stars does not read as a great product. It reads as a moderated one.

• Answer in public, quickly. The reply is not for the reviewer. It is for the next two hundred people who read that thread while deciding.
• Lead with the specific, not the apology. "The waist runs about an inch small on this cut, and we have added that to the size guide" is worth more than "sorry to hear this".
• Fix the cause where there is one. A cluster of reviews saying the same thing is free product research that a survey would have cost you money to get.
• Do not argue. You will not win, and the audience is not the person you would be arguing with.
• Remove only what is genuinely removable: abuse, fraud, or a review about a different product. "Unflattering" is not a category.

A well-answered two-star review often converts better than the five-star one above it, because it demonstrates something no positive review can: what happens when it goes wrong.

One practical warning about replies at scale. The temptation, once you have a few hundred reviews, is to template the responses. Shoppers spot it immediately. A column of identical "We're so sorry to hear this, please contact support" replies reads worse than no replies at all, because it proves nobody read them. If you cannot answer them individually, answer fewer of them individually. Ten specific replies beat a hundred generic ones, and the ones worth choosing are the critical reviews on your best-selling products.

Should you import reviews from elsewhere?

Depends entirely on whose reviews they are.

• Migrating from another review app on your own store: yes, and you should. Those reviews were written about your products by your customers. Losing them because you changed vendor is a self-inflicted wound. Carry the rating, the text, the author name and the original date across together, and never edit the content on the way.
• Pulling in reviews from a marketplace listing: no. Google's guidance is explicit about not aggregating reviews from other websites, and it also raises a plainer problem. A review of a marketplace purchase is not a review of the experience you are selling.
• Reviews of a supplier's product on a supplier's site: no. Common in dropshipping, and it is the fastest way to build a review corpus that describes shipping times you cannot deliver. It also fails on its own terms. Those reviews describe a transaction the shopper is not having, with a seller they are not buying from.

When you do migrate, verify the aggregate afterwards. A migration that drops the one-star reviews, whether by design or by a filter someone forgot about, leaves you with an average that no longer describes your products.

Where does Edge Reviews fit?

The judgement in this guide is yours: when to ask, what to say, how to answer the difficult ones. Edge Reviews handles the mechanics that decide whether any of it happens.

It triggers the request from delivery rather than fulfilment, so the ask lands when there is something to say. It puts the star rating in the email so the first step is one tap, and asks for the photo after the rating rather than before it. And it renders the reviews on the product page itself: visible, aggregated, and marked up the way the documentation above requires, rather than tucked into a tab that costs you both the conversion and the rich result.

It also imports what you already have, because the reviews your customers wrote for you last year are still true.