Chinese product reviews and ratings: earning the ones that actually convert
Short version: in China, reviews and ratings aren't a trust badge you collect after the sale — they are the conversion layer. A Chinese shopper reads the buyer photos, the follow-up reviews, and the store's rating before they read your product page, and if those don't hold up, nothing else you do matters. The catch is that since September 2024 the old shortcut — a little card in the box offering a red envelope for five stars — is illegal. So the job now is to earn reviews that convert, on two different systems that most foreign brands mistake for one. Here's how they actually work.
A pattern I see constantly: a foreign brand pours budget into KOL seeding and a slick Tmall storefront, then watches conversion stall — and can't work out why. The traffic is there. The page is beautiful. What's missing is the thing a Chinese shopper actually scrolls to before deciding: the reviews. A Western buyer treats reviews as a tie-breaker. A Chinese buyer treats them as the primary evidence, and treats a polished brand claim with no proof underneath as a reason to close the tab.
Two review systems, not one
The first mistake is lumping everything together as "reviews." In China there are two distinct systems, and they do different jobs. Confusing them is why brands over-invest in one and neglect the other.
- Marketplace ratings and reviews — the star ratings, buyer photos (买家秀, "buyer shows"), and follow-up reviews (追评) that live on your Tmall, Taobao, or JD product page. These are the close. They decide whether someone who already found you actually buys.
- Social proof and "grass-planting" (种草) — the review-style notes on Xiaohongshu and Douyin that people read before they ever reach your store. These are the discovery and consideration layer. They decide whether someone comes looking for you at all.
You need both, and they're earned differently. A five-star Tmall rating won't make anyone search for you. A viral Red note won't survive contact with a product page full of one-star complaints about sizing. Treat them as one funnel with two gates.
How marketplace ratings actually work
On Taobao and Tmall the number that governs your store isn't the average star rating shoppers glance at — it's the DSR, the Detailed Seller Rating. It breaks into three scores, each on a 1–5 scale, refreshed on a rolling window:
| DSR component | What buyers are rating | What actually moves it |
|---|---|---|
| Description accuracy (描述相符) | Did the product match the listing | Honest photos and specs; no over-promising in the copy |
| Service attitude (服务态度) | How the shop's customer service handled them | Fast, human replies on the in-store chat, especially on problems |
| Shipping speed (物流服务) | How quickly and cleanly the order arrived | Your warehousing and last-mile setup — often the weak link for cross-border |
Two things foreign brands underestimate. First, the platform shows each score against the category average with a colored marker — a below-peer score is visible, not buried. Second, the DSR isn't cosmetic: it feeds search weighting and campaign eligibility, and a Tmall flagship that lets its scores slide can lose promotional slots or, in bad cases, its standing. The score you're tempted to ignore is the one the platform uses to decide how much traffic you deserve. That's also why a low rating is a marketing problem, not just an operations one — I've argued the same about after-sales service being a channel, not a cost center.
The shortcut that's now illegal
For years the standard play was the insert card: a little slip in the parcel offering a red envelope, a coupon, or a few yuan back in exchange for a five-star review with a photo. Everyone did it. As of the Interim Provisions on Anti-Unfair Competition on the Internet — in force since 1 September 2024, and reinforced by the revised Anti-Unfair Competition Law that took effect 15 October 2025 — it's explicitly prohibited. The rules name the specific tactics: inducing positive reviews with cash, red envelopes, or coupons; hiding or burying negative reviews; and reordering so the good ones always show first.
This matters even if you'd never have written "pay for reviews" into a plan, because your Chinese TP agency or distributor may still be running the old playbook on your behalf. If it's your flagship, it's your liability. Ask directly how they're generating reviews, and get "review cashback cards" out of your parcels. The upside: the shortcut dying makes genuinely earned reviews more valuable, because the baseline is finally being cleaned up.
Earning reviews that convert, without buying them
The good news is that the legitimate levers are also the ones that produce reviews worth reading. What works:
- Engineer the unboxing, not the bribe. A card that says "thanks — tell us how it went, good or bad" is fine. A card that pays for five stars is not. Give people a genuine reason to photograph the product: packaging that looks good in a buyer show is free review bait.
- Make the first 100 reviews count. Early on, a small batch of real, detailed reviews with photos does more than a thousand thin ones. Seed with actual customers — sampling programs, KOC gifting where the review is honest and disclosed — not with a review farm.
- Answer the follow-up review (追评). Chinese platforms let buyers add a second review weeks later, after the product's been lived with. That's your durability proof. Prompt it gently through post-purchase service, and it becomes some of your most persuasive content.
- Feed the social layer separately. On Xiaohongshu — around 234 million monthly users, and where roughly two-thirds of them say platform content drives their purchases — the review that converts reads like a real answer to a real question, not an ad. The algorithm throttles notes that smell promotional, so brief creators to solve a problem, not to recite features. More on that in getting found on Red's search.
Managing the negative ones
You cannot delete your way to a good rating anymore, and honestly you never should have. A store with only five-star reviews reads as fake to a Chinese shopper — they'll sort by lowest first specifically to find out what's wrong. So the negative review isn't the enemy; the unanswered negative review is. A visible, specific, non-defensive reply — "you're right, the sizing runs small, here's the fix and here's a size guide" — often converts the browsers reading it more than the complaint cost you. It shows the brand is present and fair, which is exactly the signal thin, purchased reviews can never send.
The pattern to watch for is a cluster of the same complaint — sizing, a shipping delay, a translation error on the label. That's not a reputation problem to spin; it's product or ops feedback arriving free and in public. Fix the cause, then let the follow-up reviews record the improvement.
Bottom line
In China, reviews and ratings are the conversion engine, split across two systems: the marketplace ratings that close the sale and the social notes that open it. The old shortcut of buying five stars is now illegal and, more to the point, worthless — Chinese shoppers were already discounting it. What's left is the harder, better work: a product and unboxing worth reviewing, service fast enough to protect the DSR, honest seeding for the first hundred, and negative reviews answered in the open. Do that and your reviews stop being a scoreboard and start being your best salesperson.
If you're standing up a China storefront and want a second read on how your reviews and ratings are being handled — including what your local partner is doing in your name — that's the work I do. Reach out.