A Tmall store that opens quietly stays quiet. Traffic does not arrive because the store exists. It arrives because a brand pulls it in from several directions at once, and because the store is built to convert what shows up. This guide is for brands that already have a store live and are wondering why the visits are not there, or why they plateau after the opening push.
We work with brand teams on Tmall stores every month, and the pattern repeats. A brand spends heavily on paid tools in month one, sees a spike, then watches conversion collapse because the listing was never ready for the visitors it bought. Fix the order of operations first. Then scale spend.
Listing quality comes before any spend
Say this plainly: buying traffic into a weak product page burns money. Tmall’s ranking and its paid tools both reward listings that already convert. A page with thin images, no video, missing size or ingredient detail, and few reviews will take clicks and return almost nothing. Worse, a low conversion rate after a paid click actually damages the page’s own organic ranking, because the algorithm reads it as a mismatch between demand and product.
Before spending one RMB on ads, check five things: main image click-through against category benchmarks, at least one short video on the page, five or more genuine reviews with photos, a detail page that answers the three questions a Chinese shopper actually asks (is this genuine, does it suit me, what do others say), and a price that sits inside the range shoppers expect for the category. Sequencing matters here. Fix the listing, then buy the traffic.
On-platform SEO: title, keywords, category ranking
Tmall search still runs on a keyword-matching logic layered under an AI ranking system. Title structure matters. A title needs the core category keyword, the brand name, one or two attribute keywords (size, flavour, material), and nothing padded in for the sake of length. Overstuffing hurts more than it helps now.
Category ranking on Tmall weighs sales velocity, conversion rate, return rate, review score, and response time to buyer messages. It is not a static score. A store that answers customer chat within minutes and keeps returns low will outrank a competitor with a bigger ad budget but slower service. Sellers often obsess over ads and ignore the operational metrics that quietly decide organic placement.
Update titles and tags around major shopping festivals, when search volume shifts toward gift-related and bundle-related terms. A title that performs well in March may be flat during Double 11 because shoppers are searching differently.
Paid tools: 直通车, 钻展, and 万相台
直通车 (Zhitongche) is Tmall’s pay-per-click keyword tool, similar in logic to search ads elsewhere. It puts a listing in front of shoppers actively typing a relevant term. It works well for proven listings and badly for untested ones, because you are paying to learn rather than paying to convert.
钻展 (Diamond zhanwei, display placement) works differently. It buys visual placement across Tmall and Taobao based on audience targeting rather than search intent. It suits brand awareness and retargeting warm audiences who have viewed a product but not bought. It is a weaker tool for cold demand generation than 直通车.
万相台 (Wanxiangtai) has changed the most. According to a 2026 breakdown from Alimama, the upgraded 万相台无界版 now merges what used to be separate tools, 直通车, 引力魔方 and the older 万相台, into a single AI-driven placement engine covering search, short video, livestream, audience and product feed in one dashboard. Alimama frames it as the primary tool Tmall pushes for 618 and Double 11 campaigns, running on an internal agent engine that automates bidding and creative matching across channels (source, translated). Practically, this means brands set a goal, a budget ceiling and a target ROI, and let the system allocate spend across search and recommendation traffic rather than manually splitting budget between tools. Early in a campaign the system leans toward reach and new-customer acquisition, later toward harvesting conversions from the audience it has already warmed. Brands running 万相台 for the first time should expect a learning period before it stabilises, and should not judge it on the first week’s ROI alone.
None of these tools substitutes for the others. 直通车 finds active search demand. 钻展 and 万相台’s display side build awareness and retarget browsers. Skipping straight to the biggest budget tool without understanding which layer of the funnel it serves wastes spend fast.
Off-platform traffic: Xiaohongshu and Douyin feeding the store
Tmall’s own tools cover shoppers already searching inside the platform. They do nothing for shoppers who have never heard of the brand. That demand has to be built somewhere else and pointed at the store.
Xiaohongshu is where product discovery and trust-building happen before a Chinese consumer ever opens Tmall. A KOC review, a genuine-looking unboxing note, a comparison post, these shape the decision before the search bar is touched. Brands that run Xiaohongshu content and store promotion alongside their Tmall presence see search volume for their own brand name rise inside Tmall, which is a stronger ranking signal than almost any paid click.
Douyin works differently again, closer to impulse and livestream-driven demand. A well-placed short video or a livestream event can send a burst of traffic directly to a Tmall listing if the link path is set up correctly and the landing page matches what the video promised. Mismatched expectations between the video and the product page kill conversion fast, so the two need to be built together, not handed to separate teams. A Douyin traffic strategy built to feed the Tmall store rather than a standalone Douyin shop avoids splitting inventory and reviews across two storefronts.
Off-platform traffic is slower to build and harder to attribute cleanly. It is also more durable than paid clicks, because it builds brand search volume that keeps paying back after the campaign ends.
Member and CRM retargeting: the traffic you already own
Most brands underuse their own member database. Tmall stores can build a member tier system, issue coupons to past buyers, and push notifications through the store’s messaging tools directly to people who have already bought once. This channel costs almost nothing compared with paid ads, and it converts far better because trust is already established.
A simple cadence works: a thank-you message with a small discount a week after purchase, a restock or new-arrival note tied to the shopper’s past category, and a festival-timed coupon before Double 11 or 618. None of this requires new tools beyond what Tmall already provides in its member centre.
This is also where a Tmall store run properly compounds over time. A store six months old with a real member list will always out-convert a store that only ever chases cold traffic. If the store is new and this base does not exist yet, building it is part of the setup a Tmall store agency should handle from day one, not bolt on later.
The sequence that actually works
Fix the listing first. Then earn organic ranking through service metrics and correct titling. Then layer in 直通车 for proven keywords. Add 万相台 once there is enough data for its AI allocation to work with. Build Xiaohongshu and Douyin content in parallel to create demand the platform cannot generate on its own. Retarget the members you collect along the way. Brands that reverse this order, ads first, fixes later, pay twice for the same traffic.
Sources: 2026天猫万相台无界版超详细推广攻略大全, Alimama, translated
About the author
Harry Huang is the Founder of Ecommerce China Agency. He has spent over a decade helping international brands launch and grow on Tmall, from the first store audit through to full-funnel traffic strategy. One detail he repeats to every new client: the first thing his team checks on a struggling Tmall store is not the ad account, it is the response time on buyer chat, because a two-hour delay on messages can cost more ranking than a missing keyword ever will. Want a straight answer on where your store’s traffic is actually leaking? Get in touch with our team.
