This Double 11, merchants are doing business in consultation with AI.

By Wang Yi

AI is pushing e-commerce marketing to an earlier stage where consumer demands are taking shape.

The operation head of a sports and outdoor brand opened the advertising backend, only to find the system recommended a group of audiences he had never seen before: "gift-givers" and "minimalists".

Yet the products he sells are ski suits. Apart from ski enthusiasts and outdoor pros, the AI advertising tool he used helped him identify a group of new customers he had never thought of targeting.

Double 11 comes every year. What makes this year's event thought-provoking is that the transaction volume is not only supported by familiar regular customers, but also by new customers and the consumer power that has been newly tapped.

The results delivered in this year's Double 11 are not just a boom of a certain category, but a signal — the underlying logic of e-commerce marketing is undergoing a paradigm shift.

Not long ago, Liu Bo (alias: Jialuo), Senior Vice President of Alibaba Group, President of Tmall Division, President of Alimama, Head of Taobao Live and Content Division, publicly stated that the value of AI lies in understanding consumer demands in a deeper and more long-term manner. He regards "shifting from meeting existing demands to creating new demands" as the irreplaceable core competitiveness of brands.

Understanding why a person needs a certain product has become the starting point for brands to find their target audiences.

This shift comes at just the right time. Similar brands are looking for audiences along similar clues, and the familiar group of customers is being competed for repeatedly, so growth requires new purchase justifications. When consumers encounter problems in a certain life scenario, they may not know what to buy to solve them; brands with corresponding products may not even realize that their offerings can fit into such choices.

AI begins to help both sides build such connections, turning previously unrecognized needs into explorable business opportunities.

01. Where exactly are the "people" beyond the existing labels?

The next customer of a ski suit may not even be on the ski slope at all.

The brand mentioned at the beginning is BOGNER, a German ski sports brand. Its operation team realized long ago that people willing to pay for the brand are far more than those on ski slopes. Snowfield vacation is becoming a lifestyle, and sportswear is also worn in daily scenarios. Consumers will pay for fine workmanship, unique style and a sense of scarcity, and all these purchase justifications go beyond the basic purpose of ski equipment.

However, these judgments could not be put into practice when launching advertising campaigns. Ski enthusiasts and high-end sports crowds have ready-made interest clues; but where exactly are the people who show no interest in skiing yet are likely to consume for the brand?

A gift-giving occasion, a trip, a kind of identity expression... These real purchase motivations cannot be stuffed into the "skiing interest" label. Brands know their products can fit into more scenarios, but their methods of finding audiences have failed to keep up.

This is also the common situation across the industry. Search ads capture the demands that users actively express, and recommendation systems rely on interests and historical behaviors. If brands continue to use familiar labels, they will easily ignore consumption motivations that have not been included in the target crowds.

In this year's Double 11, the new idea put forward by Alimama through its AI targeting product Wanxiang Dianjing is to change the starting point: shifting from "who this person is" to "what this person needs, and why they need it".

BOGNER submitted this confusion to Wanxiang Dianjing. The product analyzed the video content of the brand's ads and the video behaviors of users across the entire network, and found that people who buy BOGNER products are not just purchasing skiing performance, but also expressing a certain attitude towards life. After understanding the brand's business demands, Wanxiang Dianjing deduced multiple consumer crowds that are not directly related to skiing but match the brand's tonality: gift-givers who value quality and scarcity, high-net-worth fashion crowds who refuse to wear the same outfits as others, urban commuting white-collar workers who pursue style expression, and minimalists. All these people are likely to pay for the brand's fine workmanship and design.

The maternity support belly band from Babycare is another example. This product was originally mainly targeted at pregnant and lying-in women. Starting from the "waist and abdomen support" function, Wanxiang Dianjing extended its deduction to scenarios such as long-hour gaming and live streaming content creation, and listed "e-sports girls who sit for the whole afternoon" and "live streaming creators who stay seated for hours without changing posture" as potential crowds.

The demand for waist and abdomen support already exists, but consumers may not connect it with a maternity product, and merchants may not regard these people as potential customers either.

What AI finds is the connection that has never been established between the two sides. "Creating demands" has a specific commercial meaning here: discovering new purchase justifications from existing products, so that consumers will include them in the choices for solving daily problems.

Brands are used to estimating the market by category, while consumers arrange their expenses according to the problems in their daily lives. When a ski suit enters the commuter wardrobe, it will face a whole new set of outerwear brands as competitors.

When advertising campaigns are launched around purchase justifications, the boundaries between categories will loosen accordingly. Brands need to re-estimate their market size, and re-recognize their competitors at the same time. For brands in a stock competition environment, mature products that enter new life scenarios still have the opportunity to unlock new growth space.

02. AI has found new crowds, do merchants dare to give it a try?

The crowds recommended by Wanxiang Dianjing are really unexpected — gift-givers, minimalists... These people are completely different from the brand's traditional customer base. For merchants, apart from excitement, there is also hesitation: will these people really buy the products? What if the advertising budget is wasted?

This kind of concern is very realistic. The match between the product's function and a certain demand is not enough to guarantee a deal. Merchants have years of accumulated judgment on their own customers, and when AI suddenly points to a group of people who have never appeared in the target portrait and says "they will also buy", anyone will have doubts.

The solution of Wanxiang Dianjing is to show merchants the full reasoning process of AI.

Before launching the campaign, the system provides multiple crowd plans, attached with crowd size, expected new customer acquisition effect, and the overlap rate with existing crowds. Merchants can not only see a recommendation result, but also understand why AI recommends this group of people. The reasoning link has changed from a black box to a white box.

The newly added "active expression" function during the Double 11 period goes a step further: incorporating merchants' judgments into crowd reasoning. "What problem do I most want to solve in this campaign?" "What is the selling point I most want to highlight?" Merchants state their demands, the system listens, adjusts the recommendation weight accordingly, and retains the crowds that have been proven effective.

Take Hisense as an example. In this year's Double 11, Hisense launched its new E5S TV featuring original colors, with prominent eye-protection advantages, which can bring consumers more extreme picture quality and more comfortable visual experience. The operation team shot a large number of short videos for promotion, hoping that the selling point of each short video can accurately fit the specific pain points of consumers. After reading Hisense's video content and marketing scripts, Wanxiang Dianjing in the super short video scenario extracted three scenarios: eye protection, family entertainment, and immersive movie watching, and adjusted the crowd recommendation accordingly — health-focused people see the eye protection selling point, parents see the children's eye protection content, and young people see the immersive movie watching experience. The Hisense team recognized this direction after reviewing the reasoning link, supplemented the brand's own judgment on price range and target crowds, and the system continued to optimize its recommendations.

The co-creation between AI and humans is more complex, but Alimama is well aware that what merchants need is not just an answer, but the ability to understand the answer and find more solutions.

The results are satisfactory. Data from Alimama shows that during this Double 11, from October 15 to 20, Wanxiang Dianjing brought nearly 1 billion yuan in additional transaction volume for merchants in the two scenarios of keyword promotion and crowd promotion. Among them, the new customer click-through rate of keyword promotion exceeded 90%; for merchants that used Wanxiang Dianjing in the super short video scenario, the new customer ROI was significantly better than that of traditional crowd targeting; the number of merchants that used Wanxiang Dianjing on Super Live Day increased by over 190%, their investment increased by over 250%, and the new customer GMV in the live streaming room accounted for nearly 70%.

The increment has shifted from repeated bidding within the same group of people to areas that were not covered in the past. Marketing thus takes on part of the market exploration work, helping merchants gradually confirm the growth direction through advertising feedback.

03. From a single AI product to a complete set of operation logic

The purchase justifications discovered in one advertising campaign cannot be limited to the campaign itself. They need to be incorporated into subsequent content and product recommendations to continuously influence consumers' choices.

If we only observe Wanxiang Dianjing from the perspective of an advertising tool, its significance is limited. But from the layout of this year's Double 11, it is being integrated into the more complete brand operation rhythm.

In the water storage stage, Uni Desk (the one-stop off-Taobao advertising platform built by Alimama) and Brand Power WIN (Alimama's AI brand growth model) use it to help merchants identify trends and tracks, and capture consumption intentions in advance; after entering the short video, live streaming and search scenarios, the same set of capabilities continues to participate in the matching of content and crowds; in the sales launch stage, the promotion of crowds and keywords accurately undertakes the conversion.

The same set of capabilities runs through multiple operation links, so AI's understanding of the brand will influence more rounds of recommendations. While brands discover demands with the help of AI, they also become the objects that AI needs to understand.

The logic is not complicated. When AI starts to make the "first introduction" of the brand to consumers in the background — recommending your products, it must first understand what value you have and what needs you fit in — the more accurately it understands, the more appropriate the recommendation will be. AI cannot explain a general statement like "good quality" clearly for you; only when the product description clearly states the features and user reviews confirm the experience, can AI convey your value to consumers' hearts.

The more clearly the brand expresses itself, the more powerful the "rephrasing" of AI will be. Brands with vague expressions are more likely to be excluded in AI's screening process.

Alimama calls this the "AI readability" of brand assets. In other words: in the past, the core of brand building was "how to be seen by consumers", but now there is an additional layer — "how to be accurately understood by AI".

What Wanxiang Dianjing does is, in the final analysis, to help brands and consumers find each other. When AI can accurately explain the value of the brand, it will naturally bring this value to the people who need it.

The operation proposition left by Double 11 extends to the period after the big promotion. Brands not only need to understand consumers, but also let AI understand themselves accurately. A clear and stable brand cognition will become an important foundation for continuously discovering new increments.

As mentioned at the beginning, taking one step forward from meeting existing demands to creating new demands is the core competitiveness that allows brands to go further. And now, everything starts from this Double 11 where consumer power is reactivated by AI.