Fashion Tech, turning shit into sugar?

Selling clothing can be a tough business, facing shitty problems all the time, such as:

  • low margin
  • soaring unsold inventory
  • high return rate online
  • low conversion rate offline

 

We talked about the fashion AI in action a long time ago, focusing on trends in the west.

See Link 1

See Link 2

In the past two years, a string of local Chinese startups has sprung up as well in the area of B2B fashion tech, often under the auspicious namesake of “AI”, claiming to solve all pain points, turn shit into sugar and ultimately achieve automation, personalization, and prediction. I identified below local startups in several categories as examples.

Meanwhile, Alibaba is making a stride in the wave of New Retail, attracting startups under its wing to become the official partners to help fashion brands (for instance Lily, Guess) conduct offline store digital innovation. Expectedly Tencent is following suit.

However, many fashion tech startups are still struggling for a springboard.

Recently I had an interesting conversation with the founder of Weiyi which is a startup specializing in AI stylist. The solution basically aggregates all the possible external & internal data together (like the below graph indicates), then processed by its so-called fashion brain, and is able to generate personalized styling recommendations.

The founder has a decade of experience in the fashion industry and accumulated a profound network. Even though he has easy access to key decision makers, to convince and to convert fashion brand for a pilot project is no easier task.

First, most foreign brands will shut him out of the door, because the solution would require them to open their CRM or any related internal data, this usually has to be approved by the headquarter. It is going to be a long process and rarely would centralized governance of foreign brands loosen the grip and permit such a localized experiment.

Second, local fashion brands might be the low hanging fruit as they can act fast and independent. Nevertheless, most of their e-commerce distribution is Alibaba Tmall centric, which is essentially a rather closed ecosystem to allow any integration of 3 party solution given unauthorized as an Alibaba’s official partner. So online channel might be a huge roadblock for now. The silver lining is a big chunk of sales of local fashion brands still comes from offline, therefore the founder is now negotiating with brands like Peacebird and Youngor as they are eager to upgrade their existing brick&mortar stores and his stylist solution can be perfectly embedded into the fitting room or hardware of “magic mirror”.

Third, styling is a very subjective topic. I asked the founder below question:

I have a friend who prefers the style of sexy and purchases pretty much everything under such idea. But based on her face structure and body shape, many consider maybe she should go for a more covered-up cute style. So if you run a AI fashion analysis, how would you give her recommendation? Sexy or Cute?”

The founder suggests his AI is able to find a middle ground for my friend.

But again it only works well for the big platform like Tmall or JD with massive SKUs to choose and recommend from in the first place. A quite lot of brands already have a very distinctive style, for instance, Uniqlo is famous for its plain and simple formula, and probably has nothing to do with sexy, so it is hard to find the middle ground for a single-brand to recommend a style for customers like my friend.

The founder obviously knows my point. He admits that he has to build zero to one. He needs to create more pilot projects with more fashion brands in order to hold solid leveraging power and then bring it to the table to settle a large scale pilot with big platforms like Tmall, JD, Secoo, Vipshop, LittleRedBook etc.

In many ways, the road to success still sounds hard and long for the founder.

 

By: Cecilia Wu