It is said visual search and personalized recommendation should be the future for the fashion business. We have actually seen some very prominent and promising startups in this field.
Many analysts might pull out below list to tell you it is going to be the hottest trend, already deployed by some of the leading apparel e-commerce sites in the world.
But in China, this phenomenon remains, by and large, an Alibaba’s experiment, because:
- Tmall and Taobao are the most preferred destinations for apparel shopping according to a certain survey.
- Mono-brand apparel e-commerce often relies on its online flagship stores on Tmall for traffics and conversions, instead of its stand alone e-commerce.
- Alibaba is a well-known yet rather closed ecosystem, which does not freely allow API integration with 3rd party features. On top of that, Alibaba implemented its in-house visual search technology across Tmall and Taobao many many years ago.
- Even though big Mono-brands set up their independent e-commerce presence, they usually do not have a crying need for visual search&recommendation due to small-scale coverage of SKUs.
- Strictly speaking, we do not have these dominant & vertical fashion e-commerce platforms selling multi-brands like Zalando or ASOS in the country.
Above factors can dampen the fate of those local startups which wish to earn gold from peddling AI magic for fashion industry.
Somehow JD.com, China’s #2 e-commerce site, with its burning ambition in cultivating its fashion penetration, might shine hopes for these startups.
A baby startup Infimind established in 2017, joined JD’s AI accelerator a while ago.
Infimind is founded by a recent graduate from “China’s MIT” Tsing Hua University and backed by a leading VC firm GSR Ventures.
Infimind started out by providing fashion visual technology in :
- Find similar
- Complete the look
- Shop the look
Infimind holds high expectation on the rising popularity and flexibility of Wechat mini-program on which it created a demo account in order to impress fashion brands. I tested “complete the look” feature and the result seems to be passable but not particularly satisfactory. You can watch below video.
This week I had an interesting conversation with Infimind’s CEO.
Since the startup just joined JD’s AI accelerator, it should not surprise us that it is going to launch the pilot project of “complete the look” for JD.com. Meanwhile, it is also working in progress with Secoo, China’s one-stop comprehensive luxury products and services platform invested by JD, for the same pilot.
The pricing is very affordable, as the cost per click on “complete the look” downs to only a few cents in RMB.
I challenged him regarding no strong motivation from mono fashion brand to use his company’s visual technology, he admitted that online channel or even mini-program probably have faced bottleneck so far, however many local apparel brands are gradually rolling out the offline smart store and have the desire to embed visual search/recommendation solution into the digital signage. Infimind is even in talks with Belle, China’s #1 women’s shoe retailer, pushing for the notion of “complete the look” for each pair of shoes.
Unexpectedly, Infimind CEO told that substantial demands have been coming from product auto-tagging and description for the back-end of fashion e-commerce. That is why Xiaohongshu (Little Red Book, China’s most prestigious social e-commerce platform) is now testing its solution in this part. Small to medium TP (Tmall Partner which offers end to end e-commerce operation service for brands) such as Transcosmos shows great interests as well. Nevertheless giant TP in China, for instance, Baozun should have already developed their own visual technology as Infimind CEO revealed that their discussion with Baozun did not end up in any meaningful way.
The growing pain of local visual technology startup in China will not be small, but product image recognition and auto-tagging might be the silver lining here to jump up their business.
By: Cecilia Wu