Do you have a 5-second summary of what you mean by "Growth specifics and positioning of algorithms in the StitchFix S-1"? Just the fact that they use algorithms to tailor individual boxes, or do you mean something more specific?
I currently work at an e-commerce company and there has been a ton of debate around how "algorithmically driven" Stitch Fix actually is. The general feel in the space from non-technologists is that computers cannot do this job well now and won't be able to do it well in the near future. Stitch Fix makes it a major brand point that computers are an important part of the process. So the real question is - are they making this point because it's true or because it helps their valuation (tech co. 5-10x multiples instead of ecomm 1-4x multiples).
The way Stitch Fix talks about it in their S-1 makes it seem like the latter is the priority. I'm not yet convinced that the practical value driven by algorithms at Stitch Fix is up to par with how much they talk about it.
I was interested in growth to understand both their growth rate but also to get a feel if it was driven by increasing user acquisition costs like Groupon, Blue Apron, etc or if it was organic.
I've worked in fashion before and I'm a machine learning pratictoner now.
I never tried myself but it seems quite feasible to build a decent profile of someones taste in fashion from a bit of data.
Fast fashion gave us the logistics (no more 9mo from concept to store). But Zara and friends still supply only the major trends. We're still missing for someone to reliably market the "long tail".
You could initially present a user some images to assess it's preferences and use some recommender system (collaborative filtering), a la Netflix.
That could give you an starting point. But I believe the main issue is that fashion products have, by definition, short shelf life, so you can't run the algos on SKU data. Then you can use deep learning on product images + user categorical data to try to predict preferences, maybe simple binary classification?
I guess using images as input should give better features than textual description.
> You could initially present a user some images to assess it's preferences and use some recommender system (collaborative filtering), a la Netflix.
You would have to tag the hell out of these photos, right? Disambiguating preferences is the challenge -- the user may have liked images #3 and #7, but why? Specific items, or the color palette, or the silhouette, or just the model? A post by Chicisimo on the front page addresses these hurdles [1].
I've also seen some of these image recognition apps in action -- they pick up on patterns and color very well, but struggled with silhouette.
Seeing what people themselves like doesn't sound a good way to determine what fashion they'd buy. Fashion is about a consensus largely established by a cabal of sellers; users surely want to find out what is fashionable? Fashion is mainly about what other people like.
Oh, that's largely a myth. These seller do have some influence, but it's much less important than what they want you to perceived. Fashion tastes comes from a lot of sources and some is part of your identity.
You hit the nail on the head. They also reveal this strategy through their data science "thought leadership" posts, which are seemingly meant to appeal to data folks (own your code! add value through modeling! don't worry about collaborating with engineers!) but do not reflect the realities of industrial data science.
Chris Moody knows more about this than anyone in the world, and everything he writes about on the Stitch Fix blog seems so far ahead of what other people are talking about I really do believe that they have a real technology advantage.
It seems similar to talking to people at Google in the early days. The thing that they cared about from an engineering point of view seemed weird and they language was alien. 3 years later the rest of the world hits the same problem and I remember the conversations and think "oh so this is what they were talking about".
