Is Stitch Fix an Artificial Intelligence IPO?
While people have been falling all over themselves to talk about what’s under the hood of the Stitch Fix CEO, we’re more interested in what’s under the hood of a company that was featured in our recent article titled AI Becomes Fashionable with these 10 Startups. While there was lots of media coverage about the “artificial intelligence” component of Stitch Fix prior to their IPO, people seem to have lost track of that now that the S-1 has been filed. Even the company seems to be underplaying their AI technology as a cursory look through the S-1 shows not a single mention of terms like “artificial intelligence” or “deep learning”. As it turns out though, this couldn’t be further from the truth.
As investors, we’re keenly interested in companies that are using AI as a competitive advantage so they can generate efficiencies that result in better bottom line numbers and more money in our pockets. We’re also aware that now we’re seeing all kinds of companies claiming to “use AI” when in fact it seems to be more of an afterthought rather than a core focus. Just because you use some NLP algorithms to screen scrape a few web pages doesn’t make you an AI leader. This is why we’re going to take a closer look at Stitch Fix in the context of what they are actually doing with artificial intelligence.
Founded in 2011, San Francisco startup Stitch Fix has taken in $54.5 million in funding to develop a “personal styling platform” which delivers curated and personalized clothes and accessories to you via mail. Touted as a personal online stylist, we decided to take it for a test drive and had one of our MBAs give it a go. You start out by answering loads of questions about the type of person you are, and they’ve come up with the most clever way we’ve ever seen to describe
fat big boned people:
Other fun questions include things like “are you willing to hem your pants” (fcuk no) and “how do you like your jeans to fit” (skinny! do you even need to ask that?) followed by literally 100 different pictures of shirts, patterns, outfits, etc. where we were supposed to pick what we liked. We ended up just glossing over that section out of boredom before being asked to provide our social media profiles (no thank you) so they can “get to know you better”. After you’ve chosen the frequency of delivery (we chose once a month), we’re then advised of a $20 styling fee and promptly asked for our credit card info. Sorry lads and lasses, there’s no way you’re getting a credit card number without telling us what you’re going to charge to it first. It’s at this point though that we can guess what happens next. We’ll receive a box with 5 pieces of clothing (from over 700 brands) and get charged a bunch of money, but that’s hardly where it ends.
When you receive your “Fix” as they call it, there will be five items in the box. You are free to keep any or all of items and send back what you don’t like to one of their 5 fulfillment centers (with a total footprint of 1.8 million square feet). Now here’s where the genius part comes in for those of us who don’t spend money like drunken sailors on fashion. If you keep all the items, they will discount the entire purchase by 25%. Now that’s clever people.
The business model is pretty slick and by now we’re realizing just how much this company uses data. In fact, the very first page of the S-1 filing focuses heavily on data science and “trusted relationships” (the kind that requires us to just hand over our credit info sight unseen):
These people aren’t just selling clothes, this is a “clothing as a service” platform that involves a subscription model with “run rates”, something that investors love. As of July 2017, they had 2,194,000 active clients and a repeat rate of 86%. (An active client is defined as someone who purchased a Fix in the past rolling 12 months.) What that means is that they’ve now hit almost $1 billion in revenues ($445 per active client, per year). While they stopped losing money in 2014, they reverted back and lost just a tad in 2017. Who cares about that though when you look at their simply explosive revenue growth over the past 4 years:
The S-1 points to the struggle we see being had by traditional retailers and the need for customization through technology. “Our data science capabilities fuel our business,” they say, and the average client provides them with 85 meaningful data points through their style profile (that’s more patience than we have). Over time, clients provide even more information with 85% of shipments resulting in direct client feedback. As we’ve said many times, the best artificial intelligence algorithms are the ones with the highest quality data sets, and we’re seeing that manifest itself with Stitch such as in the below example:
For example, our Delila embroidery neckline knit top is purchased 52% of the time it is included in a Fix. However, for a particular client for whom it is well suited, our algorithms may predict she is 80% likely to purchase the item if it were included in her Fix. This allows us to more efficiently tailor every Fix to each client’s specific preferences.
While they claim to have 3,400 stylists working for them, let’s not mince words here. Artificial intelligence algorithms are going to be superior to humans in every way, shape and form to predict what clients like and adjust to the whims of fickle fashionistas everywhere. Stitch has simply rocked this one, and an acquisition by Amazon (NASDAQ:AMZN) should be in the cards here. What we’d expect to see happen is that they eventually show all those 3,400 “fashion experts” the door as soon as the AI algorithms have been sufficiently trained. Of course that may put a dent in their “86% of our employees are female” metric, but what’s more relevant here is the impact of all this excess overhead to their bottom line. What investors care about is just how fast Stitch can turn this revenue stream into a profitable cash cow that will reap rewards for many years to come.
The IPO plans to raise about $100 million and will trade under the symbol “SFIX”.
If you enjoyed this article, then sign up for our free newsletter - Nanalyze Weekly. About every week, we'll send you a simple summary of all our new articles. If you didn't enjoy this article, share it on Twitter and tell everyone how much you hated it.