The Not Daily Stuntz — Stitch Fix, Amazon, and Strategy

Part of my current role is to think about strategy and positioning in terms of our capabilities in cybersecurity and data protection market. While I feel this space is critical and I am passionate about making progress in actually improving the state of security, without making everything awful to use, there is lots to learn from others in any industry.

One of the ways I get high level ideas, frameworks, concepts that I then dive into is through podcasts. There are thousands if not millions on business, strategy, leadership but one that I have enjoyed for a while is the 20 Minute VC with Harry Stebbings (https://thetwentyminutevc.com/) and a recent episode generated a lot of thoughts that I am going to write here to save my family and friends from hearing me think out loud.

Yesterday I listened to an episode from a few weeks ago with Katrina Lake, the CEO of Stitch Fix. If you need a summary, she is incredibly impressive and while I am not the trendiest person (I focus on retro Jordans and band t-shirts, mostly black) anyone could learn a lot from listening to her. For background, Stitch Fix sends “fixes” of clothes to subscribers who keep what they want and return what they dont. What the company sends is based on a stylist but more importantly data from the subscriber and feedback from other customers. Finally, podcasts are so interesting that people like Katrina with the demands on her time make an effort to participate and congrats to Harry Stebbings on building his platform. Caveat for the rest of the post, apologies if others have written on this and if I am copying others ideas, it is not on purpose, but I enjoy thinking about industries I dont know as well.

One of the main points I took away from the conversation is that Katrina doesn’t see Stitch Fix as a fashion mixed with data science company, but as a pure play data science company. She mentioned multiple times the idea that she believes they can match people to products or services outside of fashion based on the success they have had with fixes. This idea of providing people with goods they want before they even may know it is a fascinating one. Scott Galloway, entrepreneur/professor/author/podcaster/general really sharp guy, has talked recently about E-commerce moving to A-commerce (algorithmic commerce) where data from the user is used to suggest items they might not know they want yet. This has lots of interesting privacy points, but that is a separate post. When I hear Katrina talk about leveraging their core competency of data science for outside of fashion I think that makes sense, but also makes me think she is looking right at Amazon.

While Amazon has not been successful in fashion, at least compared to other industries they have gone into, and I thought might look at acquiring Stitch Fix pre-pandemic, using data to provide better products faster is their game. As I thought about what I would do if I were Katrina (she is certainly many hundreds of steps ahead of me) a few things came up that would influence my thinking around expansion.

1. Amazon is going to be able to combine all the shopping and return history, with potentially biometric data from Alexa and their wearables (leaving privacy aside just for this post). Something like Alexa heard in your voice that you were stressed and your heart rate is up according to the wearable plus you order ice cream from Whole Foods and now look, ice cream waiting at your door charged to your Prime account. Even if Stitch Fix was able to match me with a slightly better thing I want, Amazon will have the logistics advantage for anything time-related and more data sources to base decisions on. Also Amazon has all of the processing power in the world to drive models.
2. This led me to focus on expansion into things that look more like fashion: where hours or even a day or two is not much of a differentiator, somewhat high margins, brand or image matters, and where user feedback generates a flywheel (sorry for overused business term). One of the gold mines for Stitch Fix is that users are incentivized to provide detailed feedback so their next fix more closely matches what they might want. At the same time, this data goes into improving the models. It is not just five star reviews, but details that can add a lot of context and differentiate outcomes. Amazon may have a seemingly unlimited amount of data, but this detailed user feedback may be at least a short term moat (another overused term).
3. Has to fit, at least initially, in a subscriber model with a fairly low level of effort from the customers. I could be wrong, but surprise here is your new couch based on our model and good luck returning it doesn’t seem like the right direction. This would also support expansion/upsell of current fashion focused customers since the experience would be similar and I am sure they have internal workflows tailored for this type of experience.

After walking through this exercise I had a few ideas of where they could go next, but each of them had their own challenges. Wine for example meets a lot of the criteria, but is a smaller population of buyers compared with fashion which is somewhat universal and is not as easy to try and then return. Furniture has logistics challenges, etc. I look forward to following Katrina and the Stitch Fix team to see what they do next and if my thinking was remotely right, or if I should stick to the cyberz.