Beyond PMF: AI and Tech As A Fashion Industry
"You are wearing a sweater that was selected for you"
Happy Sunday and welcome to Investing in AI! I’m Rob May CEO at BrandGuard (yes we are rebranding the company to the main product name). I also produce the AI Innovator’s Podcast (reach out if you have a good guest suggestion), and am an active AI angel investor.
A few links this week… my friend Doug Levin has an AI Summer Reading post. My colleague Eric Koziol writes about whether or not AI can know our desires. And Erik Duhaime from Centaur Labs joined me on the podcast to discuss data labeling and AI.
The problem I’ve always had with how we discuss product-market fit (PMF) is that we treat it like an opening in a jigsaw puzzle. If you just find the right piece that fits, it’s all good, and your job as an entrepreneur is to figure out what the market wants. This has not really been my experience as an entrepreneur or an investor.
I see the situation more like the fashion industry. There is a scene in the movie “The Devil Wears Prada” where Meryl Streep, playing a high powered fashion designer, is trying to choose between two belts that her assistant says look so different. Anne Hathaway, who plays an admin that is new to fashion, laughs because, as she says “they look the same.” Streep then gives a lecture on fashion, the gist of which is - you think you wear that lumpy blue sweater to show you don’t care, but in fact that sweater was chosen for you by the very people in this room. The point is that our choices are often influenced in ways we don’t understand, even when we believe they aren’t.
The scene is just 2 minutes but worth watching.
I bring up this issue because I think as tech has become more mainstream, it behaves more and more like the fashion industry. For example, I’m an Android user, and have been for a long time. I’m regularly showing off cool phone features to iphone users that they don’t have (Android has had almost every major innovative feature first), who usually aren’t aware that an Android can do whatever cool thing it is I’m showing. We discuss the feature for a bit and then they lament, “yeah that’s cool but I’m so hooked on Apple because of the design.” Choosing form over function - isn’t that the definition of the fashion industry?
I see this everywhere in tech, but I don’t think we tech people like to think of ourselves as this way. We are engineers and MBAs and thoughtful critical thinkers. We are independent minded and don’t make decisions based on what’s popular or seemingly cool in the moment. Except that, watching the tech industry over the past 20 years I’ve seen it, I think we do. Tech is a fashion industry.
The reason it’s important to understand this is because it affects how you approach company building. Finding PMF isn’t just about a static market with a hole that you can plug into if you build the right thing. It’s actually about a dance between companies and markets and users. You have to influence the market as much as you have to read the market, especially if the market is nascent. Early markets can evolve in many different directions and I don’t believe one is foreordained. The early companies influence how they evolve, and if you had different companies the market could have evolved very differently.
This is an important issue for AI companies because they are creating so many new use cases in so many new markets. If you are working on things like the agent or auto-gpt spaces, we don’t have established markets for those. Finding PMF will be partially about creating PMF.
In the fashion industry, you can’t make something that is really ugly and stupid into something popular. But from all the possible things that could be popular, you can influence what gets to the top. (Research has shown this with song popularity). I believe tech is the same way.
This isn’t to say you can build anything you want and force the market to take it. You can’t. But I’m pointing out that too many companies are trying to build for the market as if it’s a static thing with a hole for a new product opportunity, and that it’s just about talking to customers and finding the right thing rather than influencing how the market will play out. You can, and should, try to influence how the market develops, and few early stage companies make the effort to do so. Those opportunities will be plentiful in new areas of AI products.
Thanks for reading.
Agree with you Rob. Getting to PMF is the mutual discovery process of you and your customers - finding use cases where your offering creates value for the buyer, and determining how that value creation is repeatable and scalable. It does not need to be bounded by the past because value creation can happen in new, unexpected ways, and AI provides all kinds of opportunities for that type of innovative value creation.