This summer I’ve been teaching product management to the tech MBAs at NYU, and like all teaching, it’s forced me to clarify my thinking. I’m going to touch upon some of those areas in the next few posts. Today’s topic is customer segments — in particular, the initial ones you choose to design for and market to.
I’ve been pushing my students to narrow their customer segments far more than is comfortable. Our natural instincts are to go for a big, broad market. But that’s not a smart place to start. For example, if you were making a new tax prep product, you wouldn’t design it for, and market it to, every tax payer. Not at first. Then what people try to do is narrow things by demographics, but actually it’s usually more useful to narrow by use cases. This is what allows you to narrow scope of what you need to build. Still, it feels very uncomfortable. It feels risky. Don’t we want to go after a big market? Wouldn’t this put all of our eggs in one basket?
Let’s do an example first, before we handle those objections. It’s a useful exercise to try to narrow an initial customer segment so far that it becomes ridiculous, and only then pull back. So let’s take our new tax prep product: all taxpayers in the world is too big and messy. ➡️ OK, how about USA taxpayers? Still too big with competition all over the place. ➡️ USA taxpayers who have multiple forms of income (i.e. complexity but a specific kind). Let’s go further: ➡️ USA taxpayers who have multiple forms of income and tend to be late in filing and paying due to that complexity. Further? ➡️ Investors who have multiple pass-through vehicles (funds, spv’s, LLCs) that struggle to file/pay on time. ➡️ Does it make sense to go further into VC vs PE vs individual investor? Arguably that is thinking too demographic rather than psychographic unless one of those segments has specific needs. Geographic region? Not sure why texas vs new york is that different unless we want to get into state taxes specifically. ➡️ How about investors who have multiple pass-through vehicles which are no longer actively managed, and so information is not being shared out in a timely manner, thus triggering the late filing. Now too narrow? Maybe. This exercise is getting to the point where it’s exposing that I don’t know enough about the market, which is useful unto itself, no?
But you can see how narrowing the customer segment exposes certain pain points and can increase focus regarding the utility of the product. You then have to decide how much your narrow segment should drive the product design vs is useful primarily for go-to-market targeting. For product design, there are two aspects: 1) designing features specifically for a narrow use case; 2) using the narrow targeting to reduce scope, minimize all the edge cases you have to cover, etc. I think the answers to both #1 and #2 are very context specific.
Picking a narrow initial target doesn’t close off going after a big market — not on its own. Land and expand! I think there is an interesting balancing act between maximizing value to your early adopter while keeping an eye on that bigger market.
Yes, you want a big market, but you cannot go after everyone all at once. Not at the start. I’m a big believer in the technology adoption curve. If you start with “everyone” as a customer, that is the same as “no one”. You can’t skip over the early adopter to your mainstream adopter — the very definition of the mainstream adopter is they are waiting for the early adopter proof points.
But what if we choose the wrong market and it’s really narrow? Isn’t that putting our eggs in a single basket?
That is where research and experiments come in. Both are needed. I always say to people, if you have a strong hypothesis about A, go test that. If you really don’t know if it’s A or B, then make the tests smaller and faster and test both. Then pick the winner and keep on testing. I’m also not saying you have to pick only one, but more focus is usually better.
In other words, starting with a narrow market only limits the size of your overall market if you let it.
Full disclosure, I haven’t thought quite this rigorously in the past. Yes, I’ve always thought about the adoption curve and the importance of the early adopter. I’ve always thought in segments rather than personas. That said, I’ve kept things more vague than the above. It was listening to my fellow prod mgt professor at Stern, Andy Breen, on this topic that challenged my arguably lax thinking. I think this extra rigor is smart.
Of course this led to a question from a student: “If I’m targeting families traveling together with 4 or more people, should I not allow groups smaller than 4 to use my initial product?”
Fair question. Like all things in product, it depends. You have more questions to answer.
- Would it reduce scope / help us ship faster if we added that restriction?
- Would it help compensate for or overcome initial weakness in the product (examples: too little activity or users, too little info or inventory, a supply/demand imbalance)?
- Would it increase our brand cachet (the velvet rope syndrome)?
If you can’t think of a positive reason to add the constraint, I don’t think I would. I would design and market narrow, but allow organic usage and see if any interesting opportunities arise.
I also found myself giving feedback to students trying to narrow things down with either irrelevant information that had nothing to do with their use cases or with inner-mind psychological things that would be very hard to detect or target. You want your segment characteristics to feel actionable — either that you can find them or they can find you.
I think all of the above applies to teams working on features, not just new products. Every feature has an adoption curve of its own — sometimes they are just far more accelerated than you get with new products. In other words, companies have curves within the curve. As you expand with new segments and new features, you constantly take on new mini-adoption curves within your bigger adoption curve.
Last, more meta, point: I think it is very useful to try to push your thinking in many areas “to the bottom.” How much scope can you reduce before it’s no longer viable? How small can you take that experiment before the data is no longer useful? How narrow can you take your initial customer segment(s) before it’s ridiculous or foolish?
It’s by going too far that we discover the edges and can better evaluate the right decision.