Why I Believe in Categories, Just Not Category Design

Imagine you’re walking down a beach and find a watch. You pick it up, inspect its intricate mechanism, admire how precisely all the pieces fit together, and conclude that something so complex and precisely engineered couldn’t possibly have arisen by chance: there must have been a designer.

Wait. That’s intelligent design. In today's post, I’m supposed to be talking about category design, but my reaction to the two ideas is surprisingly similar. I’m instinctively suspicious when people look at something complex that emerged from a messy, evolutionary process and work backwards to infer the existence of a designer. In biology, that process is natural selection; in enterprise software, it’s what I think of as the rugby scrum of the market, where technology, customers, vendors, partners, standards bodies, investors, analysts, and the media all push against each other until something resembling a category eventually emerges.

Now imagine, on that lovely beach walk, that instead of stumbling upon a watch, you stumble upon a mature enterprise software category, with a clearly defined buyer, a recognized problem, an established scope, a handful of resident vendors, and a generally accepted name. It’s tempting to look backwards at something that coherent and assume somebody designed it.

But nobody did. It emerged. It evolved. Marketing consultants may want to tell you that the whole thing was designed — and then (cough, cough) charge you a modest fee to help design the next one — but I think that’s mostly bunk.

Look, I’m old enough to remember the dawn of category creation strategy. Find a problem, build some software to solve it, assign it a three-letter acronym (TLA), preferably one ending in M for “management,” convince a few analysts to write about it, and off you go. Better yet, because the Valley was smaller then, if you had stumbled onto an interesting new problem, chances were that five other companies had stumbled onto it too, each backed by a different VC who, in those halcyon days before explicit kingmaking, might even allow a “may the best company win” mentality to prevail. But I digress.

This worked in the early days of enterprise software because there were so many unsolved problems and unserved buyers; the market genuinely needed new categories. But the proper strategy wasn’t simply to create a category; it was to create a category — ideally with high switching costs — and then dominate it. That’s what I call The Market Leader Play, a strategy I learned firsthand (and rather painfully) on the receiving end from Oracle.

Then Play Bigger came along and tried both to build upon and repackage a lot of Geoffrey Moore (e.g., “category king” seems pretty similar to Moore’s “gorilla”) while moving beyond category creation to so-called category design. In doing so, Play Bigger retroactively fit a narrative of intentional design onto what — at least in the cases I’d been close to personally — was a much more organic process.

The book got quite popular. I’ll never forget what one CMO told me nearly a decade ago when I asked if she’d read it:

"I'm a CMO and it's 2018. Of course I've read Play Bigger. Do you think I live under a rock?"

The problem is that popularity didn’t make the underlying causal story any more convincing. When I looked back at categories that either I or my friends had been a part of developing — and that includes a bunch of them — I didn’t see anybody designing them.

The Problem With Category Design

My problem with category design is how I’ve seen early-stage startups try to adopt it. I’ve seen companies build 80-slide decks and spend countless hours arguing over the precise definition of the future category — its name, buyers, use cases, boundaries, sectors, partners, and the associated positioning and branding.

I think category design leads founders to spend time on the wrong things:

• Naming the future category. Yes, you need to be able to answer the question, “What is it?” But any clear, consistent, and cromulent answer will do. You do not need to spend weeks inventing the perfect name. As Arnold Silverman, our wise and avuncular director at Business Objects, once told me: “Analysts name categories, not vendors.” It took me a while to realize that he was right.
• Building a category blueprint. This maps the people, processes, ecosystem, and products required to fulfill the category vision, including what the company has today and what it plans to build in the future. Look, I’m all for defining the yardstick by which others are measured (e.g., by writing an evaluation guide), but this is properly done much later, once the category has its own demand.
• Developing an elaborate category point of view. I think storytelling is as important as the next guy, and all startups need a POV that is bigger than their current product. The question is just how detailed, how elaborate, and how much of that POV should be about some future five-year state versus something closer to home.
• Mobilizing the company around category creation. This includes trying to get customers, investors, analysts, and the media to adopt your framing of the category, and using tactics like “lightning strikes” to condition the market. Once again, I’m all for alignment, but this risks people losing the plot. Are we trying to build a category here, or are we trying to build a company?

Beware Too Much Thinking About The Mushy Middle

My biggest problem with category design is that it encourages startups to spend too much time in what I’ll call the mushy middle. An early-stage company needs to have a strong sense of two things:

• What you are selling now (which includes the near future)
• Why you exist

The first is how you’re going to benefit your early customers. The second is why those early customers should take a bet on you. Call it your raison d’être, your company POV, or simply the reason you quit perfectly reasonable jobs to found the company, but it needs to be compelling enough that customers, employees, and investors want to enlist in the cause.

As I often say, “Early-stage startup customers buy two things: the current product and an option on the journey.”

But category design can lead you to spend a lot of time planning in the mushy middle. Play Bigger tells teams to “thoroughly imagine” the category they are trying to create and then produce four seminal documents — a category blueprint, a product taxonomy, customer use-cases, and a category ecosystem. It describes the blueprint as covering not just what you have today, but what you intend to build over the next two, three, four, or five years.

I’d rather have a photograph of today and a watercolor of the future. Be precise about what you sell now, have a compelling image of what you ultimately want to build, but stay deliberately fluid about the path between the two. The mistake is trying to turn that path into a blueprint or roadmap.

The Pinnacle of Planning: Tom Siebel's C3 Talk

This all reminds me of a talk I attended 17 years ago by Tom Siebel, who had just founded his new company C3, now known as C3.ai. (I was thrilled to discover that I’d taken detailed notes and published them here.)

Siebel was excited about getting the old Siebel Systems band back together, but then described an elaborate planning process, with months of deliberation before the company had even begun operating. A panel of 29 experts started meeting in the summer of 2008 and concluded its deliberations that December, with the product specification expected to be complete by summer 2009. That’s roughly a year just to write the spec. I also remember him talking about binders and binders of plans and procedures.

The philosophy seemed to be: take everything they had learned at Siebel, plan the new company in extraordinary detail, build the perfect product, hit the “go” button, and then grow at an unprecedented rate.

I remember sitting there thinking: you haven’t sold anything yet. You don’t even know whether there’s going to be a market for enterprise carbon-management software, and you’re already planning the company in stunning detail. Even more amazingly, I knew this was not how Siebel Systems had been built. Siebel was founded in response to the firsthand pain Tom had experienced as a sales executive at Oracle, bootstrapped financially, and launched with a lean, aggressive, tactical sales force.

And for all its grandiose and detailed planning, C3 failed utterly at its first mission, pivoted to become C3 IoT, and then made a second pivot to become C3.ai. They would have been far better off building a minimum viable product, getting it into the market, and jumping into the mushy middle to see what happens. In this case, they'd have quickly discovered that they were more than a decade ahead of the market, at which point they could have pivoted to something else.

Market Forces Create Categories: The Gainsight Example

Gainsight is often cited as an example of a successful category creator, and rightly so. The company was instrumental in defining and leading the customer success category. The interesting question is how: did Gainsight design the category into existence through a deliberate category-design process, or did it recognize a set of market forces already in motion, jump into the scrum, and then do an extraordinary job of helping the category develop and winning enough market share to lead it?

I believe the Gainsight story is very much the latter.

In fact, when then-CEO Nick Mehta and then-CMO Anthony Kennada tell the story in this excellent SaaStr talk on category creation, they open by saying that “category creation finds you.” Nick had previously run a SaaS company and had experienced the customer-retention problem firsthand when he learned about JBarra, a tiny software company working on the business problem we now call customer success. His reaction wasn't that he should go invent a category; it was that there might already be one emerging.

And the market signals were everywhere. The customer success manager (CSM) job already existed — Salesforce had helped create and name the role — and CSMs were already finding each other on LinkedIn and meeting in Bay Area offices to compare notes. Here was a new and growing group of people, with a common job and a common set of problems, who were sufficiently enthusiastic about their emerging profession that they were organizing themselves before Gainsight came along. The SaaS movement had created a buyer without an app.

Their industry analyst experience reinforced the point. When Gainsight first briefed Gartner and Forrester, the analysts naturally tried to fit the company into markets they already knew, such as proactive customer support or next-generation account management. Gainsight thought something more fundamental was happening and chose to follow that signal instead. Anthony's conclusion from the experience was: “it’s not the analysts that create the category, it’s the customers.” To reconcile that with the Arnold Silverman quote, I'd say: “customers create categories; analysts just name them.”

Then Gainsight did a spectacular job of pouring gasoline on the fire. They created Pulse not as a Gainsight customer conference — they barely had any customers at the time — but as a conference about the job of customer success. More than 300 people showed up to the first one, which Nick and Anthony describe as important validation that there really was something happening in this community. Pulse subsequently grew into thousands of attendees and dozens of local chapters. Gainsight built community, created content, evangelized the profession, hired practitioners as evangelists, and increasingly became synonymous with the movement.

Most tellingly, one of the charts Nick says they watched “all the time” was simply the number of CSMs in the world according to LinkedIn. As that number grew, his conclusion was essentially: this is going to be a huge category, we’re going to ride it, and we’re going to become the dominant player.

Market forces created the need for customer success, companies created the job, practitioners created a community around it, and Gainsight saw the opportunity early, built the application, accelerated the movement, and fought like hell to win the market.

Today, I’d call that category creation, because that’s how I think all categories get created. What I wouldn’t call it is category design. Nobody at Gainsight was sitting in a dark room with binders, trying to plan the future category into existence.

Go Sell Some Software

All this is why I’ve always said that the best way to create a category is to go sell some software. Build a technology that solves a problem. Find a buyer persona — often initially in a particular industry — who has that problem near the top of their priority list. Learn how to sell them your solution and repeat that process enough times to prove that you can do it.

Then discover and expand into adjacent markets. What other problems does your buyer have? Do buyers in other industries have the same problem (often under a different name)?

If a category starts to emerge, then absolutely help shape it and, more importantly, fight to win it. That’s when you run The Market Leader Play. I have zero disagreement with Play Bigger about the value of winning an important category – second prize really is a set of steak knives – but my disagreement is about how you get there. I believe deeply in categories, category creation, and category leadership. I just don’t believe much in category design.

So the next time you find yourself in a conversation about category design, think about that mature software category you found on the beach, and remember that it emerged through an evolutionary process, the result of a rugby scrum of forces acting over time.

Nobody designed it. Even though it might look that way.