Acquisition vs retention - two types of features we build in a B2B startup

Two important things to mention about B2B software before we start:

1. In most cases, there are two distinct types of stakeholders: decision makers (aka budget holders) and users. Decision makers care about spending less and earning more, users care about getting their job done quickly and easily (oversimplifying like hell, but please indugle me).
2. There are two distinct types of B2B software from the adoption perspective: user-driven - think Figma and Notion, and decision-maker-driven - think most of the B2B software).

Back to my realization - it happened during a 2023 planning discussion with my manager. I, being a very opinionated nuisance, was pushing for the roadmap to be driven by user research findings, product outcomes connected to business outcomes, product strategy, etc. In other words, I'm trying to follow the hottest trends in a conscious approach to product development and stay as far away from the feature-factory, biggest-prospects-driven roadmap planning (which often is a positive thing). Still, in the back of my mind, I was ready to admit to myself that I don't have a good understanding of how the true rockstar features, big innovations emerge. It concerned me that most of the hypotheses I come up with after talking to users and staring at product outcomes result in small to medium-sized improvements or features.

One major reason why I think rockstar features seem to never come out of user research is the fact that such features are rare in general. If you value high chances of success for a feature, you need to look at lots of data and churn through tens of small to medium improvements before you stumble upon an insight that will lead you to something really groundbreaking. It also could be so that a truly valuable product simply doesn't need to come up with a new "industry-shaking" feature every few months to solve users' most important problems. My kettle has been literally the same for years, but it does a perfect job of boiling water and then pouring it into a cup.

The problem is that almost no one (me included) has time to organically arrive at the key insight that will trigger the awesome feature that will delight both decision makers and users. Instead, what we have is a perpetual need to be "cool" (more on this below).

All of this got me to understanding that there are two separate types of feature work: 1 - sexy and shiny big features that are barely supported by focused research or any "proper" evidence, 2 - research-backed, meaningful user experience improvements that are hard to get excited about when making a decision to invest into a product. Naturally, type 1 or Acquisition features are for decision makers, and type 2 or Retention features are for users.

The thing is, a B2B startup can't live without either of the types. We can build improvements for the actual users all year long, but this would be an outright murder of the sales team's ambitions. Any headless CMS is a decision-maker-driven product, and the actual revenue is brought in by the sales team convincing decision makers that they should put their trust in our product. Because of that, things like the “smooth and frictionless process of creating long content entries” aren't going to fly because it's not something that is easy to demo and sell. Instead, we need something to grab decision makers' attention and create the 'wow' effect that will help us win deals (this is the being cool part).

At the same time, if we oversaturate the product with deal-winning features, the users will inevitably notice (and, most likely, they'll get to it very quickly). If you have a shiny sportscar, big muscles and five memorized breathtaking stories you stole from movies, but then can't remember their friends' names and act like an asshole in front of the parents, your relationship won't last long. Same with software, if the actual day-to-day is shit, customer retention will plummet and may even affect your sales numbers down the road (depending on the size and maturity of the market you're targetting).

What does it all lead to? To embracing randomness when deciding what to build. We need rockstar features, but we can't use the usual research methods to decide what to build (because of time constraints and difficulties associated with researching decision makers). The only way out is to get comfortable that due to the nature of decision-maker-driven adoption, those acquisition features are going to be bets that may or may not work, and the success of which is highly affected by randomness. It is possible that a feature that costs six months to develop will flop tremendously in terms of usage, but then will be the last deciding factor in our favor in a quarter-saving deal. Because of that, the added value of focused research is minimzed, and the guess of a product manager is as good as anyone else in the company.

Now a few words on bets: This article from Shreyas comes very much in time for this post of mine - it would be absolutely unfair to say that bets that are largely based on instincts are a bad thing. Great instincts are often developed by continuously getting in front of your users, studying the market, adopting new technologies as you basically train your brain like you would train an AI to later make lightning fast decisions that seem to be random guesses to outsiders. This way at times you don't require any focused research (related to the quarterly OKR or the new strategic bet) to make a solid decision about what to build next.

With retention-focused features, the approach is the opposite: since we care deeply about features and improvements to actually improve users' lives (and positively impact the related metrics), we need to increase our chances of success by lowering the uncertainty and perfectly understanding the user needs and wants. In this case, the value of research and conscious decision-making is amplified, and the success of such features is less random.

Ultimately, this two-feature realization leads to peace of mind. I know that the big and shiny is a necessity and I'm coming to terms with guesses. You can still get excited about these features by building them the right way and increasing the chances that users will actually benefit and find the decision-maker fishing hooks helpful in their routine. The most brilliant idea, with no execution, is worth $20.