Lessons I learned at Prolific - part 2
This is my second post on lessons I learned at Prolific. This time it explores, decision making, the factors that drive success, scaling challenges, and risk.
Some of these will be topics to dive into in further posts, let me know what you find interesting, or would like to hear more about.
At some point there might be a part 3 around hiring, people, and growing teams. and culture.
Business models are as important as product
Sometimes you hear people say “ideas are worthless, execution is everything” - this is wrong.
Execution is important, but succeeding in a startup, is a combination of fast execution, good decision making and good ideas. The two skills are in tension, the best founders need to sit at the right place on the spectrum for execution versus smart decisions. And the right place to be depends on the company and context.
Another phrase you hear is “first time founders think about product, second time founders think about distribution”. I might also tack onto this “investors think about differentiation and defensibility”.
YCombinator is heavily focused on executing on the product fast - “talk to your customers, build something they want, keep shipping fast.”
It makes sense, because this comes first in the process, without some kind of product-market fit, you are going nowhere. Their risk profile is also different, they are investing in huge batches of companies. Markets and models can be figured out later to capture whatever momentum emerges.
You’ll hear a lot less discussion about pricing or business models at YC. But business models can be powerful from the start, and getting them right can be the highest impact decisions you will make as a startup.
I mentioned in Lessons part 1 - the powerful growth driver for Prolific was the network effects of the marketplace platform. Because we focused on individual researchers rather than institutions first, we had the ability for customers to use the platform without any sales process. We continued to making tweaks to the way the pricing and revenue model worked as we grew, with big compounding effects. In combination this allowed us to grow entirely from revenue, in the early years.
Most early stage startups are not very defensible. Does it matter?
At the start it’s often very unclear how a startup might become defensible, or how it’s different from others.
Many forms of defensibility arise from making good choices about the business model. Examples are network effects, platform effects, partner integrations, regulatory, sales/services lockin, proprietary data etc.
Simply continuing to move fast is a form of defensibility and differentiation in itself. and in the unpredictable future market of new AI startups, it might be the only approach you have. Still you want to be wary of rapidly building a commoditised, low margin business, or you’re not going to like where you end up.
One way to start ahead is to begin in an unattractive (perceived as small) market, like academic research data in 2014. Niche markets can be off putting, but gaining strong product-market fit in a niche, can often be a great springboard to wider adjoining markets.
Few things matter a lot
Only a few important events and decisions really made the difference to success, surrounded by a lot of noise. This was hard to see at the time, but hopefully these posts might help you spot some of those inflection points ahead.
Outcomes for startups follow a power law distribution.
The power law fundamentally shapes how startup founders and investors should approach decision-making.
Resources, capital, talent, and risk should not be allocated evenly in startups, but concentrated where they can create exponential impact.
The tricky bit is to see which decisions might have asymmetric consequences, and make them well.
I recommend reading the Black Swan and spending a lot of time at underground poker games. You can also practice using a decision making framework like this one favoured by Patrick Collison to learn what decisions might fit this profile. If you develop any edge at decision making, then making more good decisions will move you ahead of the competition.
Learn to get a sense of the few decisions that need a lot of time spent on them, and avoid getting bogged down by those that don’t.
At Prolific, among thousands of choices, only a few made the difference to direction of the company. - about business model, product features, which markets to focus on (embracing AI), how to fund the business (bootstrapping then YC, and later a big series A).
Unfortunately what tends to happen is the opposite, in what’s know as “the bike shed fallacy”
"A nuclear power plant proposal receives minimal discussion and quick approval because of its complexity. A simple bike shed proposal triggers lengthy debates about materials, colors, and design since want to contribute opinions"
Time spent discussing an issue becomes inversely proportional to its monetary value or importance
The decision making culture at your startup will influence this, who gets to make specific decisions, how democratic are they?
You don’t have to get very big before the answers to this are not clear. You’ll experience a subtle shift as you go from a tight small initial team, where a lot of strategic alignment is implicit (around ~20) and some hierarchical structure begins to emerge. For any RACI framework there will also be subtle cultural political framework of consensus and alignment. Startups can feel like they’ve run into mud at this point.
Pain can't hurt you, but what you can't feel might kill you.
A lot of things that are on fire should be ignored for the few things that matter. People will get very upset about some of this, but you need to remain focused. You also need to keep realigning your team's focus, while building a strong decision making culture.
Are you even seeing the decisions that are implicitly being made, and drifted into, or are you intentionally surfacing them? One practice we introduced at Prolific was intentional decision logs for major decisions.
Most people are familiar with the concept of technical debt, but startups also accumulate organisational debt.
You need to decide which fires need to be left to burn. You are going to have to get used to the discomfort of ignoring most non-existential risk yourself. You will also have to convince your team it’s fine, to keep them executing. Otherwise the lack of confidence and alignment can itself lead to chronic culture problems.
As companies grow, and more experienced people join, there’s also a natural growth of process, and formality. Unless you continually assess what’s actually helping at the current stage of growth, there’s a danger this turns to bureaucracy and stagnation, that will be fatal. It’s a dangerous siren song, because at the time if feels like what a “grown up” company run by experts should be doing, and hence a sign of success.
Brian Chesky described something similar in his now famous Founder Mode talk.
The reaction to this, where a Founder seizes more direct control, for a hard reset, can also lead to a painful cycle of destructive evolution.
Early decisions can live with you a long time, and have severe consequences. Some are inevitable and conscious, some create invisible fragility, some can be avoided.
Technical shortcuts and manual processes might be the only viable way to get a product in front of your initial customers. “Doing things that don’t scale” is rational as you work towards product-market fit.
Hiring decisions, and cultural shifts (intentional or not) will stay with you and echo on much longer than you might think. The effects get bigger the larger your company grows, something I’ll explore in a later post. You need to make sure your team aren’t learning the wrong lesson, and continuing to “fight the last war” when a new threat lies ahead.
Company formation and governance decisions can have long term consequences that may even become existential. Look at the challenges OpenAI is having fixing these right now.
Where to set up your company, how to arrange your cap table, and shareholder/board agreements are critical. This may not seem very relevant when starting out, and isn't always predictable. This is an area where you need a really strong reason if you choose to diverge from standardised mainstream approaches. Standard legal structures have evolved through an process that killed other startups, so take advantage of their lessons.
