June 2022
I love a good experiment.
Experiments test risk. Experiments guide prioritization. Experiments drive growth and experiments keep failures small.
Some 10-person companies can ship 100+ experiments every day. I’ve seen it 🙋♂️. New product features, ad copy variants, internal processes exploration, meeting formats, team celebrations... Shipping experiments creates momentum, and momentum creates energy. For the right collection of people, walking into an office that thrives on experiments feels closer to playing a video game than working.
The reason why so many teams avoid experiments is that many experiments fail. Without the proper framing, failure bruises egos. At NASA we accepted a robotic failure rate of 50% for deep space missions - for high-profile missions we’d send two redundant spacecraft to the same destination (MER, Voyager, etc) in hopes that we’d find success with one spacecraft. In some risk-tolerant environments today, perhaps a 99.9% failure rate is acceptable for certain initiatives. It is a deliberate decision how much risk to wager on experiments. But too many teams leave this up to chance. The best teams develop a portfolio strategy to manage experimental risk: A small number of high-risk, high-reward experiments, some moderate-risk experiments, and some obvious highly-predictable experiments. And they re-evaluate their portfolio health frequently. If you lead a team, this portfolio review is your job.
As an employee, the next time you interview at a company, test their culture of experimentation. What is their experimentation volume and how do they manage their experimental portfolio? How do they surface results to the organization? How are experiments prioritized? Ask team members on your interview panel to walk through a recent experiment. Experimentation autonomy is a great proxy for broader workplace autonomy. If your appetite for experimentation doesn’t match that of the company, you should consider whether it’s a good fit.
As an executive, how do you develop a culture of experimentation? Lead by example: Communicate and celebrate both successful and failed experiments. Give your team permission to run their own experiments. Talk about experiments with your customers. This topic deserves its own post.
As an early-stage founder, your goal is to quickly validate and survive long enough to upgrade to hyper-growth. How do you know you’re making progress? How can you ensure your team is focused on the right things? How do you find 1000x opportunities? Run more experiments with a clear hypothesis and focal metric. Failing to build a culture of experimentation with your early team at an early stage is tough to recover from later.
I’m surprised that we still call early-stage venture-funded companies ‘startups’. If customers aren’t lined up around the block yet, the entire company is still an experiment. A ‘startup’ is closely linked with founders ego, we get emotionally invested in the success of the company the moment we describe a company as ‘my startup’. When that first TechCrunch article arrives, suggesting that you’re changing an industry, people get locked into the stated mission. When in reality you might just have an MVP and a few customers. Re-framing a startup as an experiment helps to insulate egos - if the experiment fails, it’s not personal and you have permission to run another experiment. Experiments are a vehicle to test a hypothesis, not measure self worth.
Which brings me to Staffparty. We’re doing things a bit different. My goal is not to build a startup, my goal is to validate that a handful of products that we want to see in the world provide significant value to customers. If we do that, we get permission to build a startup. If not, we’ll go run some other experiments.