Help me run experiments

I optimize for failing fast

Madhavan Malolan
madhavanmalolan
2 min readMar 22, 2019

--

The one question that has ever since changed my flirting with new ideas is

“What will you learn if it doesn’t work out?”

Everyday we have dozens of ideas, a handful that pass our filters but very few that get our time and effort.

How do we choose what ideas to work on? I focus on leverage. We’d pick the shiniest idea that the team is excited by and one in which we have some leverage. Something that is a part of the team that makes us well placed to tackle the problem.

Commit

We have ideas every hour, each shinier than the one before. Often times the old ideas come back in new packaging. It is hard to pass on all of them, especially when we’ve seen how hard perusing the current idea is. It feels like this new idea will be much easier. Grass is greener on the other side.

We commit. We ignore all ideas that isn’t in direct line of attack of what we’re doing. Shoot them down, with a heavy heart.

Focus

I believe my odds of making this startup work is one in a thousand. To make that 0.1% count, I need to make sure I ship something great. To ship something great, we need to focus really hard on making that one thing work.

We have to make it work.

I needs more than the team’s 100%. 100% is not good enough. We don’t run parallel experiments. Running a parallel experiment will split our bandwidth between experiment 1, experiment 2 and the context switching. We’d be giving less than 100%. We need to increase that number, not reduce it.

Fail fast

There are some assumptions we’d make with an idea. How can we invalidate these ideas the fastest? The odds of them getting invalidated are higher than being validated. We might as well play by the statistics.

If we fail fast, we can give closure to the experiment and move on to a new one with no regrets.

How do you run experiments?

I’m writing a weekly blog about how I’m running my company in real life. Please do help me out if you think I can do something better! I’d love to discuss.

Follow me here for updates. If you prefer to have updates mailed to you once a week you can subscribe here : http://eepurl.com/gkBrLL

--

--