Just push your brain until it hurts!

March 2, 2013

Like any other startup, we are constantly trying to solve ill-defined complex problems. Sometimes I spend weeks and weeks just trying to figure out what I’m optimizing for, let alone how to do it. Then it hits you. You have that awesome moment of clarity! You draw it on a whiteboard, snap a photo of it with your iPhone. Think about it a bit more later that night. Re-draw it on a piece of junk mail on your kitchen counter. Snap another photo of it. And voila… the next great idea is right there staring at you!!! You run to the office the next day. All excited to get the project started. It’s developed, tested and released to a percentage of users in a sprint or two. I should feel good once it’s launched, right?

Wrong! That first part was actually the best part. The shitty part is about to start. When rubber hits the road, you face one of the following 4 scenarios — sorted from the most likely to the least likely:

Scenario #1– The idea falls flat. Complete failure.

Scenario #2– The idea doesn’t quite work but you still have a tiny sliver of hope — because of an unexpected twist. That is, customers are doing something quite different than what you originally imagined. And that new direction might be slightly net positive.

Scenario #3– The idea doesn’t quite work but you still have hope. This time, you feel like your original hypothesis might still be correct. But you don’t know the correct combination of parameters to make it work. Even worse, you are not sure whether that combination exists.

Scenario #4– The idea works exactly as intended.

#4 happens so rarely that I won’t waste much time on it. And in the unlikely event that it in fact happens, you probably know how to celebrate! Even if some optimization is to be made, believe me, it’s a million times easier to optimize something that is already working.

Scenario #1 is the most likely scenario. Unlike popular belief, this is the easiest scenario to deal with. There is nothing to be done except to deal with your own emotions. If you’re a pro, you know the drill: Go home. Get yourself a glass of scotch on the rocks. Wrap yourself in a blanket (or do whatever you need to do to look miserable). Pity yourself. Sit in the dark and watch some cheesy reality TV show on hulu until you fall sleep. Tomorrow is another day. You’ll get over it. I always have.

Scenarios #2 and #3, however, kill me! You’re not getting much traction on the first version of the product but you can feel that there’s a tiny opening to push your way in a new dimension. The idea didn’t quite work but it’s not dead yet. You just have to continue examining it! Cannot… let… this… thing… go… yet!

This is mentally taxing for two reasons:

1- You don’t know for sure that there’s something there. You just have a hunch. And continuing to push yourself and others based on just a hunch is difficult. “Am I beating a dead horse?” Like I said, it’s a lot easier to work on something that’s already getting traction!

2- Most of the time the problem is so large and there are so many parameters/conditions to play with that it becomes mentally intractable. You cannot possibly test all possible combinations. Even worse, it’s never a smooth continuous function so that you can “hill-climb” your way out of it. You’re walking on a non-continuous surface in a large multidimensional space. You move one parameter slightly in one direction and all of a sudden it changes paradigm. And you will never know if you have fully optimized the system. How would you? What is it that you’re optimizing exactly? Is there a god? Where do babies come from?

Now to make matters worse, add the fact that you cannot simply move on from the problem either (like you would in scenario #1). Scenarios #2 and #3 are where most of the real progress is made. I would say at least 80% of overall forward movement. This is the stuff. So what do I do?

My first thought is to crawl into a ball and cry. But that won’t get me anywhere.

So, over the years, I’ve manufactured some rationale to make me feel better. And when I make myself feel better, I can continue to push in the dark. And more often than not, when we continue to iterate on a half-alive idea, we end up finding a bright solution — to my own surprise most of the time. Here’s what I say to myself. They work on me — sugar pill or not:

I DON’T have to reach the optimum point!!! Most real-world problems are too big to ever be fully comprehended and solved. You’re not even sure if you’re seeing the full picture, or if you’re solving for the right variables. But here’s the good news: I just need to find a configuration that works slightly better than what we had last month! Who cares whether or not I’m digesting the whole picture? Or how much better it can get? All I should care about is to find ONE configuration that works slightly better than before. Just squeeze a tiny win out of it. Then move to that new point even though it’s a tiny negligible win. The next step will reveal itself to you magically! It always does.

– If we’re hurting our brains trying to scratch the surface of a problem, even if we find some sub-optimal approximate solution, we might still be light years ahead of our competitors. Yes, I cannot fully solve that problem. But over the years, every time we have talked to other companies who are facing similar problems, we were always surprised how much deeper we attempted to dig into those problem than others.

So here’s the key: A problem doesn’t have to be fully understood, dominated, and solved for you to succeed. Just push your brain until it hurts! Other people are probably not pushing themselves that hard. People don’t like hurting their brains!

%d bloggers like this: