From Luca Dellanna’s EconTalk appearance on Compulsion, Self-Deception and the Brain:
Try a lot of things and see what works, see what sticks. And, at the beginning, don’t necessarily look for effectiveness. Look for what sticks, what you keep doing. And then, only in a second step, you can look for effectiveness.
One common mistake in the startup world is to optimize before having found a product/market fit–a product that the market really wants, that they start pulling. And, only then you can think how to advertise it better, how to make it cheaper, and so on. But, if you try to optimize too early, there is just too much friction because you don’t have a good product enough. You don’t have a product that the market wants; and you push it, but the market doesn’t get it.
And, the same applies in some way to our habits. If you try to look only for the habit which is the most effective, but it’s not a habit that your brain is, like, receptive to, then you might waste a lot of effort.
I agree. I’ve been involved in a lot of idea generation sessions, one this week even, where this happened.
There’s a strong tendency for people to want to think down the road to the finished product or write-off an idea if they can’t immediately see how it can fit into the current system or if they think it will cost too much to make.
It’s often hard to reel them back to step 1. Forget about finished product, cost troubles or how it will fit in the current system. Step 1 is to figure out if you got something that works, on a really small scale.
This is another hurdle. Lots of folks still want to jump to step 3 or 4, imagining what a market test would look like.
Think smaller. What’s the minimum you would need to do for a proof of concept and very small pilot? That could be as simple as mocking up the idea on some paper and talking to 5 or 6 random friends about it.
Many think this is too small. They want fully baked market research.
Some of the tendency to jump over the early steps comes from folks who just don’t know better. In big companies with big resources, they large test is considered a gold standard.
But, I think some of the tendency also comes from another incentive. I’ve worked in companies where people got brownie for simply proposing smart sounding ideas, but had not ownership of whether the idea got tried or worked, or not. That was someone else’s job. But, they basked in the glow of of their great sounding ideas.
Talk is cheap. Have them put their money where their mouth is.
They don’t want to test the idea because they want credit for coming up with the idea. Actually giving it life, even in a small way, terrifies them because it might fail. So they try to optimize for what the final product will look like because they know that finding out is in some far out future and nobody will remember who had the idea by the time everyone finds out it’s a dud.
In one company that I worked in, field management had lots of discount ideas, many of which have been tried many times without. We had a hard time getting them to pay attention to that info as they were too busy accepting the praise for having an idea.
When field managers came to me with ideas, I started asking if they would fund a trial from their budget.
About 90% of ideas died on the spot. When faced with the prospect of paying for their idea, they became more interested in the results of previous attempts, calculated the hole that would put in their budget and refrained from spitting those ideas out in meetings for brownie points because I might ask them to fund it.
One time a field manager did have an idea that hadn’t been tried before. We asked if he’d be willing to fund it to try it.
I was skeptical of the idea, but he was taking the risk. Also, I know I can be wrong, so I didn’t let my skepticism get in the way.
It turns out it worked, spectacularly. He never let me forget my skepticism. I never let him forget that I didn’t let that stand in the way of trying it, nor did it keep me from digging to discover why it worked.
To Luca’s point, when we tried, it wasn’t fully baked. There were lots of problems in execution. It was late getting out the door. There were systems problems.
But, there were strong signs that he was onto something. None of those early problems hurt the signal that. The signal was one of the clearest signals I’ve seen in all my years of looking for things.
Folks who haven’t seen such signs tend to get excited about 1-2% bumps. I don’t anymore. They can have those. I’ve seen too many statistically significant 1-2% bumps that evaporate into the noise of the real world when they are rolled out and they have to resort to the smart sounding, but empty defense, “just think how much worse it would have been if we hadn’t done it.”
These are the same folks that want to optimize before finding out if they have something.