Link Search Menu Expand Document

Don't half-ass many things

Half-ass one thing many times.

The purpose of a startup is to find a new and better way to do something

Before figuring out the product-market fit, startups tend to flail and try many different things. Often they prudently use the 80-20 approach.

But constantly doing new things is bad

You end up with too many processes and features, none of which exactly hit the spot. All of them have been 80-20ed.

Many different features or processes also take time to maintain. 20 different processes are more difficult to maintain than one process with 20 steps.

It also becomes more difficult to recognize when something is working.

And because you have to juggle so many things, your attention isn’t focused on it either.

It’s also less straightforward to train and onboard your users and your team.

Every new thing has a fixed cost regardless of whether it works or not

Every new initiative effectively has a starting penalty. It takes time before an iteration that works is developed.

While it’s much easier to start a new initiative than to find a 1% improvement in an existing one, it’s also not the right thing to do.

Instead, it’s better to zero in and focus on what already works and improve it a bit instead of grasping at a new thing that might work.

So, instead, build and re-evaluate more frequently

Once a feature is built, give it a limited amount of time before it starts paying off.

If it does not start paying of you should consider removing it entirely.

If it does, then it qualifies for another round of 80-20 to improve it.

It’s this sequence of 80-20 improvements to processes and functionality that build a successful product. Not many 80-20s on many different things.

Related Lessons

Further reading

  • The Prune Date
    Features come and go frequently during the lifetime of a product. Except that they often don't actually go but just stop being used.
  • Your product is a joke
    Lessons from improv about honing in on specific benefits

Suggest an improvement to this page (