I recently came across a product called Retool that positioned itself as “the best product for building internal tools without coding”.

That’s a super bold pitch by them, because I really see their product as useful for many other use cases. So why would they narrow it down? Well, no product has ever pitched me as a replacement for our internal tools, and now I’m really curious to use it for that.

This is opinionated product positioning at it’s best.

Tell me who the product isn’t for

Take onboarding. A lot of products will show you tooltips and ideas about what you can do, but nobody ever tells you what you can’t do. That means you’ll have to either test it extensively or just guess if the product is for you or not.

Lookback’s beginnings were a lot like that. We were pursuing so many different use cases for the product (QA, feedback, user research, etc) that eventually we had to say: STOP! Let’s pick one and run with it. We confused our ‘end game potential’ with ‘the way to get there’.

So the trick is to be more opinionated. Build a product that’s very clear on:

– who the product is not for
– what the product cannot do
when not to use the product

Like in this example from Canva, a design app. When creating a new design, I can only choose from one of their template categories. There’s no “custom” option.

It’s an elegant way of communicating that if my category isn’t available, the app isn’t for me.

Adding “website design” as a category in Canva is probably trivial, but I’m guessing that it sits just outside of their comfort zone for what the app should do. So they’ve made the opinionated choice to not include it.

Or, take a utility app like Board Game Stats. The name may not spark joy in me, but if you’re a board gamer, you’ll immediately feel right at home. Compare that with startup names like “zwift”, “sgnl” and “xensr”, and, yeah, you get the idea.

Chances are that if your product is capable of doing more than it says, I’ll start using it for more — but you need to help me find and love the best use case first.