How much effort should early-stage startups invest in backend scaffolding versus taking on technical debt?

Question

Accepted Answer

Hey. First, thanks for all your participation in the community—it's great seeing you so active.

It's an interesting question, and I think my answer may not be typical, so take that for what it's worth. But you shouldn't spend a lot of time on your support and your scaffolding until you know exactly what it is you're building.

For example, on a Hollywood set, they don't spend a lot of time making sure something is going to last 30 years or that it can support day after day of being in the attic. It just needs to look right. They create the impression of a busy Western town.

At some point in your product development, you're absolutely just trying to create that same impression. You build a busy Western town because you're trying to see how customers will behave. You're trying to see how they react and which features they use, and that does not require a long-term investment. If you make sure the scaffolding is strong right away, sometimes you end up building that strong scaffolding in entirely the wrong place.

Another good example we talk a lot about is when people build new office complexes or school campuses. They purposely don't put in the sidewalks at first because you're just guessing where the sidewalks should be. They let people walk and create their herd paths in the grass, and then after a few years, they use those paths to know where people actually walk and what they need to build.

So my answer is yes: accumulate technical debt. At some point you're going to realize, "Oh my gosh, this is how they're using the product, where they're using it, and what's important." Then you're going to have to go back in to create that scaffolding, create that framework, and build that strong underlying layer you need.

There you go. There's my note. And by the way, as we're recording this, you know I'm actually real.