Building for Yourself Isn't Selfish

I work in tech and I hear this word thrown around constantly: dogfooding. Terrible name. Incredible concept.

It means use your own product every single day. Be the target user. Live with what you shipped.

I built Edicek because I needed a way to capture ideas fast. I use it multiple times a day. Every time I open it, I notice what's clunky. Not from analytics dashboards or user feedback forms. I notice because it's slowing me down right now.

Something I do so often takes way too many clicks. A feature I thought was clever turns out to be confusing when I'm rushing. These things frustrate me internally before they frustrate anyone else.

That's the entire point.

Getting real feedback from users is extremely hard. They won't tell you what's broken. They'll just stop using it.

But when you are the target audience, you're not building into the void. The product grows with your needs. You feel the pain yourself.

And that's when extraordinary progress happens.

TL;DRUse your own product every day. When you're the target user, you notice what's clunky because it's slowing you down right now, not because analytics told you so. You fix problems before they frustrate anyone else. The opposite is obvious too—my Volvo's infotainment is baffling because developers built it from an office, not from the driver's seat. If you're building something, it should hurt you first before it hurts users.

When nobody dogfoods

The opposite is just as obvious. You can tell immediately when the people building a product don't use it themselves.

A few months ago I built pleasevolvofixthis.com. It's a website where I listed over fifty things that are broken or don't make sense in my Volvo EX30. Mostly infotainment issues.

I drive this car every day. And every single day something frustrates me.

The UI decisions are baffling. Basic controls buried three menus deep. Features that worked in older models just gone. Temperature display that disappears when you need it most.

It's obvious the developers don't drive these cars.

They're building car controls from some office in Sweden. They get a Jira ticket, implement it, mark it done, move on. Probably testing on an Android tablet mounted to a desk.

I do similar work. I know exactly how this happens. And it's such a shame.

This is dogfooding in reverse. Anti-dogfooding.

If you're building infotainment for a car, every developer and PM should have that car. They should drive it daily. Everything they build should hurt them first before it hurts users.

Not on a tablet in a warm office. In the actual car. In winter. In traffic. When you're rushing.

These friction points need to frustrate them the same way they frustrate us.

That's why I built the website. Not to complain. To show what happens when the people building the product never use it.