I had heard the term "dogfooding" for years before I learned where it came from.
In the 1970s, Alpo ran a television advert where the spokesman fed his own dog the Alpo dog food. The dog loved it. The message was simple: if it is good enough for my dog, it is good enough for yours. The term stuck, and somehow made its way from pet food advertising into the vocabulary of every product team in tech.
It is a strange bit of etymology. The world of startups moves fast, obsessed with the new, yet we still use a term from a dog food advert over fifty years ago.
But the reason it stuck is because the concept is fundamental. If you are not using your own product, you are building blind.
Dogfooding catches things that testing misses. Automated tests check whether something works. Dogfooding tells you whether it is any good. There is a difference between a feature functioning correctly and a feature being pleasant to use. You only feel that difference when you are the one using it.
It also builds empathy. When you experience the same friction your customers do, you prioritise differently. That annoying extra tap to complete an action stops being a minor issue in the backlog and becomes something you actually want to fix.
The best product teams I have worked with treat dogfooding as non-negotiable. Everyone uses the product. Not just the PM, not just the designers. Engineers, leadership, support. When the whole company feels the rough edges, those edges get smoothed faster.
There are limits, of course. Sometimes you are not your own target customer. A B2B enterprise tool might not fit neatly into your own workflow. But even then, there are ways to get close. Use the product in test scenarios. Shadow customers. Sit with support and watch what people struggle with.
The principle remains the same. The closer you are to the experience of using your product, the better your product will be.