If you run a technology startup, chances are that you know what dogfooding is. Don't underestimate the utility of the practice. We have seen the value of using one's own products demonstrated again and again, both in our own experiences and by other companies.
OrangeSlyce, my last startup, developed a marketplace where small business owners could browse and hire local graphic design students for freelance projects. I did not regularly use OrangeSlyce, opting to do design work myself or hire people I already knew – following exactly the old-school method OrangeSlyce competed with! The few times we tried dogfooding, it didn't go well: the design students' work was middling, the process a bit awkward, and some features went unused. Had these problems been spotted earlier, maybe OrangeSlyce would have ended up a different and more successful service.
From this comes our first surprising lesson: Dogfooding may be the ultimate form of customer feedback. Surveys, GetSatisfaction/Uservoice, usage statistics – all fine ways to peer into people's heads. But using your own product is like mind-melding with a motivated user. It’s instant feedback.
I first met Brian Chesky at a Y Combinator event where he spoke on Airbnb's early days, including the classic tale of Obama O's. But the most interesting tidbit was Brian's dogfooding story: Every night over the last month or so, he had found a place to stay using Airbnb itself! It paid off. Brian directly discovered how payment transactions were a poor experience. On the face, it seemed logical for guests to pay the host directly for the stated price. But Brian discovered how awkward it felt not having correct change, or even to simply exchange money in person. (Can you imagine a typical guest writing Airbnb headquarters to complain that paying is awkward?) Airbnb then learned how important it was to handle payments online.
Our second dogfood lesson: Sometimes pain cannot be anticipated before it is experienced. Problems can hardly be solved before they are recognized.
Here at Blogic, we are firm believers in the value of eating our own dogfood. This blog post is not only our first company blog post, but also the first published on our Blogic blogging platform. And it's already been instructive, forcing us to prioritize certain features more highly than we did initially. For instance, both of us prefer to write content in Markdown. While Blogic doesn't yet support blogging via Markdown directly (you have to convert to HTML first), we feel this need more strongly now. Not everyone sees RSS/Atom news feeds as an absolute must-have, but we found it hard to imagine our own blog not shipping with a news feed – or without Feedburner integration – so those features got prioritized, too. This post is written in the first person, emphasizing the importance of supporting multiple authors. Et cetera.
Here we find another benefit (and potential caveat). Our last surprising dogfood lesson: Dogfooding helps kill stupid platonic ideals. Really using our own software challenged some simplified assumptions and misconceptions we had about our users. “Do they really need this feature?” Sometimes the answer is “yes, from day one.”
And the potential caveat: Dogfooding may influence you to end up building your dream product, not your target market’s. But you know what? Sometimes, that’s okay. At Blogic, we are lucky enough to be building something we hope is valuable to other startups just like ourselves. Paul Graham has advised startups to first solve their own problems, and we are doing that here at Blogic. We'll just have to keep in mind that not everyone will have the exact same use cases we do. But we don't think that will be hard; that's what makes it interesting.