We’re building yet another recipe site (with a few twists). Clearly madness, I can hear you thinking. How many people do you think will use this? Well, so far we’ve got exactly 1 user (one of us) – however this is 1 more user, who really wants and cares about this product, than many other corporate software applications I’ve worked on.
How can one real user beat hordes of hypothetical companies ready to pay anything for your 
software? To start with, I seem psychologically incapable of caring what an imagined person or, even more so, an imagined company *might* want one day – that is, if the stars align correctly and you slip through the corporate software sales process (think feature checklists and friends in the right places). In fact, the only reason I’ve been able to develop anything inside a corporate that I was proud of (note: proud = positive difference to a real person and sharing the excitement) is because I’ve been in positions where I build things to directly help the people I work with every day.
However, back to the point (temporarily I’m sure). With one user in the same room, we can talk, share the problem and see the software used in a real scenario. This gives me a tiny chance of avoiding creeping featuritis. For example:
Geek: “Wouldn’t it be better if we had wiki-like versioning for recipes?”
Non-geek: “Ah sure, that would be cool – I wouldn’t ever have to worry about losing anything.”
Geek: “But what would be even better is if we had multiple users having different views of the same recipe but able to incorporate snippets of each other’s changes as desired.”
Sounds somewhere between good and insane (in the membrane, insane in the brain [Cypress Hill]). Thinking about it from the hypothetical perspective, it’s very easy to think that this would be a really cool feature and we all know that cool feature = people love your software and $.
However, when you mention this to your one real user, she gets really quiet and her eyes glaze over, as she’s too polite to say “Sorry, I think I must have misheard you… you want to spend the next 6 months building a distributed source control system into a recipe site – so I can enter text!?!?” (Her exact unspoken quote).
The simple message – one real user gives you a small chance to continually work at cutting away the deadwood that will otherwise overwhelm and kill your project.
Tags: development, software