Sunday, February 26, 2006

The Trouble with Dog Food

Eating your own dog food is a popular practice in the software industry. It's a good idea, but like all good ideas it can be taken to harmful extremes. With that in mind, a few thoughts about the dark side of dog food:
  • The more a developer uses the product the more comfortable they'll make that product for themselves. This is great if the developer is representative of the target user, but a problem if they're not. The thing is, they're more often not.
  • Extreme dogfooding, where you use your own products to the exclusion of all others, gives you a warped perspective on the market. You should be familiar with all of the products in your space, and you should use the ones you like the best. . . After all, that's what your customers will do. Anything else is just lying to yourself. There's a fine line between eating your own dog food and drinking your own Kool-Aid.
  • In a large company with multiple teams, remember it's only eating your own dog food when it's the product you're working on. When you're not in control of the product in question, then all you're doing is beta testing.
So go ahead and eat your own dog food, but make sure it's part of a balanced meal, and remember that it's no substitute for actually paying attention to your real users.


I would generally agree, particularly given your last qualifier... always balancing your own findings during Alpo-munching projects against what your audience/customers are telling you from their own experience, objectively- and maintaining subjectivity in your project/discovery process to avoid becoming your own shill.

In fact, I don't think you can really pull off projects like this at all without addressing external validation points at every possible step- otherwise it's essentially a rigged experiment that will always skew towards your own workflows and practices - which are in turn highly skewed from the real-world by your direct access to
the development teams building the 'dog-food' products, so to speak.

The only way I can do a 'dog-food' project (and I've done a LOT over the years) is from within a completely separated team/group from the product team(s) building the product(s), and essentially 'becoming a new customer' - all the while openly discussing these findings with external customers to balance out any myopia I may have fallen into from my internal vantage point.

That being said- on a personal level most of my fastest ascents up a learning curve have been when putting these types of arbitrary restrictions around myself... so I don't see myself stopping the practice any time soon. ;-)
True, true. It's all about balance.
Post a Comment

Subscribe to Post Comments [Atom]

<< Home