After posting about Dogfooding the other week (which if you haven't read yet I recommend doing so now, it's really a short read), one of my favorite development focused podcast shows covered the same exact topic while reviewing another blog post on the topic.
After both listening to the above episode and also talking with Scott Kaye about the topic, there were a number of interesting takeaways that I had:
New Customers
Dogfooding is great at improving the product for existing customers, but what about attracting and improving the conversion for new customers?
Scott got me thinking about this more.
Dogfooding can still be a useful exercise for attracting new users, see for example the note from my original post on how Graphite forces their employees to go back through the account creation flow every so often by deleting their account on a scheduled basis.
Famously also, I think either Stripe or Shopify have shared that they have new hires build some kind of product using their offering (usually focusing on a specific part of their offering). They then have these new hires feed back in improvements and fixes to make their feature better for others and also improved documentation.
The key insight I think is to make sure to play the role of a new user as part of the dogfooding - often this means having you start fresh without an account, but it could involve other aspects too!
In combination with user feedback
From the episode, they talked about collecting user feedback - while I talked briefly about this in the previous post as well - I wanted to re-iterate that dogfooding alone won't magically make your product/feature better across all axes.
One pitfall of only dogfooding and not listening to real customers is that you may end up building a bit of a monoculture or a single solution to a subset of the problems that your customers may have. Sure the resulting product may fit all your needs, but you may also have greatly reduced the percentage of potential users in the process.
I believe that you should do both - dogfood actively but also listen to and act on customer feedback. While sometimes it can be difficult to actually understand what users may want (speech is a lossy medium after all), but none the less it's still worth capturing and listening to.