September 7, 2022
Last year, the product development team at Underdog.io had a frustrating realization while working on a new customer-facing feature. We were building a simple tool that allowed companies to save searches in their Underdog.io dashboard:
"Looking for backend developers that know Python and JavaScript, consider themselves to be members of an underrepresented demographic in tech, and want to work remotely? Create a saved search and we'll notify you of new matches."
This tool would accomplish two goals for us: (1) give users a fast and accurate way to filter down our feed of active candidates, and (2) make our internal team more productive by reducing the number of manual searches that we run for customers. Seems pretty straightforward, right? We talked to users and then planned, built, and released the feature only to see it receive little adoption from the bulk of our users.
Upon reflection, it turns out that we built the product version of something when what our audience actually liked was the services version of that thing. They wanted someone to do the work for them, not learn how to use a tool that makes it easier for them to do it themselves. This is a recurring issue for software-enabled businesses like Underdog.io: the path of least resistance often looks more like a service than a feature. The more you try to productize aspects of your work and increase your gross margins, the more pushback you get from paying customers.
Based on our experience at Underdog.io, the best way to bridge the gap is to focus your efforts on properly onboarding and training new customers, the ones that only know the productized version of the service, so that they can use our platform on their own. For existing customers, we found that the best way to overcome inertia was actually to build even more tooling; in this case, it was an admin tool that allowed us to pre-populate saved searches into users’ accounts and lighten their upfront workload. We still have some stragglers, but we’re slowly chipping away and pushing them toward the self-serve model.