Earlier this week I was speaking to a friend working on a new service offering for customers. It was a subscription plan with weekly checkin calls and as we were speaking, there was a want to automate.
I love automation, it’s why I have a bunch of apps that can replace text and make things happen in the background… but it has to be the right time.
When we don’t know what our customer wants, automation is crippling. This is the time to take a page from Paul Graham and do things that don’t scale. Once we figure out what works, then we can hard code that, until then… duct tape and having SCI (swivel chair integration) of a bunch of people manually doing work is what makes the most sense sometimes.
Once we have it right, then we can get fancy with teams and scaling and all of the stuff I find a lot more fun.
What should you be testing and NOT automating?