Sometimes it makes more sense to pay for products than to build them yourself. This might sound obvious, but it can be a hard thing to realize in the context of a bootstrapped company like Zapier.
One of our core values is "Don't be a robot, build the robot." The basic idea is you shouldn't do repetitive tasks yourself because you can build a tool to do those tasks for you. This is how we try to operate internally, and it's also the main service we provide to our customers.
Erik Aybar, an engineer here at Zapier, recently suggested this value might not be enough anymore. He wrote an internal blog post: "We should take this even further, when appropriate: 'Don't build the robot, buy the robot.'"
There's something to that, for small and medium businesses and rapidly growing tech companies alike. Let's dig into why.
The building mindset
Zapier is not a small business, but I think we have more in common with small businesses than we do with other large tech companies. We bootstrapped our way here, the way small businesses do. We also became profitable by selling a service to customers, which (for some reason) is unusual by Silicon Valley standards.
Part of that might be that we're not entirely a Silicon Valley company. Bryan Helmig, co-founder and CTO at Zapier, suggested our origin story is probably part of why we've always defaulted to building things ourselves:
Coming from a fairly humble midwestern upbringing, it's really hard for me to shake the bias of "You want us to pay for that?!? I could do that myself in like 2 weeks AND I'd have fun doing it!" But there's an opportunity cost associated as well as the long-term maintenance cost of building everything yourself.
It's easy to think of tools you build as "free" and tools you buy as "costing money," but that's not the case. As Helmig points out, any tool you build internally will need to be maintained, and that's far from free.
So this has prompted some internal thoughts about the "build the robot" value. When does it make sense to do manual work yourself, when does it make sense to build a bot for the job, and when does it make sense to pay someone else to automate things for you?
Helmig broke it down like this:
When to be the robot: You're working on something new and unproven or impossible for robots to perform, or there is a finite amount of complexity you could de-risk by just rolling up your sleeves.
When to build the robot: You've got a team to support it, you've read this xkcd comic, and you're creating a differentiated Zapier capability—or, there's a Zap for that.
When to buy the robot: You're doing something pretty standardized that most companies need to do, and you don't think we should compete with external specialist firms that provide such capability.
We're very much still figuring this all out. Every company needs to. That's what makes it interesting.
When to buy the robot
The Zapier Blog, which you're reading now, was until last year hosted on a Django-based platform our team built internally. It worked, but maintaining it was a lot of work, so we migrated to Contentful. It wasn't easy, but it's a switch that's going to save us time and money in the long-term.
Businesses make these sorts of decisions all the time. Judith Miller, who used to own a nail bar, switched from a thrown-together spreadsheet to a proper CRM for tracking customers, and it made a huge difference.
And it's not just about buying software: sometimes hiring a person or agency to do something for you can also make sense. John Ross, who runs a test prep company, ditched his accounting software and hired a bookkeeper, and everything ran more smoothly after that. Our customers are having similar conversations. Most build their own automations using our service, which is generally pretty straightforward, but some people hire Zapier experts to build and maintain their Zaps for them.
Sometimes it makes sense not to pay for software you can build yourself. Sometimes it doesn't. Where that line is depends on your business, but it's a line worth thinking about.
