Rippling founder Parker Conrad on the opportunity in building “compound startups”

Y Combinator CEO Garry Tan points out that the classic startup advice is “do one thing and do it extremely well,” but Parker took a different approach building Rippling that he calls the “compound startup.” He set out to solve multiple problems for their customers simultaneously.

Parker elaborates on this:

“I think it actually resembles a lot of software companies that were built 20+ years ago. If you look at companies like SAP, Oracle, and Microsoft, they actually look like compound software businesses… What happened is I think there was this moment in time where it really was possible to do one narrow thing because there was so much greenfield territory in software as everything shifted from on-prem to the cloud. You could go out and do this really narrow thing and turn it into a SaaS company. Almost all of these things ended up — at least for a period of time — being worth low-single-digit billions of dollars. And it was really easy to find a niche.”

But Parker believes this approach is suboptimal because it limits the types of problems you can solve for your customers — especially big customers.

“You can actually just wipe out a lot of work and make things function much better if you can take on a whole host of interrelated applications and build one comprehensive solution.”

Parker argues that if you can solve multiple problems for your customers with parallel applications, your product/market fit becomes much more powerful and harder to displace.

The other argument in favor of the compound startup approach is that sales and marketing has gotten much harder and less efficient for software businesses:

“You see this at every stage, but over the last five years, public companies are spending 50% more on sales and marketing, but they’re adding 10% less in new ARR with that increase in spend. Fundamentally, the business model for a lot of software is just broken. There’s too many businesses.”

Parker also believes AI sales agents will completely destroy outbound as a channel if they do eventually end up working.

Full video: Y Combinator “How To Build The Future: Parker Conrad“ (Jan 2025)