Build vs. buy: the honest version
We are the kind of company that benefits when you decide to buy. We sell products and services. So you should probably read this with that in mind. With that said, here is the honest version of the build-vs-buy question, including the cases where we tell clients to build.
Buy if all three are true:
- The category exists and has at least one mature, well-operated vendor.
- Your business doesn't compete on the underlying capability — you compete on the layer above it.
- The cost of the off-the-shelf option, fully loaded, is less than two engineers' time annually.
If those three are true and you still want to build, you're either underestimating the maintenance cost of the build (most common), or you're underestimating the integration cost of the buy (less common but real), or you have an org-internal reason that has nothing to do with technology.
Build if any of these are true:
- The capability is your product. Payments processing for a payments company. Inference infrastructure for a model company. Ad serving for an ad-tech company. Don't outsource the thing you sell.
- The off-the-shelf options are immature, abandoned, or fundamentally misaligned with your operating model. This happens more often in vertical workflows than in horizontal infrastructure.
- You have a long-tail of customizations that would force you to fork or fight the vendor anyway. At some break-even point, "build it yourself" becomes cheaper than "fight a vendor's roadmap forever."
The third option that nobody talks about: hire someone to build it for you, then own it. This is what most of our software-development engagements actually are. The client is in a category where buying isn't viable (immature vendors, vertical specifics) and building isn't viable (small team, other priorities). So they hire us to build a custom system, hand it over, and they own it from day one. It's not technically "build" and it's not technically "buy." It's the option people forget exists.
A few things we tell clients when they're stuck on the question:
- Maintenance cost is roughly 2–4× the build cost. If you wouldn't fund 4× the build budget over five years to keep the thing alive, you shouldn't build it.
- Vendor lock-in is real, but vendor churn is more common. "What happens if this vendor disappears in two years" is a question more startups should ask their three-month-old vendor.
- The team that's building loves to build. Of course they want to build. That's not a vote on the merits.
If you're stuck on a specific build-vs-buy decision, our consultancy team will run it as a one-week scoping engagement. Fixed price, written recommendation, no kickbacks anywhere in our portfolio. We'll tell you to buy from a competitor if that's the right answer. (We do, regularly.)