Decision Library
Build or buy, counted in engineers as well as dollars
Build versus buy is usually argued as license fees against salaries, and both sides undercount. A build carries years of upkeep and the work it displaces. A purchase carries integration, configuration and the price of leaving. A Decision Room compares building, buying, assembling and going without on the whole cost, with the team you have.
More than two options
Build the capability yourself. Buy a product and adopt its way of working. Assemble: buy the core and build the part that makes you different. Hire a contractor to build it and hand it over. Go without for now, using a manual workaround or a simpler export.
The middle options often win, and they are the ones most often left off the table.
Core and context
A useful test is whether the capability is part of how you win customers or simply something that has to work. Building is easier to justify for the first, because control and speed of change are worth paying for there. For the second, such as payroll, ticketing or standard reporting, a mature product carries years of other customers' bug reports that your build would have to rediscover.
Counting the whole cost
Price engineering time at its loaded cost, salary plus benefits and overhead, and include what that team would otherwise ship. Add upkeep, which teams often plan as a share of the build cost each year, and keep that share labeled as your assumption. For a purchase, add implementation, configuration, the integration you must maintain, and price rises at renewal.
Compare over three years at least, since a one-year view flatters whichever option is cheaper to start.
How the room argues ownership
Each path gets a Champion, going without included. The Craftsman checks whether the build estimate matches what similar work took this team before, and whether the people who would maintain it are likely to stay. The Gray Wolf hunts the costs of buying that a quote omits, and the Grandfather prices the exit from each path after three years: rewriting a build, or migrating off a vendor.
The engine works the three-year sums from your own figures. The Verdict names the leading path, the estimate it is most sensitive to, and the figure that would flip it, and the engineering lead records the Human Decision.
What to weigh
Time to value
When a working capability reaches users, counting integration and training.
Whole cost over three years
Loaded engineering time, upkeep, licenses, implementation and renewal increases.
Control
Whether you can change the part that matters most when you need to, without asking anyone.
Lock-in
The cost of leaving a vendor or replacing a build once data and habits depend on it.
Differentiation
Whether the capability helps win customers or simply needs to work reliably.
Where it goes wrong
Build forever
The internal project grows until it resembles the product you declined to buy.
Shelfware
A purchased system is licensed, half configured and quietly abandoned.
Hidden displacement
The build delays the roadmap work customers were actually waiting for.
Reporting: build, buy, or export
A product team needs customer-facing reporting. Building it would take three engineers four months at a loaded $15,000 per engineer-month, or $180,000, plus upkeep the team estimates at $36,000 a year. Buying an embedded reporting product costs $48,000 a year plus $20,000 to integrate. Over three years, building costs $252,000 if upkeep starts in year two, and buying costs $164,000. A third option ships a CSV export in two weeks and waits. With weights of time to value 30, whole cost 25, differentiation 20, control 15 and lock-in 10, buying leads, since customers expect reporting but rarely choose this product for it. The Verdict names what would favor building: reporting cited as the deciding factor in more than a quarter of lost deals.
Questions
- Our engineers say the build takes a few weeks. How is that weighed?
- As an estimate, which the Craftsman compares with how long similar work took the team before. If no history exists, the Verdict carries lower confidence and names the estimate as the figure most likely to change the answer.
- Who makes the final build or buy call?
- Your engineering and product leaders. The Verdict is analysis, and their choice is kept beside it in the Decision Passport.
Bring Argumont a Decision worth challenging.
Start with a paid 60 day pilot on your own real Decisions.