If you run technology for a university department, a non-profit, or a mid-market company, you've probably had this debate more than once. Do we build this with our own people, or bring in an outside team? It almost always gets framed as a cost question, and that's the first place good teams go wrong. The reassuring part is that this exact decision has been studied for decades, and the findings tend to match what experience teaches the slow way.
Cost is the wrong first question
An in-house hire and an outside engagement rarely cost the same thing for the same reason, so lining up their hourly rates tells you very little. A full-time employee is a fixed, long-term commitment that gets cheap per hour once they're up to speed, though they're slow to hire, hard to replace, and often quiet between projects. A partner costs more per hour but flexes with the work: you pay for a burst of senior capability and stop when it's done.
Economists have a name for what you're weighing, and it isn't price. The field is transaction cost economics, and its founder, Oliver Williamson, earned a Nobel for a plain idea: the real cost of any make-or-buy decision includes the oversight and coordination around the work, not just the invoice. A 2006 meta-analysis gathered transaction-cost studies covering tens of thousands of real decisions and found the framework holds up well for choosing between building in-house, buying it in, or partnering.1 So the urge to compare rates is understandable, and it's still the wrong starting line. The better questions are about time, risk, and what happens after launch.
Where each option genuinely wins
Neither choice is better in the abstract. They suit different situations, and the research is fairly specific about which.
Williamson's work points to one factor above the others: how specialized the work is to you. When a job leans on deep, particular knowledge of your systems and it comes up again and again, keeping it inside is usually cheaper and safer. When the work is more standard, the open market handles it well. Two well-known management papers add the strategic version of the same point. Prahalad and Hamel argued that a company should hold on to its core competencies, the handful of things it has to be excellent at, and find good outside partners for the rest.2 Quinn and Hilmer made that practical for sourcing: concentrate your own people where you can be genuinely distinctive, and bring in specialists everywhere else.3
Lean toward in-house when…
- The work is continuous and central to your mission, not a one-time push
- Deep, daily knowledge of your systems matters more than raw speed
- You can realistically hire and keep the specific skills you need
- The work is steady enough to keep a full-time person genuinely busy
Lean toward a partner when…
- You need a specific outcome on a deadline, not a permanent role
- The project needs skills you don't have and can't justify hiring for
- You want senior strategy, design, engineering, and AI without four separate hires
- Speed and a fresh outside read matter for this particular problem
The abandonment risk is the one to take seriously
Ask a lean IT team why they hesitate to bring in outside help, and you'll usually hear a version of the same worry. Someone parachutes in, builds something clever, then leaves, and your team is stuck maintaining a system it never fully understood. That fear is well founded, and there's good data on how these arrangements actually end.
Researchers call the reversal backsourcing: pulling work back in-house after it was outsourced. It happens a lot. One often-cited study found that 34% of discontinued outsourcing contracts get brought back inside.4 A 2006 study of 160 IT managers looked closer, comparing firms that stayed with their vendor, firms that switched to a new one, and firms that backsourced.5 What separated the groups had almost nothing to do with price.
| Outcome | Product quality | Service quality | Relationship quality | Switching costs |
|---|---|---|---|---|
| Continued the relationship | 5.45 | 5.26 | 5.32 | 4.45 |
| Switched vendors | 4.86 | 5.01 | 4.12 | 3.46 |
| Backsourced (brought in-house) | 3.92 | 4.21 | 4.42 | 3.15 |
Two things jump out. The firms that brought the work back in-house rated product and service quality lowest, so backsourcing reads as a response to work that simply wasn't good enough. The firms that switched vendors told a different story: the quality was fine, but the relationship and the ties holding them to that vendor had worn thin. In neither case was the hourly rate the deciding factor. Quality and the working relationship were.
That is the risk worth planning around, and it runs both ways. A cheap engagement that leaves you with something you can't operate was never actually cheap. Neither is an expensive one that never passes along what your team needs to know.
The middle path most teams miss
The choice was never really binary. For a lean team, the best outcome is often a partner who works with your people instead of around them: one who brings the senior capability you're missing, ships the thing you need, and then hands over enough knowledge that your team can own it and build on it.
There's a reason this matters beyond good manners. A long-standing idea in management research, absorptive capacity, holds that an organization can only take in outside knowledge to the extent it already has some footing of its own to receive it.6 In the same spirit, Feeny and Willcocks studied which capabilities a company should keep in-house even when it outsources heavily, and placed the informed buyer near the top: enough retained understanding to direct the work and judge it honestly.7 Keeping a little capability inside is exactly what lets you rely on help outside.
That's deliberately how we work at Designmine. We document as we go, train your team, and treat "your people can run this without us" as part of what we deliver. It's also why we'll say so plainly when a problem is better solved by a hire. If that's the honest answer, you deserve to hear it from us.
A simple way to decide
Strip the research down and the same three questions keep surfacing. They line up almost exactly with what the academics measure. Is this work continuous or bounded? Continuous, specialized work leans in-house; bounded work leans toward a partner. Do you have, or can you realistically hire, the exact skills it needs? If not, a good partner is faster and less risky than a long search. And once it ships, who needs to own it? If the answer is your team, make sure whoever builds it leaves them able to run it.
Answer those honestly and the right call usually gets clear, whichever hourly rate looked cheaper at the start. If you'd like a second opinion on a decision you're weighing, we're glad to give you a straight one.