Should I outsource it or build it in-house?
Updated 17 September 2026
This is a complete worked Executive Decision Brief for the build-or-buy decision every small business meets: whether to develop a capability in-house or pay someone who already has it. It shows the board recommending outsource the first version, bring it in on evidence — with the CTO's dissent about control and knowledge kept on the record.
Cerno is a private AI boardroom that runs consequential decisions through a structured process and returns a brief like the one below. This example is illustrative; your own would reflect what the capability is, how often you need it, and how close it sits to what customers pay you for.
Outsource the first version to a supplier, under a defined scope with the handover written into the contract. Bring it in-house only once it's proven continuous, close to the core of what you sell, and you've seen what good looks like. The hire is the second decision.
- ›The capability is needed now for one project, with a plausible but unconfirmed ongoing need.
- ›No one in the business currently has the skill; a hire would take two to three months to find and longer to be productive.
- ›A capable supplier is available at a known price for the first project.
- ›The first project reveals whether the need is continuous.
- ›The supplier will hand over cleanly if asked at the start.
- ›The business can afford a fixed salary later if the need proves real.
Scope the first project tightly and get a fixed price. Put the handover and ownership of all outputs in the contract now.
Watch what it actually takes — hours, skills, tools. That's the job description if you ever hire for it.
Test three things: is the need continuous, is it core, do you now know what good looks like? Three yeses is the trigger to build in-house.
Keep outsourcing, on a retainer if the need is regular. Revisit in six months.
How this brief was reached
The advisors formed independent positions, a Devil's Advocate challenged the leading one, and the vote landed 3–2 for outsourcing first — close, with two dissents that make a coherent case for the opposite answer.
The majority's reasoning was about which door is which. Outsourcing a defined project is a two-way door: it ends when the project ends, and you owe nothing further. Hiring for the capability is much closer to one-way — a salary, a person, an obligation, for a need that today is confirmed for exactly one project. When one option is reversible and the other isn't, and the need is uncertain, the reversible one goes first and produces the evidence the irreversible one requires.
The CTO's dissent is the strongest argument for building now, and it's right if the capability turns out to be core. The brief's answer is that "if" — the whole recommendation is a way of finding out whether the CTO is right before paying as though they are. The COO's dissent about queueing behind a supplier's other clients is real and is why the retainer appears in the actions.
Confidence is moderate. The board is confident that outsourcing first is the safer sequence; it is genuinely uncertain whether the capability will prove core, which is precisely why it declined to decide that now.
What the board weighed
One project is not a need. The business needs this capability for one confirmed project and a plausible pipeline. Hiring against a plausible pipeline is how first hires go wrong — see the first-hire brief. The supplier converts a fixed-cost bet into a variable cost that scales with the need as it actually materialises.
You can't specify a job you haven't watched. The board's second point was about information. Nobody in the business knows what this work involves — how long it takes, what tools, what a good result looks like. A supplier doing it well, once, produces all of that. A hire made without it is a hire made blind.
Core is a fact about the customer, not the workload. The CTO's "if it becomes core" was sharpened by the board into a test: is this close to what customers pay for, or is it something the business needs in order to operate? The first is worth owning. The second — accounts, hosting, the website — is worth buying from someone whose core it is.
Lock-in is a contract term, not a fate. The fear that outsourcing traps you was addressed at the point where it's cheap: the first contract. Ownership of the outputs, the logins, the documentation, and a handover clause — written when both parties are keen, not when one wants to leave.
If your situation differs
If the capability is unambiguously core — it's the thing customers buy — the CTO's dissent is the majority and the question becomes how fast you can hire, not whether. Outsource while you hire, not instead of it.
If the need is occasional and unpredictable, there's no in-house version that makes sense; a supplier on call is the answer indefinitely, and the decision is about which supplier.
If cash is tight and the need is real, a part-time or contract hire sits between the two — the reversible form of building in-house — and the first-hire brief covers how to structure it.
The frameworks behind it
This decision turns on reversibility — outsourcing is the two-way door that produces the evidence the one-way hire needs — and on the base rate for what building capabilities actually costs, which a supplier's first invoice tells you more honestly than your own estimate.
Build or buy? Put your version through the boardroom and get a brief like this with your actual capability and your actual need.
Run this decision in CernoFrequently asked
Doesn't outsourcing mean I never learn to do it myself?
It means you learn what it takes without paying to learn how. A good supplier shows you what the job involves, how long it really takes and what it costs — which is exactly the information you need before deciding whether to build the capability. Outsourcing the first version is the cheapest way to find out whether you want to own it.
What's the test for bringing it in-house?
Three things at once: the work is continuous rather than occasional, it's close to what customers actually pay you for, and you've watched a supplier do it well enough to know what good looks like. Miss any one and in-house is a fixed cost for a job you don't yet understand.
How do I avoid being locked into a supplier?
Own the outputs and the account logins from day one, and write the handover into the first contract rather than the last. Lock-in comes from not having asked, not from outsourcing.