Deciding what should stay in-house
Part of a retrospective. The retrospective month groups the topic; it is not an earlier publication date. Published 18 September 2026.
Outsource defined delivery work only after deciding which judgement, information and transition capability the organisation must retain.
The choice between internal delivery and outsourcing is often presented as though the whole service must sit on one side. That can obscure a more useful distinction: who performs the work and who remains capable of directing it.
An organisation can buy specialist operations while retaining service ownership, architecture decisions and commercial judgement. It can also employ engineers internally while relying heavily on a supplier for the knowledge needed to recover a system.
I would assess the actual capability and dependency, not just count employees and contractors. The objective is an arrangement the organisation can operate, challenge and change.
Start with the decisions you need to understand
Identify the business knowledge and technical judgement required to direct the service. Someone must be able to explain what good performance looks like, assess a proposed change and recognise an unacceptable result.
Consider how closely the service is connected to the organisation's distinctive processes. Frequent business changes may require internal understanding that is difficult to specify once in a contract. A more standard service may be easier to buy, provided the requirements and interfaces are clear.
Do not assume internal staff must personally perform every task to retain ownership. They do need enough capability to ask informed questions and evaluate the answers. If only the supplier can define the problem, estimate the solution and declare it complete, the customer has little independent basis for a decision.
Map the retained responsibilities before seeking prices. That helps avoid comparing a fully staffed internal service with a supplier quote that excludes management, integration or business support.
Compare the full arrangements
Include recruitment, development, leave cover and specialist support in the internal option. Include transition, contract management, retained staff and exit costs in the external option.
Check the assumptions about demand. A supplier may offer valuable flexibility for occasional specialist work. A service with frequent changes may incur charges or waiting that the headline price does not reveal. Neither observation establishes a universal preference; examine the proposed terms and actual work pattern.
Use realistic scenarios rather than a precise forecast unsupported by evidence. Ask how the arrangement behaves if demand increases, a key person leaves or an important change arrives at short notice. State where prices are confirmed and where they remain estimates.
Avoid counting redeployed staff as a cash saving unless the financial consequence is real. Releasing their time may be worthwhile, but specify the work they will take on and the skills needed. Outsourcing can fail economically if the organisation pays a supplier while retaining all the old work internally through duplication or unclear boundaries.
Keep sensitivity separate from location
Sensitive information needs appropriate controls regardless of who employs the people handling it. Internal delivery does not automatically make a service secure, and an external service is not automatically unsuitable.
Assess access, information handling, subcontracting and the ability to obtain evidence. Involve security, privacy and legal specialists where the proposed service warrants it. Check the actual arrangement and obligations rather than relying on broad labels about where a supplier is based.
Decide what the supplier needs to access to perform the work and how that access will be controlled and removed. Confirm how incidents are reported and how the organisation can investigate a concern. These details should inform the choice before the contract is signed.
If a requirement cannot be met, treat it as a genuine constraint. Do not assume a commercial deadline changes the underlying obligation. Equally, do not impose an unsupported restriction merely because it makes an internal preference easier to defend.
Retain ownership in a practical scenario
Imagine a fictional business buying managed infrastructure operations. The supplier handles monitoring, routine maintenance and agreed incident response. An internal service owner sets the business requirements and reviews performance, while an internal technical lead assesses architecture changes.
That can be a sensible division if the internal roles have time, information and authority. It becomes fragile if those roles are removed on the assumption that the supplier now owns everything.
The business still needs someone to prioritise competing requests, approve consequential changes and decide whether service limitations are acceptable. Those are customer decisions, even when the supplier provides expert advice.
Write the boundary down with examples. Who authorises an emergency change? Who decides which application is restored first? Who funds work outside the agreed service? Supplier reviews should test whether the boundary works in practice rather than only inspect contractual targets.
Check whether you could change your mind
Before approving the arrangement, walk through a plausible exit. Identify the records, configuration, access and expertise another team would need. Check what can be transferred and the assistance available under the proposed terms.
A right to terminate is not a complete transition plan. The organisation may still need time to recruit, migrate information or establish a replacement service. Understand that work while the supplier has a commercial reason to discuss it constructively.
Keep the plan proportionate. You do not need a second full operations team waiting unused. You do need enough independent knowledge to commission and supervise a transition. The same principle appears in succession planning: a named substitute is useful only when the capability can be demonstrated.
The decision can then be specific. Buy this defined operational service, retain these responsibilities and review these dependencies at an agreed point. That is more useful than declaring the organisation for or against outsourcing in general.
Working on something similar?
If this connects with something you’re working through, I’m happy to talk about where you’re stuck and whether I can help.
See how I work