Hire or outsource? Test the work before choosing the structure.

Why hire-or-outsource decisions go wrong

Most hire-or-outsource debates are really debates about certainty. A hire feels safe because the person is “ours”. Outsourcing feels safe because you can leave. Both feelings describe a real risk, but neither is evidence. The costs also differ in kind: a hire commits salary, management attention and a slow, sometimes legally regulated exit; an agency or contractor commits less per month but leaves knowledge, quality control and priorities partly outside your control.

The usual mistake is to choose the structure first and then argue for it. Choose instead from three properties of the work: how long the need will last, how much context the work requires, and how easily you can specify and check the output.

• Leans toward hiring: the work shapes your product or customer relationships, demand has been steady for many months, and quality depends on context that takes weeks to learn.
• Leans toward outsourcing: the scope is bounded and can be written down, demand is seasonal or uncertain, the skill is specialised and needed part-time, and the output is easy to inspect.
• Leans toward a hybrid trial: the signals disagree. This is the most common situation, and the reason a small test is worth running.

How to answer the three questions honestly

1. Which part of the work must the company own?

List the tasks in the workstream. For each one, ask whether losing the person who does it would cost you customer trust, product understanding or negotiating position. Only those tasks argue for an internal owner. Often only a slice of the work has to stay inside; the rest is execution that can be specified.

2. What evidence shows the workload will last?

Use invoices, tickets, tracked hours and pipeline, not forecasts. Look at the last twelve months month by month and separate steady demand from peaks. If you cannot name the customers or contracts behind the next twelve months, treat demand as unproven, however busy it feels today.

3. What becomes hard to reverse?

For a hire: notice periods and legal obligations (which vary by country), onboarding time, recruiter fees and the effect on the team if the role ends. For outsourcing: knowledge that stays with the supplier, switching costs, minimum contract terms and dependency. Write each item as “cost and time to undo”. A tidy list often changes the decision by itself.

“I do not know yet” is a valid answer. It tells you which measure the trial must produce.

Running the four-week trial, step by step

1. Pick one real workstream with a visible outcome, not a side task nobody will miss.
2. Name one internal owner who is accountable for the result, and one external specialist with a written scope: deliverables, hours and price.
3. Agree the measures before you start: cycle time, quality (rework or reviewer score), hours spent coordinating, and knowledge captured in documents or playbooks.
4. Set two checkpoints, at the end of week two and week four, with a written stop rule.
5. At the end, compare with the previous four weeks as a baseline and write a one-page decision: hire, keep the hybrid, outsource fully, or stop.

A good trial costs a fraction of the first quarter of a hire and can be finished without embarrassment in either direction.

Common mistakes

• Testing a task nobody cares about, so the result proves nothing.
• Letting the contractor do everything so the internal owner learns nothing.
• Comparing cost per hour instead of cost per outcome, and forgetting coordination time and rework.
• Skipping the baseline, so “better” is a feeling rather than a comparison.
• Deciding on the last day, in the mood of the last day.
• Using a contractor arrangement that a local authority could treat as employment. Check the rules where you operate.

Second example: support for a growing product

A founder answers customer questions personally and wants to hire a full-time support lead. The three questions show that product knowledge must stay inside, but ticket volume has doubled only in the two weeks after each release. The founder keeps one internal owner for escalations and books a part-time contractor for the post-release peaks for four weeks, measuring response time, resolved-on-first-reply rate and hours spent briefing the contractor. Only then does the full-time role get a job description.