Automation Agency vs In-house Engineer: a Cost Breakdown
What an in-house automation engineer in Europe costs versus an agency engagement, fully loaded numbers, when each makes sense, and the hybrid model.
The agency-versus-in-house-engineer question is rarely answered honestly. Both sides have an incentive to oversimplify, the agency wants the contract, the engineer wants the role. The honest framing is a fixed-versus-variable cost trade with a continuity dimension on top, and the right answer depends on whether automation is core to your product or a cost line on your operations.
The fully loaded cost picture
A senior automation or platform engineer in London, Berlin, or Amsterdam costs €120-160k per year all-in by the time you add benefits, taxes, equipment, and ramp-up time. That is the headline. The less-quoted numbers are the time-to-hire, typically three to six months for a senior, and the failure rate, which the industry quietly admits is around 30% in the first 18 months.
Agency engagements run on a different curve. A single shipped workflow is €5-25k, four weeks. A multi-workflow build is €40-120k, six to ten weeks. Light retainers are €4-9k per month. The agency model trades higher per-unit cost for fixed scope, fast time-to-first-ship, and zero exposure to hiring risk.
When in-house wins clearly
Hire in-house when automation is core to your product, not a cost centre. If your team is shipping automation as the thing customers buy, an internal engineer compounds. They learn the domain, build internal tooling, and the cost is amortised over years. Hire in-house when there is a steady backlog, at least 50% of an engineer's time, year-round, on automation work. Below that, the engineer ends up half-bored and half-blocked, which is a fast track to attrition.
Hire in-house when continuity matters: when the systems being built will be modified weekly, when context-switching across an agency is more expensive than carrying the salary.
When an agency wins clearly
An agency wins when the work is bounded. A defined number of workflows to ship. A re-platforming. An integration build. Anything with an end state that someone can describe in a single document. Agencies are also the right answer when the skills needed are specialised, LLM orchestration, vector search, prompt engineering at scale, and you do not have the volume to justify a full-time hire.
And agencies win when speed matters. A four-week ship versus a four-month hire is the difference between catching a market window and missing it.
The hybrid model most teams should run
The pragmatic answer is often both. An agency builds the first generation of systems, instruments them, and trains an in-house operator to run them. Six to twelve months in, that operator either takes ownership of the systems and the next backlog, or hands a clean codebase to a new engineering hire. Either way, you have avoided a six-month hiring cycle for work that pays back in eight weeks.
The mistake is to imagine the agency-vs-engineer choice as permanent. It is a phase decision, not a forever decision.
Where to read more
The answer page on agency vs in-house covers the same ground in tighter form. For pricing context, the cost-in-Europe answer sets honest ranges.
If you are weighing this decision for your own team, send a short note describing the work and the cadence. We will tell you which side of the line we think you are on, even when the answer is not us.
One workflow, four weeks, measurable lift.
Send a short note about the process you want to automate and the metric you want to move. We respond within one working day with a fit assessment, rough scope, and price range.