The build versus buy question for content comes down to one choice: hire someone in-house, or buy content engineering services. Most teams answer it by comparing a salary to a retainer and picking the smaller number. That math ignores the real difference, which is what you keep when the money stops.
Most teams still think they are hiring a writer. They are not. The role has shifted into an engineering function, and that change rewrites every number in the comparison.
Price the system, not the seat.
TL;DR
- The decision: hire a permanent operator, or buy a system that outlives the engagement.
- The cost: one in-house content engineer runs $120K to $220K a year before tools and benefits.
- The market: content marketing manager listings are down 73% since 2023, so the talent you want is rare and expensive.
- The rule: hire when you need an embedded operator, buy services when you need the system installed and running.
The role changed, so the math changed
Content roles used to be about producing pages. You hired a writer, gave them a keyword list, and counted the posts that shipped. The job description now reads differently. Ahrefs analyzed 20 US content engineer job descriptions and found 85% require building an AI-augmented content pipeline.
Another 70% require owning SEO, AEO, and GEO. And 65% require prompt engineering. Pipeline building is more universal to the role than writing itself. So the $120K to $220K hire you are weighing is not a content producer. You are paying a systems builder who also writes.
The shift is not cosmetic. Google’s own AI optimization guide, published in May 2026, draws a hard line between commodity content and everything else. The pages that win AI search are the ones a machine built on purpose. That is engineer work, not writing work.
The production versus systems split is the whole ballgame. A writer produces articles. A content engineer produces the system that produces articles. Hire the first and you get output that stops when the person stops. Buy or build the second and you get output that compounds. The job descriptions above are not asking for a faster writer. They are asking for someone who installs the machine.
That is the first number most teams get wrong.
You are not hiring a writer anymore. You are hiring someone to build and run a machine, and that is a different person at a different price.
The hiring market is not on your side
Even if you decide to build in-house, the market is working against you. LinkedIn’s labor data shows global hiring down 30% from pre-pandemic levels, while applications per applicant are up 30%. More candidates are chasing fewer roles, which means slower fills and longer gaps on your team.
The specific role is getting scarcer, not just busier. Averi.ai reports content marketing manager listings are down 73% since 2023. The writer and producer titles are contracting while the systems title grows. That means the few people who can do the job can name their price.
A salary is not the full cost. Add benefits, payroll tax, and the tool stack, and the loaded number lands well above the posted band. A $180K base salary is closer to $240K all-in before that hire ships a single page.
A bad hire costs more than the salary. You lose the recruiting months, the ramp months, and the pipeline that should have been filling while you waited. By the time you realize the hire cannot build the system, you have paid for a year and you have pages, not a machine.
Building in-house right now is a months-long project before the first draft ships.
| Consideration | In-house hire | Content engineering services |
|---|---|---|
| Cost | $120K to $220K a year plus tools | Fixed build, then optional retainer |
| Time to value | Months to recruit and ramp | Weeks to first output |
| What you keep | One seat and its knowledge | A system your team runs |
| Main risk | They leave, the system leaves | You still need an operator |
What the service actually delivers
A content engineering service is not a writing retainer with a fancier name. The deliverable is a running system: the tools, prompts, roles, and gates that produce and improve content on a schedule. You are buying the machine, not the articles.
That install happens in a defined sequence.
A short map of what already earns traffic and citations, plus what is dead weight. It names your best queries and the gaps a system should fill.
The tools, prompts, and roles that will produce the content. This is where the service builds the pipeline instead of just staffing it.
The checks that stop generic or off-brand drafts from shipping. A system without gates is a content firehose, not an engine.
The loop that tells you which queries are moving and which need a rewrite. You measure position and pipeline, not drafts shipped.
A good service hands that system to your team and teaches one operator to run it. The engagement has a finish line. What you keep is the machine, and it keeps producing after the vendor leaves. An in-house hire keeps producing only while they stay, and they take the system with them when they go.
An agency sells output. A service sells the system that produces output, then teaches you to run it. That is why a service has a finish line and an agency has a recurring invoice. Both are legitimate, but only one leaves you with an asset.
When in-house still wins
This is not a one-sided argument. There are three cases where hiring wins, and you should be honest about which one you are in.
First, when you already have a proven operator. If someone on your team runs a content system today, promoting them is cheaper than teaching a vendor your business. Second, when the work is sensitive enough that an outside tool cannot touch your data. Third, when you need a permanent internal voice rather than a system that hands off to whoever runs it.
Most teams are in none of those three. They want output now, a system that survives turnover, and no headcount fight with finance. That is the buy signal.
If that sounds like you, stop pricing the hire and start pricing the install.
The decision is rarely permanent. Plenty of companies buy the build first, then hire an operator to run it once the system is documented and the volume is proven. That path is cheaper than hiring first and hoping. You are not choosing a side for life. You are choosing what comes first.
What I actually run
I can write this because I run the system every day. Three properties, one operator, and a nightly loop that produces, evaluates, and ships. The kokasexton.com library sits at 253 posts as of November 2026. The LinkedIn algorithm study behind my numbers covered 2,279 posts from 246 creators.
So when I compare a hire to a service, I am comparing a system that already runs against a promise that someone can build one. A hire starts from zero. A service starts from the machine I already operate every night. That is the honest difference, and it is why the service usually wins on time to value.
I set my own floor for this role at $180K, and I still lean toward the service. Not because the person is not worth it. Because a single hire is a point of failure, and a documented system is not. One person gets sick, goes on leave, or takes another offer. The system does not.
My bias is to hand you a system you can run without me. A service that needs you dependent forever is renting you a seat, not building you an engine. The build is the product. The retainer is optional.
The reason it works is the loop, not the effort. The system decides what to write, writes it, gates it, and ships it while I sleep. A hire reproduces my output. A service reproduces the loop.
I wrote the longer version of the role split in the AI content engineer post. I also broke down what AirOps sells versus what I actually run.
If you want the system installed inside your company rather than quoted as a proposal, start with the done-for-you content engine. It is a four-week build with weekly deliverables.
Hire when you need an embedded operator and already have one. Buy content engineering services when you need the system installed, running, and documented. Most teams are in the second camp, and the math says stop paying for content twice.
Sources: Ahrefs, What Is Content Engineering (85% of 20 US job descriptions require an AI pipeline, published May 19, 2026), Averi.ai content engineer role analysis ($120K to $220K salary band and a 73% drop in content marketing manager listings, 2026), LinkedIn Hiring Assistant 2 (global hiring down 30%, applications up 30%, published September 2026), and Google AI optimization guide (May 2026). Figures are vendor-reported and directional.
