In-house, agency, or fractional. They’re not three prices for the same thing.

You have a revenue operations problem and three ways to pay for it. Hire someone. Retain an agency. Bring in a fractional operator.

I am the third one, so read the rest with that in your pocket. I'll try to earn it by being straight about when the other two are the right answer — which is more often than anyone selling you one of them tends to admit.

They are not three prices for the same thing

That's the mistake underneath most of these decisions. Someone builds a spreadsheet, puts a day rate next to a salary next to a retainer, and picks the small number.

But you aren't buying hours. You're buying a combination of three things that don't come bundled together: seniority — has this person done it before, or are they about to learn on your CRM; availability — are they there on the Tuesday it breaks; and continuity — will the person who designed it still be around in eighteen months to explain why.

No option gives you all three. That's the whole decision. Everything else is negotiation.

In-house: continuity, at the price of seniority

A full-time hire is the only one who is genuinely yours. They sit in the standup. They hear the sales floor complain, unprompted, which is a data source you cannot buy. They accumulate context, and context is most of what makes an operator good after year one.

The catch is arithmetic. At your budget you'll get one person, and one person can be senior or they can be affordable. The senior RevOps operators who've architected a multi-country CRM from scratch are expensive, scarce, and mostly already employed. So you hire keen and mid-level, and then you've hired someone who will do their learning inside your production system.

Sometimes that's fine. It's genuinely fine when the architecture already exists and the job is running it — where the work is throughput, not design.

The second catch is that you've created a single point of failure with a laptop. Everything they know that isn't written down leaves with them, and almost none of it is written down. We've covered what that costs when it goes.

Hire in-house when: the system is designed and needs running; there's enough steady work to fill a week, every week; and you can either pay senior money or you genuinely have time for someone to grow.

Agency: capacity, at the price of seniority you can name

An agency solves the thing neither of the others solves: volume. A migration with four workstreams, a website rebuild running next to a CRM rebuild, six countries at once. You need parallel hands and someone else's project manager. That's real, and it's worth paying for.

The failure mode is famous enough to be a cliché, which doesn't make it less true. The person on the pitch is not the person on the work. You meet the partner, you sign, and then a delivery team you haven't met inherits a scope document. They are often good. They are almost never the person you were sold.

The other structural problem is incentive. An agency's model wants scope, and scope wants tools. An audit that concludes “turn three of these off and change how you run your Mondays” is a bad quarter for them. It isn't dishonesty. It's what the business model rewards, and business models beat intentions every time.

Hire an agency when: the work is genuinely parallel, you need a delivery machine more than a single brain, and you have someone in-house senior enough to hold them to the outcome rather than the deliverable.

Fractional: seniority, at the price of availability

A fractional operator is the senior person you can't afford full-time, for the fraction of the week you actually need them. The design work — what the data model should be, which stage means what, where the handoffs go — is not forty hours a week. It's a hard eight, and then execution.

So you buy the eight.

The honest cost: they are not there on Tuesday. If your problem is that tickets are piling up and someone needs to work through them, a fractional operator is an expensive way to buy throughput, and you should hire. They also have other clients, which is a genuine trade — it's why they've seen your problem forty times, and it's why they aren't in your standup.

And there's a failure mode nobody mentions: you can end up with an excellent system that only the fractional person understands. That's the same key-person risk as the in-house hire, wearing a nicer invoice. The only defence is documentation and handover, built in from the start rather than promised at the end.

Bring in fractional when: the problem is design rather than volume; you need someone who has done it before and will say “don't”; and the honest answer to “is there forty hours of this a week” is no.

The combination nobody sells you

These aren't exclusive, and the arrangement that works most often is a mix.

Senior design, done once, properly, by someone who has done it before. Then an in-house person — who does not need to be senior, and who will now grow into a system that was built correctly rather than inheriting a mess. Then an agency, occasionally, when there's a burst that needs six pairs of hands.

The reason this isn't pitched to you is that it's nobody's whole deal. The agency wants the retainer. The recruiter wants the placement. The fractional operator would rather stay than document themselves out of a job.

Notice which of those three I am while you read that sentence.

How to actually choose

Ask one question, and be honest about the answer: is your problem design, throughput, or volume?

Design means nobody can tell you what a stage means or where a lead goes. Throughput means everyone knows and there's simply too much of it. Volume means several large things must happen at once.

Design is fractional. Throughput is in-house. Volume is an agency.

Most teams reach for the wrong one because they diagnose by budget instead of by problem. They hire a junior to fix an architecture question, or retain an agency to do work one senior person could have designed in a fortnight, and then conclude that RevOps doesn't work.

If you can't tell which of the three you have — that's a design problem. And now you know what to do about it.

The rest of what people ask before they decide — scope, pricing, how an engagement actually runs — is answered here.