A marketing agency, an in house team, or neither

What each option is actually buying

Strip the pitch away and the three options are buying three different things, only one of which is hours.

• An in house hire buys accumulation. Every month the person knows more about your customers, your product and why the thing that failed last year failed. That knowledge is the asset, and it walks out with them.
• An agency buys range and continuity. Several people with different skills, a process that predates you, and cover when one of them is ill. The cost is that your account is one of several and the knowledge stays with the supplier.
• Neither buys the removal of the work. Some of what you are trying to staff is not judgement at all, it is a rule somebody applies by hand every week, and a rule can be written down once.

Where the crossover actually sits

Two conditions decide this, and neither is the monthly figure.

The first is whether the work has an end. A rebuild of your tracking, a launch, a site, a first working paid media account: these have a shape and a date, and hiring somebody permanently to do a thing that finishes creates the awkward second act where the role has to be invented. The second is whether the work is narrow or wide. One competent hire covers about two of the five activities in a marketing retainer properly. If you need all five and cannot afford three people, the range argument favours an agency and it is not close.

Put those together and the honest rule is that continuous plus narrow favours a hire, continuous plus wide favours an agency, and finite favours a project regardless of who does it. Most businesses discover which one they are by trying the wrong one first.

The cost comparison people get wrong in both directions

A salary is not the cost of an employee. Add the employer's contributions, the tools that one person needs and nobody budgeted for, the recruiting cost amortised across an uncertain tenure, and the management time of whoever the person reports to. That last one is invisible on a spreadsheet and expensive in practice, because a junior marketer with nobody to learn from is a training obligation as much as a resource.

A retainer is not the cost of an agency either. Add the media budget, the tools you still pay for yourself, and the hours your own people spend in review calls, supplying assets and approving things. Engagements that fail on the client side almost always fail here, because nobody costed the internal time and so nobody protected it.

Compare fully loaded against fully loaded or do not compare at all. Half the arguments about which is cheaper are really arguments about which costs were counted, and we have written the equivalent comparison for technical work in an in house team against a partner.

How each one fails

The in house failure is isolation. One marketer, no peer to check the work, no exposure to what is happening in twenty other accounts, and a slow drift towards whatever they were already good at. It is rarely dramatic. It shows up as a year in which nothing was tested that the person had not tested before.

The agency failure is attention and incentive. Your account is one of several, the senior person who sold it is not the person in it, and the arrangement quietly rewards keeping the retainer rather than finishing the work. Neither failure is dishonesty. Both are what the structure produces if nobody watches for it, which is why the questions in how to choose an agency are mostly about structure rather than about talent.

The third answer, which nobody is selling you

Before you decide who does the work, describe the work honestly for a fortnight. Most people are surprised by what comes back.

A large share of what gets framed as a hiring decision turns out to be a list of tasks with fixed rules: pulling the same numbers from four places into one report every Monday, re-typing enquiries from a form into the CRM, sending the same three follow-up messages at the same three intervals, chasing the review request nobody remembered to send. None of that needs judgement. It needs a rule, written once, that keeps applying itself.

That is not an argument against hiring. It is an argument for hiring against the right list. Take the fortnight's tally, cross off everything with a fixed rule and see what is left. If what remains is a full role, hire. If what remains is a few hours of genuine judgement a week, you were about to pay a salary to protect a spreadsheet. The tally is described here in full, and it costs nothing to run.

Doing two of them at once

The arrangement that works most often is not one of the three. It is a small internal owner, an outside supplier for range, and the repeatable half removed entirely so neither party is billing for it.

The internal owner does not need to be a marketer. They need to hold the context, answer questions inside a day and be accountable for the number. An agency with no internal counterpart spends its first two months guessing, which you pay for, and its reports go unread, which you also pay for.

The reason to remove the repeatable half first is that it is the part that makes both of the other options expensive. Reporting, data entry and follow-up are the tasks most often described as somebody's whole afternoon, and they are the easiest to hand to a system that runs in your own accounts. What is left after that is a smaller and much clearer hiring question.

Removing the repeatable half, and what that actually takes

The third answer above, removing the repeatable half so that neither party bills for it, has a timetable rather than a vision behind it. The audit takes the first week and produces a written list of what genuinely repeats and what only looks as though it does. One item is live inside a fortnight, usually the report somebody assembles by hand on a Monday. The second and third arrive across month two. That order matters, because a business that removes the reporting first can see what the other two are worth before committing to either.

The removable half is not the whole job, and pretending otherwise is how this arrangement fails. Anything published in your name, any change to spend and any reply to an unhappy customer wait for human approval. A workflow meeting a case it was not built for hands it to a person with the record attached. The small internal owner in the arrangement above is the person standing on the other side of that stop, which is why the arrangement needs one and why an hour a week is a floor rather than a comfortable estimate.

The published version of removing a repeatable half is Quillon, an Australian training resources company whose delivery line was rebuilt as a single audited automation of 34 AI nodes, moving gross margin from under 35 per cent to 94 across 94 units in two months. That is a services business starting to behave like software, which is the strong form of the argument above rather than an analogy for it. Moiz Khan owns automation architecture at Wobble and Quillon is his own business, which is why that entry carries more detail than any other on the work page.

Common questions

Is it cheaper to hire in house or use a marketing agency?

Neither, reliably, because the two are rarely compared on the same basis. A salary is not the cost of an employee once contributions, tools, recruiting and the manager's time are counted, and a retainer is not the cost of an agency once media, your own tools and your internal review hours are counted. Compare fully loaded against fully loaded, or the comparison is measuring accounting rather than value.

When should I bring marketing in house?

When the work is continuous and narrow enough that one person can cover it properly, and when the knowledge accumulating month by month is worth owning. If you need media buying, creative production, content, measurement and strategy at once and cannot afford three people, the range argument favours an agency and one hire will quietly cover two of the five.

Can I use an agency and have someone in house as well?

That is the arrangement that works most often. A small internal owner holds the context, answers questions within a day and is accountable for the number, while the agency supplies range. An agency with no internal counterpart spends its first two months guessing at things somebody in your building already knows.

What does an agency give me that a single hire cannot?

Range and cover. Several people with different skills, a process that existed before you arrived, and somebody still working when one of them is away. What it cannot give you is accumulation: the knowledge of why the thing that failed last year failed stays with the supplier rather than with you.

What is the third option, if it is neither?

Removing the work. A large share of what gets framed as a hiring decision is a list of tasks with fixed rules, such as assembling the same weekly report, re-typing enquiries into the CRM and sending the same follow-up messages at the same intervals. Those need a rule written once rather than a person applying it every week, and crossing them off first makes the hiring question much clearer.

How big does a business need to be before an in house marketing team makes sense?

Size is the wrong variable. The one that matters is whether there is enough continuous work of a similar kind to fill a role in eighteen months' time, once the obvious backlog has been cleared. Plenty of small businesses have that in one channel, and plenty of larger ones do not have it in five.