Sooner or later every small business asks it: do we hire an agency, bring marketing in-house, or keep doing it ourselves? The question usually gets framed as a budget decision, and it usually gets answered by whoever is most tired that week.
It is not a budget decision. It is a decision about what you are buying — capacity, expertise, or accountability — and those three come apart more cleanly than most comparisons admit. This guide separates them, and then adds the option that did not used to be on the table.
The question is usually asked wrong
"In-house or agency" is posed as a fork with two branches, and it inherits an assumption nobody states out loud: that doing it yourself means it does not get done. Under that assumption the fork is real. You either buy the hours from an agency or you hire the hours as salary, because one person cannot produce a month of content, a landing page, a video, and a report.
That assumption is the part that changed, and it changed recently enough that most of the advice on this question was written before it did. The comparisons you will find are usually two columns — agency versus in-house — with a cost table underneath. They are not wrong. They are just answering a question from a world where the third column was empty.
You can date the assumption. Ask what a single marketer could produce in a week five years ago and the honest answer is: one good thing, or three mediocre ones. Ask it now and the answer depends entirely on how much of the week is spent producing versus deciding. That is not a claim that the work got easy. It is a claim that the ratio moved, and the two-column comparison was built when the ratio was fixed.
There is a second thing wrong with the framing. "In-house" and "agency" describe who employs the person, which is an org-chart distinction, not an operational one. What actually differs is what you get and who is on the hook for it. Two businesses can make opposite choices here and both be right, because they were buying different things and never said so.
So the useful version of the question is not "which one is cheaper." It is: which of the things an agency sells do you actually need to buy? Some of them have got much cheaper to produce yourself. One of them has not moved at all.
In-house vs agency marketing: what an agency actually sells
Strip the pitch deck away and an agency is selling three distinct things. They are usually bundled, they are priced as one line, and you probably need fewer than three.
1. Capacity
The ability to produce a volume of work on a schedule — posts, pages, creative, campaigns, reports — without your own week absorbing it. This is the largest share of most retainers and the least discussed, because "we will do the work" is not a compelling slide.
This is the part that has changed. Producing a first draft, formatting a deliverable, generating variants, assembling a monthly report — the mechanical half of marketing delivery has got dramatically cheaper for one person to do. Not free, and not judgement-free, but no longer the reason you need to hire five people.
2. Expertise
Knowing which channel to use, what a good ad account looks like, how a launch is sequenced, which of the twenty things you could do this quarter is the one that matters. Pattern recognition from having seen many businesses, not just yours.
This has moved much less. A model can tell you what a media plan usually looks like. It cannot tell you that this particular channel is dead in your particular vertical because everyone tried it eighteen months ago and it stopped working. Specialist expertise remains genuinely purchasable and genuinely valuable — and it is the thing most likely to be worth an agency fee on its own.
3. Accountability
Someone whose job it is to have an answer on the first of the month. Not a person who reports to you, but an organisation with a contract, a named account manager, and something to lose if the work does not happen.
This has not moved at all, and it is chronically undervalued. If marketing is the thing that slides every time something urgent happens, buying accountability is a rational purchase even at a premium. You are not paying for the work; you are paying for the work to be somebody else's problem.
Which are you buying?
Write down which of the three is your actual reason. Most businesses that answer honestly find they wanted one of them and paid for all three — usually accountability, sometimes expertise, rarely capacity in the way they assumed.
What doing it yourself actually costs
The in-house column is normally filled in with a salary and some software. Both are the easy numbers, and neither is the expensive part.
Attention, not money
Marketing done by the founder or the operator competes with everything else that person does. The cost is not the hours — it is that the hours are fragmented, and marketing is one of the few functions that gets materially worse when it is done in twenty-minute slices. A campaign built in fragments looks like a campaign built in fragments.
The demand-generation gap
Here is the cost nobody puts in the table. When you fill your own week with delivery, the work that generates the next quarter's demand is the thing that does not happen — because it is never urgent. Agencies are not immune to this either, but at least their capacity is ring-fenced.
This is the strongest single argument for buying capacity, and it is an argument about your calendar rather than your budget.
The switching cost you pay every week
A solo marketer — or a founder wearing the marketing hat — pays a tax that neither an agency nor an employee pays: every return to the work costs re-entry. Where were we on the campaign, what did we decide about the offer, which version of the copy was current. An agency's account team holds that state between sessions. Working alone, you rebuild it each time, and the rebuild is invisible so it never appears in anyone's estimate.
The practical consequence is that the third option works far better with written state than without. If you go this route, the single highest-return habit is writing decisions down where the next session can find them — not for anyone else's benefit, but so that Tuesday you does not have to reconstruct Monday you.
The knowledge that stays
The one thing in-house genuinely wins on: what you learn stays. Every campaign an agency runs teaches the agency something. Every campaign you run teaches you. Over three years that difference compounds, and it is the reason companies eventually bring marketing in-house even when the maths says otherwise.
Three answers, and when each one is right
Not a ranking. Three genuinely different answers with genuinely different conditions.
| Option | Right when | Wrong when |
|---|---|---|
| Agency | You need specialist expertise you cannot build, external accountability, or genuine multi-channel scale — and your budget clears the threshold where you get senior attention | Your budget is below that threshold. Under it you are buying junior execution supervised by someone with forty other accounts, at a price that assumed you were buying seniority |
| In-house team | The volume is sustained, the knowledge is worth compounding, and you can wait out a hiring cycle | Hiring takes longer than your window. A role filled in four months does not help with a quarter that started last week |
| One person plus AI | The deliverables are describable and repeatable, and the judgement stays with a person who understands the business | You need real strategic depth, external relationships, or someone to be accountable when you are not. Automating delivery does not create judgement, and it definitely does not create accountability |
That third row is where a one person marketing team now lives, and it is worth being precise about what it is and is not. A marketing team of one with good tooling can produce the output of a small agency retainer on the deliverables that are describable — the content, the pages, the reports, the variants. It cannot replace the person who notices that the whole strategy is pointed at the wrong segment. That person is still you, and freeing up your week is only useful if you spend the freed time being them.
A related case worth separating: the one person agency, meaning someone who sells this capability to clients rather than using it on their own business. The economics are different — you are selling the output, so the hours you save become margin rather than free time — but the constraints are identical. The describable half scales; judgement and accountability do not, and clients notice which one they are getting.
The failure mode of the third option is not "the work is bad." It is "the work is fine and nobody is thinking." If you go this route, protect thinking time explicitly, because nothing else will.
How to actually do the arithmetic
Most in-house-versus-agency cost comparisons are wrong in the same way: they compare an agency's monthly fee against a salary, which is comparing a fully-loaded price against a partially-loaded one. A salary is not the cost of an employee, and an agency fee is not the cost of the work.
Here is a comparison that holds up. Three numbers, all of which you can get today.
Step 1: your real hourly cost
Not your salary divided by 2,080. Your target income, grossed up for tax, plus overhead, divided by the hours you can actually bill or actually spend on marketing — which is far fewer than the hours you work. The pricing calculator in our rates guide computes this; it was built for freelancers pricing client work, and the arithmetic is identical here.
Step 2: the agency fee in hours
Divide the agency's monthly fee by that hourly number. The result is the question in its honest form: at my own cost, how many hours a month is this fee worth?
If a retainer converts to more hours than you could realistically spend on marketing yourself, the agency is buying you capacity you do not have — which is a good reason to hire them. If it converts to a handful of hours, you are paying a large premium for expertise and accountability, which may still be correct, but you should know that is what you are buying.
Step 3: what you would actually do with those hours
The step everyone skips. If the answer is "I would not do the marketing anyway, it would just slide again," then the honest comparison is not agency-versus-yourself. It is agency-versus-nothing, and agency wins that one every time.
This is why accountability is worth paying for and why so many businesses that "could do it themselves" should not.
If you choose to do it yourself: the first month
Choosing the third option and then improvising is how it fails. The businesses this works for treat it as a system from day one.
Week one: decide what you are not doing
Pick two channels. Not five. A one-person operation running two channels properly beats one running five badly, and the second version is what everybody does by default. Write down the three you are not doing, so that saying no later is a decision you already made rather than one you have to make again every week.
Week two: build the repeatable pieces once
Identify the deliverables that recur — the weekly post set, the monthly report, the landing page structure, the email format — and build each one properly, once, in a form you regenerate rather than rewrite. This is the entire leverage of the third option. If month two involves rebuilding what you built in month one, you have not chosen the third option, you have just chosen more work.
Week three: set the standard you will be held to
The thing an agency gives you that you must now give yourself is the first of the month. Pick a date, put a recurring report on it, and treat it as an external commitment. A monthly report has a shape — a verdict, performance against a goal, what was done, the data, next month, what you need from others — and writing it forces the review that would otherwise never happen.
If you are doing this for clients rather than for your own business, that same report is the deliverable they are paying for, and it is worth doing properly.
Week four: find where the hours went
Look back at the month and split the time into two buckets: the mechanical half — producing, formatting, assembling, converting — and the judgement half. If the mechanical half is dominant, that is the half worth attacking with tooling. Orkas runs a set of agents locally for exactly that half: writing the document, building the workbook, producing the variants, pulling the research together. It does not decide what matters. That was always the part you were keeping.
Four ways this decision goes wrong
1. Comparing a retainer to a salary
An agency fee includes overhead, tools, management, and margin. A salary includes none of those. The comparable number is fully-loaded cost per useful output, and it is usually closer than the headline figures suggest — sometimes in the agency's favour.
2. Thinking the saving is money
Bringing marketing in-house rarely saves money in year one. It buys control and compounding knowledge, and it costs attention. If the case for going in-house rests on the saving, it is a weak case and it will not survive a busy quarter.
3. Underrating accountability
The most common regret is not "we overpaid the agency." It is "we brought it in-house and then nothing happened for five months." Accountability is the least visible thing an agency sells and the one you notice most when it is gone.
4. Treating "we can do this" as "we should do this"
The third option makes many things possible that were not. It does not make them worth your time. The test is not whether you can produce the thing — it is whether producing it yourself is the best use of the only week you have.
Frequently asked questions
Is in-house marketing cheaper than an agency?
Usually not in the first year, and the comparison people make is normally unfair to the agency. An agency fee is fully loaded — it covers tools, management, overhead, and margin — while a salary covers none of those. Add employer costs, software, recruitment, and the ramp-up period during which output is low, and the gap narrows sharply. In-house wins on control and on knowledge that stays with you, not on price.
When should a small business hire a marketing agency?
When you need specialist expertise you cannot realistically build, when you need genuine multi-channel capacity, or when marketing keeps sliding because nobody is accountable for it. That third reason is the most common real one and the least often stated. The counter-condition is budget: below a certain threshold an agency assigns junior execution to your account, and you pay a senior-looking price for it.
Can one person do all the marketing for a business?
For a small business with two focused channels, increasingly yes — for the describable, repeatable half of the work. A one-person marketing team with good tooling can produce content, pages, reports, and variants at a volume that used to require several people. What one person cannot do is be in two places, hold specialist depth across every channel, or be accountable to themselves when a quarter gets busy. Plan for those three gaps explicitly rather than discovering them.
What does a marketing agency actually do?
Three things bundled into one invoice: capacity (producing the volume of work), expertise (knowing which things are worth doing), and accountability (having an answer on the first of the month). Most clients need one or two of the three and pay for all of them. Working out which one you are actually buying is the fastest way to decide whether the fee makes sense.
How do I compare the cost of an agency to doing it myself?
Convert the fee into hours at your own real hourly cost — target income grossed up for tax, plus overhead, divided by the hours you can genuinely spend on marketing. Then ask whether you would actually spend those hours on marketing if the fee disappeared. If the honest answer is no, the comparison is not agency-versus-yourself; it is agency-versus-nothing.
The two-column comparison was never really about cost. It was about what you are buying and whether you have somewhere else to get it. Capacity has become much easier to produce alone; expertise has not; accountability has not moved at all. Work out which of the three you actually need, price the rest honestly, and the decision usually makes itself — and if it turns out to be the third column, a month of trying it will tell you more than any comparison table, including this one.