Choosing an agency · 12 min read
Agency, employee or freelancer: which one your marketing actually needs
Three ways to get marketing done, with genuinely different cost structures, risk profiles and ceilings. This includes the cases where an agency is the wrong answer — which are more common than agencies tend to mention, and easy to identify in advance.
Written by Zubair Afzal, FounderUpdated
The framing that works
Project, function or capability
Most comparisons of these three options argue about cost. Cost matters, but it is the second question. The first is what kind of thing you actually need, because each model is genuinely good at one of them and mediocre at the others.
A project has a start, an end and a definable output. A site migration. A rebrand rollout. Rebuilding a paid search account from scratch. Projects suit freelancers and contractors, because you are buying a bounded piece of skilled work and you do not need the person afterwards.
A function is ongoing work that requires context you cannot easily transfer: knowing the products, the customers, the internal politics, why the last three things failed. Functions suit employees, because context accumulates and stays.
A capability is expertise you need intermittently, at a level you cannot justify hiring for. Technical SEO on a large site. Digital PR. Analytics implementation. Nobody needs a full-time person for those in a mid-sized business, and hiring one produces an expensive specialist with nothing to do half the year. That gap is what an agency is genuinely for.
Read your own requirement against those three. If it is clearly a project, hiring is expensive and slow. If it is clearly a function, an agency is renting you something you should own. If it is a capability, employing it is uneconomic. Most businesses need some of each, which is why the strongest arrangement is usually a combination rather than a choice.
The three models on the dimensions that actually differ
Cost structure, risk, breadth, depth and ramp time. No figures, because the real numbers vary by market and role more than any average would survive.
| Dimension | Agency | In-house employee | Freelancer |
|---|---|---|---|
| Cost structure | Variable. Scale up or down with notice. No employment liability, no equipment, no tooling cost. | Fixed and fully loaded: salary plus taxes, pension, tools, recruitment, cover and management time. | Variable and usually the lowest per hour, because you are paying for one person with no structure behind them. |
| Breadth of specialisms | Wide. Several disciplines available without hiring for each. | Narrow unless you hire several people. One generalist covering six specialisms covers none of them deeply. | Usually one specialism, sometimes two done well. |
| Depth of context about your business | Moderate. Improves over time but never matches someone in your meetings every day. | Highest. Knows the products, the customers and why the last attempt failed. | Varies. A long-term freelancer can build deep context; a project one will not. |
| Speed of ramp | Weeks. Onboarding is a known process and the skills are already in place. | Months. Recruitment, notice period, then a learning curve before output is trustworthy. | Days to weeks. Usually the fastest start of the three. |
| Bus factor | Moderate. Cover exists in theory; in practice the risk concentrates in whoever holds the account context. | Low if documented, high if not. A single undocumented marketer is the same risk as a single freelancer. | One. If they are ill, busy or gone, everything stops until you replace them. |
| Capacity ceiling | High. More capacity is a commercial conversation, not a hiring round. | Hard. One person has a fixed number of hours and adding another takes months. | Low. Bounded by one person, and they have other clients. |
| Where knowledge ends up | Partly with the supplier. Mitigate by insisting documentation and accounts stay in your ownership. | With you, if you write it down. With the individual, if you do not. | With the individual, almost always. |
| Best suited to | Capabilities needed intermittently, multi-channel programmes, and situations that need senior judgement more than hours. | The ongoing function: strategy ownership, internal coordination, brand and institutional memory. | Bounded projects and single-specialism steady-state work. |
How each one fails
The characteristic failure of each model
- The agency account drifts into activity reporting.
- Work continues, reports arrive, tasks are completed, and nothing commercial changes. Retainers create an incentive to stay busy rather than to finish. The defence is agreeing up front what outcome the retainer is for and reviewing against that, not against a task list.
- The single in-house marketer is asked to be six people.
- One person is expected to run SEO, paid, social, email, content and the website. They do all of it adequately and none of it at a level that competes with a rival who has a specialist on each. This is the most expensive version of in-house, and it usually ends with the person leaving.
- The freelancer disappears at the worst moment.
- Not through bad faith — through illness, a bigger client, or burnout. With no documentation and no shared accounts, work stops entirely. The mitigation is unglamorous: you hold every login, and there is a written handover you could give a replacement tomorrow.
- Nobody internally owns the outcome.
- Suppliers are appointed, work happens, and no single person inside the business is accountable for whether marketing produced anything. This is the failure that makes all three models look like they failed, and it is the only one that is entirely within your control to prevent.
Our position
Where we think an agency is and is not the answer
An agency makes sense when
- You need several specialisms and cannot justify hiring for any of them individually.
- The work needs senior judgement at key moments more than it needs constant hours.
- You need to move faster than a hiring process allows and the requirement may change within a year.
- You have an internal owner who can brief, decide and hold the supplier to an outcome.
Do not hire an agency when
- You need one channel run steadily and nothing else — a specialist contractor will be cheaper and better.
- Your budget is low enough that you would be buying servicing capacity rather than expertise.
- Marketing is inseparable from your product and the context lives with your product team.
- You have already decided to build the function internally, in which case an agency delays the thing you are trying to build.
Questions
Resourcing questions we get asked most
Is an agency more expensive than hiring someone?
Compared on headline numbers, usually yes. Compared on fully loaded cost, often no. An employee costs salary plus employer taxes, pension, equipment, software licences, recruitment fees, holiday and sick cover, management time, and the months of reduced output while they learn your business.
The more useful comparison is cost per unit of specialist capability. One salary buys one person. A retainer of similar size buys fractional access to several specialisms, but less of each and none of them exclusively.
When is a freelancer the right answer?
When the work is a well-defined project, a single specialism, or a steady-state activity that does not need coordinating with anything else. A site migration, a paid search account that just needs competent management, a content programme with a clear brief — all good freelance work.
Freelancers are also the right answer when your budget cannot support an agency at a level where you would get senior attention. A good freelancer at your budget beats a junior account manager at the same price.
When is an agency the wrong choice?
When you need one channel run steadily and nothing else; a specialist contractor is cheaper and usually better. When marketing is inseparable from your product, because the context an agency needs is the same context your product team holds. When your budget is small enough that agency overhead consumes most of it.
It is also wrong when you have decided to build the function internally. An agency in that scenario delays the capability you are trying to build, and the knowledge leaves when the contract does.
What does agency overhead actually buy?
Coverage across specialisms, continuity when an individual is unavailable, senior review of junior work, tooling you would otherwise license yourself, and exposure to problems across many businesses rather than one.
That is a real product, but it is only worth paying for if you actually use it. If your account only ever touches one specialism and one person, you are paying for a structure you are not consuming.
How do I stop the account being handed to a junior?
Ask before signing who will do the work rather than who will manage it, how many accounts that person carries, and what happens when they leave. Get the names in writing, and make senior involvement a stated part of the scope rather than an assumption.
Also ask what proportion of hours are delivery versus account management. A ratio heavily weighted to management is a structural answer to your question, whatever the pitch says.
Can I combine the models?
Yes, and it is often the strongest arrangement. A common pattern is one in-house owner who holds strategy, relationships and institutional knowledge, plus external specialists for the work that needs depth or does not justify a full role.
The failure mode is having no internal owner at all. Without one, external suppliers make decisions in isolation, nobody holds the whole picture, and the knowledge leaves with the contract.
Not sure which of the three your situation calls for?
Describe the requirement and we will tell you honestly which model fits, including when the answer is a freelancer or a hire rather than us. We would rather be useful now than badly matched later.
Last updated