The decision to bring brand in-house or outsource it is rarely made on its merits. It is made on the founder's last bad agency experience, or the head of HR's preference for headcount, or a board member's strong opinion. The right answer depends on the company's stage, the work in front of it, and one or two operational variables most teams don't think to test.
The wrong way to make this decision
Most leadership teams approach this question by listing pros and cons of each model. The list converges on platitudes — "in-house has more context, agencies have more range" — that don't help anyone choose. The decision then defaults to whoever is loudest, or to whichever model the company tried last and found wanting.
The right way is to start with the work the company actually has in front of it for the next twelve to eighteen months. Different brand work has different shapes — some is best done in-house, some is best outsourced, and most companies need a hybrid that is biased toward one model based on stage and rhythm.
The four kinds of brand work
Brand work falls into four operational shapes. The right model depends on which kind dominates the company's near-term need.
- Strategic articulation — positioning, narrative, principles, brand strategy. Episodic. Done deeply, occasionally. Best outsourced because it benefits from outside perspective and is hard to do well from inside the company you are trying to articulate.
- Identity creation — building a coherent visual and verbal system from the strategy. Project-shaped, with a defined start and end. Best outsourced for the build, in-housed for the maintenance.
- System maintenance — keeping the identity coherent at scale, governing exceptions, updating components, training new contributors. Continuous. Best in-housed.
- Application work — the daily output of decks, emails, ads, social posts, packaging, customer-facing artefacts. High-volume. Best in-housed where the volume justifies it; outsourced where it doesn't.
Most companies have all four kinds at different intensities. The model should follow the dominant work, not the org chart aspiration.
The stage test
Pre-Series A
Almost certainly outsource everything. The volume of work doesn't justify a permanent hire, the founder is still close enough to do the maintenance themselves, and the strategic work needs the outside perspective most. Hiring an in-house brand lead at this stage usually buys a single person doing four jobs badly.
Series A to ~£5m revenue
Hybrid. Outsource the strategic and identity work. Hire one strong brand operator — usually a senior designer or a brand manager with strategic literacy — to own the system maintenance and the application work. The in-house person becomes the agency's day-to-day partner and the company's institutional memory.
~£5m to ~£20m revenue
Build out the in-house team for application and system maintenance. Keep the strategic and identity work outsourced. The internal team grows to handle the daily volume; the outside partner stays on retainer for the work the in-house team cannot or should not do alone.
£20m+ revenue
Mostly in-house, with selective outsourcing for specific moments — a major repositioning, a new business unit launch, a category-defining piece of work. The agency relationship shifts from supplier to specialist consultant for the highest-leverage work.
Two operational variables most teams miss
Beyond stage, two variables disproportionately determine the right answer.
The cadence of application work. If you are producing more than ~10 substantial customer-facing artefacts per week (emails, decks, social, landing pages), in-house starts to pay back hard. Below that, outsourced is cheaper and more flexible.
The depth of internal alignment. If the senior team is genuinely aligned on the brand and able to articulate it clearly, in-house works well — the team has a clear thing to defend. If alignment is weak or contested, in-house brand work absorbs the disagreement and gets stuck. Outsourced work is sometimes better at this stage because the agency can hold a position the internal politics cannot yet.
What good looks like in practice
The companies that do this well share a few patterns. They name a single senior brand owner regardless of model — internal lead or external partner — and hold them accountable for coherence over time. They review the model annually rather than quarterly, because brand work has a slow rhythm. And they resist the false economy of "in-house is cheaper" — a single mediocre in-house brand lead costs more than a strong agency, accounting for salary, on-costs, recruitment risk, and the work that doesn't get done because one person cannot cover four kinds of work.
Common pitfalls
Three patterns recur. The first is premature in-housing — companies hire too early, pay for a permanent role they don't yet have the work for, and then either under-utilise the hire or stretch them across work they cannot do well. The second is permanent outsourcing — companies past £10m revenue still using an agency for daily application work, paying agency rates for what should be in-house production. The third is the unmanaged hybrid — both an in-house team and an agency, with no clear decision rights between them; the work either duplicates or falls between the two.
What This Looks Like in Practice
The patterns we describe here apply across our work — from Antidote Africa, where the strategic and identity work was best done with us as the outside partner because the proposition itself needed articulating, to mature brands like BGR, where the long-running relationship is calibrated to the work that genuinely benefits from outside perspective rather than to volume. The model is not a fixed answer; it is one we revisit deliberately as the company changes shape.
Closing
The in-house vs outsource question has no universal right answer — but it has a wrong answer for almost every company at almost every stage. The right answer is to identify the dominant work, calibrate to the stage, and review annually. The wrong answer is to default to whichever model is currently fashionable or whichever the founder last had a bad experience with.
If you are weighing this decision and would value an honest outside perspective on which model best fits your stage and work, we are happy to talk.