In-house team or agency for hospital digital: choosing the split
In-house team or agency is rarely a clean binary in a hospital group. Own the thinking, the clinician relationships and anything touching patient data. Own the daily machinery of enquiries, CRM and the website. Rent media execution, production and specialist builds where demand is uneven. Split by continuous against episodic work, keep the internal team small and senior, and fix accountability before contracts.
Every growth leader in a hospital group inherits a version of this argument. Marketing wants more people on payroll. Finance points at the agency retainers. Someone from the board asks why we pay an outside firm to write social posts. Someone else asks why we employ a team that briefs an agency to do the work anyway.
In-house team or agency is put as a binary, and it almost never is one. The useful question is where the line sits between them, what each side is accountable for, and which parts of the work should never have left the building in the first place. Getting that split right matters more than the headcount number that usually dominates the discussion.
I have run both sides. I have built internal teams that made agencies redundant in some areas and useless in others, and I have watched internal teams turn into brief writers who added a fortnight to every piece of work. What follows is how I would draw the line now.
Start with what only you can do
Some work cannot be outsourced because it depends on access, not skill. Knowing which consultant will actually return a call on a Saturday. Knowing that the new unit is short of beds in one specialty and desperate for volume in another. Knowing which unit head will block a campaign because of a conversation last year. Knowing the real capacity position before a campaign goes live, which I have argued is the constraint most marketing plans ignore in capacity is a marketing constraint.
That knowledge is the core of the internal job. Strategy, prioritisation, the relationship with clinicians and unit heads, the argument with finance, the decision about what not to do. None of it travels to an outside firm, however good they are, because it lives in corridors and in trust built over time.
The second thing that belongs inside is anything that touches patient data or consent. Enquiry handling, the CRM, the contact centre, follow-up, anything that reads or writes a patient record. Not because agencies cannot be trusted, but because accountability for that data sits with the hospital and should not be one step removed.
What agencies are genuinely better at
It is easy to become cynical about agencies, and easy to forget what they solve. They solve peaks. A launch, a festival campaign, a new service line, a sudden need for video in three languages. An internal team sized for peaks is idle most of the year, and an internal team sized for the average will miss every launch.
They also solve craft depth. A good performance media specialist, a good film director, a good designer, a good production crew: these people have chosen careers where they work across many clients and many problems. A hospital group can rarely offer them the variety that keeps them sharp, and the ones who do join often stagnate. The same goes for technical specialisms that a hospital needs a few times a year.
There is a less comfortable point in their favour. An outside firm can say things to a hospital’s leadership that an employee cannot, or at least can say them without risking a career. I have used that deliberately: a finding that lands badly from the marketing head lands neutrally from someone with no stake in the politics. It is not a reason to outsource judgement, but it is a reason not to treat external partners purely as a pair of hands.
And they solve the outside view. An internal team stops seeing its own hospital after a while. Everyone knows the doctor, so nobody notices that the doctor page makes no sense to a stranger. An agency reading the website cold will tell you things your team stopped being able to see.
In-house team or agency: where the line actually belongs
My default split has three layers, and it has held up across different groups.
Own the thinking. Strategy, the plan, the priorities by specialty and unit, the budget argument, the measurement definitions, the relationships. This is the internal team’s job and it should be senior, small and close to the business.
Own the machinery. Anything with a patient in it and anything that runs every day: the CRM, the enquiry desk, the contact centre, the website as a booking product, the review and listing estate, data and reporting. These are operational assets, they degrade without daily attention, and an agency contract is a poor way to hold that attention.
Rent the making. Media buying execution, film and photography, design production, larger content pushes, specialist technical builds. This is where external capacity earns its money because volume is uneven and craft is deep.
The failure mode to watch is the internal team that owns none of these and instead owns the briefing. A team whose entire function is to translate between the business and an agency adds delay and no judgement. If your organisation chart looks like that, the problem is not the agency.
The multi-unit complication
In a group with several hospitals, this stops being one decision. Units will have their own local arrangements, sometimes with a firm a unit head has known for years, and those arrangements are rarely visible centrally until someone tries to count them.
The instinct is to consolidate everything immediately. I would resist that in the first pass. Local knowledge in a catchment has value, especially for regional language work and community activity, and a unit head who has a partner they trust will cooperate more than one who has had a central roster imposed on them. What must be central is the standard: the brand, the claims rules, the data handling, the measurement definitions and the approval of anything that mentions clinical outcomes. How I have handled the practical side of that is in building an agency roster across hospital units.
Then consolidate where scale genuinely pays. Media buying across units is the clearest case, because fragmentation costs money and makes attribution impossible. Film and photography is next, because consistency shows. Local activation is usually the last thing worth centralising, and sometimes never.
What it costs to hold capability inside
The honest comparison is not salary against retainer. Holding capability inside means recruitment, management time, training, the risk of a single person leaving with a system in their head, and the slow drift that happens when a specialist has nobody to learn from. It also means you carry the cost when demand drops.
Against that, an internal team’s knowledge compounds and stays. An agency’s knowledge of your group walks out with the account executive who moves on, and you brief the replacement from the beginning. That handover tax is real and nobody puts it in the comparison.
The question I would actually ask is whether the work is continuous or episodic. Continuous work belongs inside, because the cost of re-explaining context every time exceeds any saving. Episodic work belongs outside, because idle internal capacity is the most expensive thing on the list. Most disputes about this are really disagreements about whether a given activity is continuous or episodic, and that is a much easier argument to settle.
Hiring the internal side in the right order
If you are building the internal team, the order matters more than the total. I would hire a senior person who can hold the relationships and say no before I hired any specialist. Then whoever owns the enquiry to appointment machinery, because that is where money leaks daily. Then content and brand, because that is continuous and requires clinical judgement about wording. Specialists last, and only where the work is genuinely continuous.
The senior hire is the one most groups get wrong, usually by recruiting a channel expert into a leadership seat. I have set out what I look for in hiring a head of digital for a hospital, and the short version is that the job is mostly persuasion inside a clinical organisation, not media skill.
What I would avoid is hiring a junior team quickly to justify cutting agency spend. You get a team that cannot do the work, an agency that is now unpaid but still needed, and a growth leader spending their days on rework. The team should grow behind demonstrated demand, not ahead of it. The wider version of this argument is in building a digital team inside a hospital group.
Judging whether the split is working
Once a split is in place, most groups never test it. They renew contracts, argue about headcount at budget time, and never ask the plainer question of whether the arrangement produces better work than the alternative would.
The signals I watch are behavioural rather than financial. How long it takes to get something out of the door when a unit head asks for it on a Tuesday. Whether consultants will talk to the people making content about them, which they will not do with a stranger who reappears every few months. Whether anyone can explain last month’s enquiry movement without a deck being prepared. Whether the same mistake gets made twice, which is the surest sign that learning is not being kept anywhere.
I also look at what happens in a crisis or a launch weekend. Those are the moments that reveal who is actually available and who is inside the organisation’s confidence. An arrangement that works beautifully in ordinary weeks and disappears in the difficult ones is not the arrangement you thought you had bought.
Put those signals in front of the executive committee once a year with the budget, not instead of it. The point is not to relitigate every contract annually. It is to keep the split as a conscious choice rather than a habit, because the business changes underneath it. A group that opens two units, adds a service line or moves into a new city has changed which work is continuous, and the split should follow that rather than the other way round.
How I would choose
If I took over a hospital group’s digital and growth function tomorrow, I would own strategy, the enquiry machinery, the website, the data and the clinician relationships with a small senior internal team, and I would rent media execution, production and specialist builds. I would keep one strong external partner rather than a long roster, and I would keep the number of internal people lower than most marketing leaders would like.
The reason is simple. A small internal team that decides and a good external team that makes will beat a large internal team that does everything adequately, and it will beat an outsourced function that has no judgement behind it. The senior internal seats are worth paying for. The production seats usually are not.
The one situation where I would go further inside is a group running continuous, high volume content and campaign work across many units and languages. At that scale the handover tax on external production becomes the dominant cost and an internal studio starts to make sense. Below that scale it is a vanity build.
And the one situation where I would go further outside is a group with no senior internal leader in place. Do not build a team under a vacancy. Hire the leader, let them choose, and accept a year of heavier external dependence in the meantime.
What to change in the next quarter
List every external arrangement across the group, including the ones units pay for locally, with what each one does and who signs off the work. Most groups have never had this list in one place, and the duplication in it usually pays for the first internal hire.
Then mark each activity continuous or episodic. Anything continuous that currently sits outside is a candidate to bring in. Anything episodic sitting inside is a candidate to release. Do this with the marketing team in the room rather than to them, because the conversation is about their jobs and pretending otherwise damages trust you will need later.
Finally, fix the accountability before the contracts. Write down who owns each outcome, not each activity: who is answerable for enquiries, for booked appointments, for brand consistency, for the group’s visibility in search. An outside firm can be responsible for work. Only someone inside can be accountable for a result. Once that is clear, and only then, decide what you keep and what you buy. Vendors are easier to manage when the outcome already has an owner, which is the point I keep returning to in managing vendors without becoming their project manager.
Questions people ask
It is the split of hospital digital and growth work between salaried staff and outside firms. In practice it covers three separable layers: the thinking, meaning strategy, priorities and relationships; the machinery, meaning enquiries, the CRM, the website and reporting; and the making, meaning media execution, film, design production and specialist builds. Most disputes happen because these layers get discussed as one decision.
Anything that depends on access rather than skill, and anything touching patient data. Knowing which consultant responds, which unit needs volume in which specialty and what the real capacity position is cannot be briefed out. Enquiry handling, the CRM, follow-up and anything reading or writing a patient record belong inside because accountability for that data sits with the hospital and should not be one step removed.
Peaks, craft depth and the outside view. Launches, festival campaigns and multilingual video create demand no internal team can be sized for without being idle the rest of the year. Specialists in media buying, film and design stay sharp by working across many problems, which a single hospital group cannot offer. And an outside team reading your website cold notices what your own people stopped seeing.
Look at what the internal team actually produces. If its main output is briefs and approvals, you are paying for translation rather than judgement, and the delay it adds is real. Look also at who is answerable when enquiries fall. If the honest answer names an outside firm, the accountability has drifted outside the organisation and no contract will pull it back.
The costs nobody lists. Recruitment, management time, training and the risk of one person leaving with a system in their head on the internal side. On the external side, the handover tax each time an account team changes and your group has to be explained again. Ask instead whether each activity is continuous or episodic, because that distinction predicts the true cost better than either rate card.
Do not consolidate everything on the first pass. Local arrangements in a catchment carry knowledge, especially for regional language work, and a unit head with a trusted partner cooperates more than one handed a central roster. Centralise the standard instead: brand, claims rules, data handling, measurement definitions and approval of anything clinical. Then consolidate media buying, where fragmentation clearly costs money.
Seniority first. A leader who can hold clinician and unit relationships and say no comes before any specialist. Next, whoever owns the enquiry to appointment machinery, because that is where money leaks daily. Then content and brand, which is continuous and needs judgement about clinical wording. Specialists last, and only where the work is genuinely continuous rather than occasional.
Not quickly, and not junior. Hiring a cheap team to justify ending a retainer usually produces a team that cannot yet do the work, an agency that is still needed, and a growth leader spending their days on rework. Let the team grow behind demonstrated demand. The saving is real when capability is genuinely replaced, not when the invoice simply moves onto payroll.
A hospital is not the obvious employer for digital specialists, so the offer has to be about scope and learning rather than only pay. Plan for retention of the few people who hold critical systems, documented handovers so knowledge does not walk out, and career paths that do not require leaving to progress. Also plan honestly for the conversations that follow when work moves inside or outside.
Any outside firm touching patient or enquiry data needs a written purpose, a lawful basis, access limits, retention terms and a clear end of contract position. Campaign reporting that includes identifiable information should be restricted, and approval of clinical claims must stay with the hospital. These belong in the agreement at signing, not in a review after something has already gone out.
It should. Work moves inside as it becomes continuous and moves outside as it becomes occasional. A sensible rhythm is a review each year against the same question rather than a restructure each time a budget is squeezed. What should stay stable is the ownership of outcomes, because moving that around every year is what makes teams and partners defensive.
A few weeks of honest listing. Record every external arrangement across the group, including those units pay for locally, with what each does and who signs off the work. Then mark every activity continuous or episodic. Most groups have never had that list in one place, and the duplication it reveals often pays for the first internal hire on its own.
Deciding the structure before deciding accountability. Once someone inside is clearly answerable for enquiries, booked appointments, brand consistency and search visibility, the question of what to keep and what to buy becomes straightforward. Decided the other way round, the group ends up with contracts, an internal team and nobody who can be asked why the numbers moved.
