Running agencies across multiple hospital units: the roster model
A hospital agency roster usually grows by accident as units hire local firms and the centre adds specialists. Design it deliberately instead. Start from who decides what, pick a roster shape that fits those decision rights, give each shared asset one owner, run one medical review route for all patient content, measure agencies on common definitions and keep every account in the hospital’s name.
Ask a hospital group how many agencies it works with and the honest answer is usually “we would have to check”. The group brand has a creative agency. Performance media sits with another. Two or three units hired their own local firms before anyone at the centre noticed. A video studio makes doctor content, an SEO partner looks after the website, and someone in the international desk has a freelancer running social posts in Arabic and Russian.
None of these decisions was wrong when it was made. Together they produce a hospital agency roster that nobody designed: overlapping scopes, duplicated fees, inconsistent medical review and a brand that looks slightly different in every city. The fix is not to fire everyone and appoint one agency. It is to design the roster the way you would design any other part of the operating model, deliberately and around how the group actually makes decisions.
This piece is about that design. Managing an individual vendor well is a separate skill, covered in managing vendors without becoming their project manager. Here the question is structural: how many agencies, doing what, briefed by whom, across how many units.
How the sprawl happens
Agency sprawl in a multi-unit group follows a predictable pattern. A unit opens or is acquired, and its leadership wants visible marketing quickly. The central team is stretched, so the unit hires a local agency it trusts. That agency does reasonable work and builds a relationship with the unit head. Two years later the group has several such relationships, each defended by the unit that owns it.
Meanwhile the centre adds specialists as new needs appear. A performance agency when digital spend grows. A content studio when doctor videos become important. An online reputation partner when reviews start to matter. Each is hired for a real gap, and each is scoped on its own, without reference to the others.
The costs are mostly invisible. Two agencies bid against each other on the same search terms. A unit’s local firm uses an old logo. Doctor profiles are rewritten three times by three teams. Medical claims pass through one review process at the centre and no review at all in a unit. And every agency reports success in its own metric, so nobody can compare them.
Three shapes a hospital agency roster can take
In my experience, most workable rosters fall into one of three shapes. The right one depends less on agency capability and more on how much authority the centre holds over unit marketing.
One lead agency with specialists beneath it
A single agency holds the relationship and subcontracts or coordinates specialists. The group gets one point of accountability and one integrated plan. The risks are cost layering, a lead agency that is weak in some disciplines but will not admit it, and dependence on one firm’s continuity. This shape suits groups with strong central control and a relatively uniform brand.
A specialist roster run by the centre
The central team contracts separate agencies for creative, media, content, search and reputation, and integrates them itself. This gives the best talent in each discipline and clearer commercial terms, but it needs an in-house team capable of briefing, integrating and refereeing. Without that team, the roster fragments again within a year.
A hub with local partners
The centre owns brand, media buying, the website and core content through a small set of agencies. Units keep a local partner for on-ground activation, camps, community events and regional language work, within guardrails set by the centre. This is often the most realistic model for Indian groups spread across very different cities, because it respects the fact that a unit head knows the catchment better than anyone at the centre.
My own preference, for most groups I have seen, is the hub with local partners. It admits a political reality rather than fighting it, and it puts the expensive, scalable work where scale actually helps.
Start from decision rights, not from agencies
Before choosing a shape, write down who decides what. Who owns the brand? Who sets the media budget for each unit? Who approves medical content? Who decides which specialties get promoted in which city? Who can hire a vendor?
The roster should mirror those answers. If unit heads control their own budgets, a roster that routes all briefing through the centre will be bypassed. If the centre owns the brand, a unit cannot be allowed to commission its own identity work. Most roster failures I have seen are really decision-rights failures, and the question sits inside the wider design discussed in designing the growth operating model around shared digital assets.
The politics are real. Moving a unit’s favourite agency out of the roster feels, to the unit head, like losing control. Handle it as a change in who does what, with a clear promise of what the unit will receive in return, and a date when both sides will review whether the promise was kept. See the politics of centralisation for how that conversation tends to go.
One source for the assets every agency touches
The single most useful structural decision is to define a small set of shared assets that only one party can change, and that every agency must draw from. In a hospital group, these usually include the brand identity and its rules, the master list of doctors with approved profiles and photographs, the approved description of each specialty and centre, the website and booking flows, and the tracking setup that defines what counts as an enquiry.
Each shared asset needs one owner. Agencies can propose changes, but they cannot make them directly. This sounds bureaucratic until you have watched two agencies publish different consultation timings for the same doctor, or a local firm run an ad with a phone number that bypasses the contact centre.
How the brand flexes across units also needs to be settled before the roster is. If the group has not decided whether units carry the parent brand, a co-brand or their own name, every agency will interpret it differently. The thinking in brand architecture across a multi-unit group should come first.
Briefing across units without a traffic jam
Once the roster is designed, the daily risk is that everything funnels through one overloaded central marketing manager. Units feel slow service, start going around the process, and the old sprawl returns.
What works better is a simple briefing calendar. Each unit gets a regular planning slot with the central team and the relevant agencies, where it brings its priorities for the coming period: specialties to push, doctors joining, camps planned, capacity available. The central team translates those into briefs, checks them against group campaigns and capacity, and sends them to the agencies in one consolidated form.
Urgent requests still happen, and a fast lane for genuine urgency, such as a new doctor joining or a service disruption, keeps units from inventing their own routes. Everything else waits for the calendar. Units accept this far more readily when the calendar is reliable.
Medical review has to cover the whole roster
Every agency on the roster creates content that patients will read as the hospital speaking. That makes medical and legal review a roster question, not an agency question. A single review route, with named clinical reviewers for each specialty and a clear turnaround, should apply to every piece of patient-facing content, whichever agency made it and whichever unit asked for it.
Local partners are where this most often breaks down. A camp poster, a regional language video or a social post in a unit’s own account can carry claims nobody at the centre has seen. Build review into the local partner’s contract, not just into the central process, and make it easy: a short turnaround, a named reviewer and a shared library of pre-approved statements for common specialties.
Agencies should never be the last line of defence on clinical accuracy. That responsibility stays with the hospital.
Comparing agencies fairly
A roster only improves if you can see which parts of it are working. That requires common measures across agencies, set by the hospital, not by each agency’s own report.
For most roster members the measures that matter are the ones closest to the patient: qualified enquiries by specialty and unit, conversion to appointments, cost per appointment where spend is involved, content approved and published on time, and quality scores from medical review. Creative agencies are harder to measure directly, but brand tracking, share of search and the performance of their assets in paid channels are reasonable proxies.
- Agree the definitions of enquiry, appointment and attribution window across the roster before the year starts.
- Give every agency access to the same dashboard, so nobody works from a private version of the truth.
- Review the roster as a whole each quarter, not agency by agency, and look for overlap and gaps.
- Hold one annual review where each agency’s scope is reconsidered, not just renewed.
Budget allocation across units sits alongside this. If you want a structured starting point for how spend might be split, the hospital marketing budget calculator helps frame the conversation with unit heads and finance.
What the in-house team must hold
Every roster shape depends on a central team that can do certain things without an agency. When that team is missing, agencies fill the gap, and the hospital slowly loses control of its own marketing.
At minimum, the in-house team should write briefs, own the shared assets, run the briefing calendar, manage the review route, hold the dashboard and make budget calls. It should also hold the institutional memory: which campaigns worked in which city, which doctors are comfortable on camera, which specialties have capacity and which do not. Agencies change. That knowledge should not leave with them.
I would add one more capability: someone who understands the contact centre and CRM well enough to see what happens after an agency’s work produces an enquiry. Agencies can generate interest. Only the hospital can see whether that interest became an appointment, and only the hospital can fix it when it did not. A roster that ends at the lead form is a roster measured on the wrong thing.
If the central team is small, choose the roster shape that asks least of it, and grow the shape as the team grows. Designing a specialist roster for a small team is a common and expensive mistake.
Commercial terms that do not reward complexity
Agency fee structures can quietly push the roster in the wrong direction. A lead agency paid a margin on everything it subcontracts has little reason to simplify. A media agency paid on spend has little reason to argue for less of it. A content studio paid per piece will make more pieces.
Where you can, tie a portion of fees to shared outcomes that the whole roster influences, such as appointments from digital channels or approved content published to plan. Keep scopes explicit and non-overlapping, and ask each agency at renewal what it would stop doing if the budget were cut. Their answer tells you how well they understand their own value.
Also protect the group’s ownership of its assets. Ad accounts, analytics properties, website code, content libraries and CRM data should sit in the hospital’s name, with agencies given access, not ownership. Changing an agency should never mean losing history.
A roster review you can run this quarter
If the current roster grew by accident, a clean first step is a short review rather than a big pitch process. List every agency and freelancer working for any unit, with scope, contract owner, fee structure and the accounts they control. Mark overlaps. Mark gaps. Mark where assets sit in an agency’s name.
Then settle three decisions with the CEO and unit heads: which roster shape fits your decision rights, which shared assets have one owner, and which single review route applies to all patient-facing content. Only after that should you consider changing agencies.
You will probably find that a few relationships need to end, a couple of scopes need to be merged and one or two local partners are doing valuable work the centre never noticed. That is a good outcome. A roster that is designed, even imperfectly, will serve patients and units far better than one that simply accumulated.
Questions people ask
A hospital agency roster is the full set of external agencies and freelancers that support a hospital group’s marketing, such as creative, media, content, search, reputation and local activation partners. Designing it means deciding how many agencies to use, what each owns, who briefs them, which shared assets they draw from and how their work is reviewed and measured across units.
Not necessarily. Many groups keep local partners for on-ground activation, camps and regional language work, because unit teams know their catchment best. What changes is that the local partner works within group guardrails for brand, medical review and tracking. In return, the unit should receive faster central support for media, website changes and doctor content.
Not always. A lead agency simplifies management but may add a margin on subcontracted work and can be weak in some disciplines. A specialist roster can be more cost effective and higher quality, but only if the in-house team can brief and integrate the agencies. The right answer depends on your internal capability more than on agency rates.
The CFO should look for overlapping scopes, fee structures that reward activity rather than outcomes, and assets held in agency names. A clear list of every agency with its scope, contract owner and fee basis is the first step. Tying part of the fees to shared outcomes, such as appointments from digital channels, helps align spend with results.
Use one review route for all patient-facing content, whichever agency created it and whichever unit requested it. Name clinical reviewers for each specialty, set a clear turnaround and keep a library of pre-approved statements. Build the review requirement into every agency contract, including local partners, so no content reaches patients without the hospital checking it first, whatever the deadline pressure.
The hospital should own every ad account, analytics property, website codebase, content library and CRM dataset, with agencies given access rather than ownership. This protects history and bargaining power. If an agency relationship ends, the group keeps its data, audiences and performance records, and the next partner can start from a known baseline. Put this in every contract from the start.
The review itself can be done within weeks: listing agencies, scopes, owners and overlaps is mostly a matter of collecting contracts and access lists. Agreeing decision rights with unit heads takes longer, because it involves politics. Transitioning agencies should follow natural contract renewal points where possible, to avoid disruption to live campaigns, patient enquiries and the doctors who have built working relationships with particular agency teams.
IT should be involved wherever agencies touch the website, booking flows, tracking, forms and patient data. Agencies often request access that creates security or DPDP risk. A clear access policy, with named accounts, limited permissions and a process for removing access when an agency leaves, is part of good roster design and protects patient data.
Use shared measures where possible, such as qualified enquiries, appointments and cost per appointment for media, and on-time approved content for studios. Creative work is harder to measure directly, so use proxies like brand tracking, share of search and how creative assets perform in paid channels. Review the roster as a whole rather than each agency in isolation.
Mostly, a reliable service. If units receive timely support through a regular briefing calendar and a fast lane for genuine urgency, they have little reason to hire their own vendors. Where units bypass the process, it usually signals slow central service or unclear decision rights, and both should be fixed before tightening rules. Enforcement without better service simply drives the workarounds out of sight.
Frequent pitching is expensive for everyone and damages continuity. A better habit is an annual scope review for each agency, where the hospital asks what it would stop doing, where it overlaps with others and what it has learned. Run a full pitch only when a relationship is clearly failing or the roster shape itself is changing, and give incumbents a fair chance to compete.
The board rarely needs agency detail. It needs to know that marketing spend is controlled, that brand and medical content are governed consistently across units, and that the group owns its data and digital assets. A short note on roster structure and any material changes, within the regular growth report, is usually enough. Detail belongs with the executive committee, not the board.
More than most groups expect. Someone has to run the briefing calendar, own shared assets, integrate agency plans and referee disagreements. A specialist roster needs the most in-house effort, a lead agency model the least. Underestimating this effort is the most common reason carefully designed rosters fragment back into sprawl, usually within a year of the redesign.
Start with a complete list of every agency and freelancer, including scope, contract owner, fees and the accounts they control. Mark overlaps, gaps and assets held in agency names. Then agree roster shape, shared asset ownership and one review route with the CEO and unit heads before making any changes to the agencies themselves. Most of the value comes from those decisions, not from the choice of agencies.
