PMax Asset Groups by Service Line for a Multi-Specialty Hospital

This post is for the digital lead or performance marketer at a multi-specialty hospital who already runs Performance Max and has noticed that one campaign now serves cardiology, orthopaedics, oncology and maternity from a single pot of assets. It assumes you know what an asset group is. It answers a narrower question: how to divide one hospital's PMax setup by service line so that each line has its own message, landing pages, signals and reporting, without fragmenting budget into pieces too small to learn from.

Our earlier article on when Performance Max suits healthcare lead generation covers whether to use the campaign type at all. This one assumes you have decided yes, and deals with structure. The hospital's page architecture matters here too, so it helps to have read how service-line pages are built around the patient's next decision and how patient journeys differ by department.

Nothing below is a performance promise. The structure is a way of making results legible. Any clinical wording in assets must be approved by the treating department before it goes live.

Why one mixed asset group fails

A single asset group for a whole hospital forces the system to assemble ads from headlines about knee replacement, cancer screening and antenatal care in one pool. The combinations can be incoherent, and the landing page the campaign sends traffic to cannot be right for all of them. More importantly, reporting collapses. If the campaign reports forty enquiries, you cannot tell whether they came from a maternity package or a bariatric consultation, and you cannot tell whether the cost per enquiry is good for either.

Different service lines also differ in what an enquiry is worth, how long the decision takes and who responds. A health-check package may close in days. An elective joint replacement may take weeks of second opinions. Pooling them under one cost-per-lead target hides the difference and lets the cheap, fast line flatter the average.

Campaign or asset group: the decision rule

Asset groups share a campaign's budget, bid strategy, location targeting and conversion goals. That gives a clean rule for what to split where.

• Same budget logic, same conversion goal, same locations: separate asset groups inside one campaign.
• Different conversion goal or value, different branch locations, or a budget you must protect: separate campaigns.
• A line with too little volume to learn: keep it as an asset group inside a larger campaign rather than starving it in its own campaign.

A hypothetical example, clearly illustrative: a 200-bed hospital in Pune might run one PMax campaign for consultation-booking service lines across its main campus, with asset groups for orthopaedics, cardiology, gastroenterology and women's health. A separate campaign would cover its satellite clinic in another part of the city, because the location target and budget differ. The oncology line, where enquiries are rarer and more sensitive, might sit in a campaign of its own with its own goal, or stay out of PMax entirely and run on Search.

A blueprint for each service-line asset group

Each asset group needs the same five parts. Writing them in a table before touching the account keeps the departments and the agency aligned.

Keep the asset group's name identical to the service-line name used in your CRM or enquiry sheet. When the account's asset group report and the enquiry log use the same label, reconciliation takes minutes rather than a week.

Search themes: topic phrases, not a keyword list

Search themes let you tell the system what your customers search for. Google describes them in its Performance Max help documentation, which also notes that audience signals are suggestions and that the campaign may show ads to relevant audiences outside them. Treat search themes the same way: they steer, they do not fence.

For a service-line asset group, write themes at the level a patient would phrase a need: the procedure name, the symptom cluster, the specialist title, and the city. Keep the themes for one line out of the others. If the orthopaedics group carries a theme for "heart check", you have defeated the point of the split. I could not confirm a current numeric cap on search themes from Google's own pages, so check the limit in your account rather than relying on a figure from a third-party guide.

Search themes are also where queries that should never be bought need care. Procedures that are legally or ethically sensitive may be restricted. Read Google's healthcare and personalised advertising policies before adding themes for such procedures, and have the compliance owner sign off.

Audience signals and the health policy constraint

Google's personalised advertising policy lists Health as a sensitive interest category and states that advertisers promoting it cannot use advertiser-curated audiences. The policy table says Customer Match, your-data remarketing segments, audience expansion and lookalike segments are not supported for Health ads, while predefined Google audiences such as in-market segments remain available. Which of your service lines count as Health under the policy is a question to settle with Google's policy documentation and your account team, not to assume.

In practice this means a hospital asset group may not be able to carry the audience lists that retail accounts take for granted. Where the policy blocks a list, the signal for that group comes from the content you control: specific assets, a specific landing page and specific search themes. This is another reason to keep groups narrow. A narrow group with a precise page gives the system more to work with than a broad group with no usable list.

Landing pages decide whether the split pays off

The split only works if each group lands on the right page. Send the cardiology group to the cardiology hub or a specific procedure page, not to a general "book an appointment" form. Where the hospital has no good page for a line, build the page before building the asset group. A paid click arriving at a department page that lists every machine and no route to a booking will waste the structure you created.

Check three things per page: it loads quickly on a mid-range phone, the phone number and booking action are visible without scrolling on mobile, and the page names the same service as the asset group's headlines. The Performance Max service page covers how we approach this alongside the rest of paid search.

Brand traffic will contaminate the numbers

A hospital's own name is the most searched term in the account. Left alone, PMax will happily claim those searches because they convert cheaply, and the service-line cost per enquiry will look better than it is. Google's help page on brand settings for Search and Performance Max explains brand lists and exclusions, which apply to Search and Shopping inventory in Performance Max. Set the exclusion before you compare service lines, otherwise the lines with the strongest brand pull win by accident. A dedicated article in this series will go deeper on that mechanism; here, the point is simply that service-line comparison is only fair on non-brand demand.

How to read the split

Once the groups run, review them on a fixed cadence, not daily. Asset group reporting in PMax shows asset ratings and group-level performance. Use it for three decisions only.

1. Which assets to replace. Replace assets the report rates low after the group has had enough impressions to be judged, and keep the wording approved by the department.
2. Which group is under-supplied. If a group receives almost no impressions, check the final URL, the search themes and whether the audience restriction is blocking delivery.
3. Which group should graduate. A line with steady, qualified enquiries and a distinct goal may deserve its own campaign.

Do not judge groups on cost per lead alone. Join the enquiry log to the account by asset group name and look at how many enquiries became booked and completed appointments for each line. A line with a higher cost per enquiry but far better completion can be the better line. If you need a method for joining that data, our piece on offline conversion tracking for clinics is a starting point.

For a worked method with your own data, export the last quarter's enquiries by service line, count enquiries and completed appointments per line, and then compare those counts against spend per asset group. Do the arithmetic yourself; do not borrow benchmarks from another hospital's account.

Common failure modes

• Overlapping search themes across groups, so two lines compete for the same queries and neither reads cleanly.
• A shared landing page across groups, which makes the groups cosmetically separate and operationally identical.
• Too many groups for the budget. If a group cannot reasonably collect a few conversions a week, merge it with a related line.
• Unapproved clinical claims in auto-generated assets. Review any generated text before it serves, and switch off options you cannot review.
• Changing everything at once. Alter one group at a time so the report can tell you what changed.

Where to start

1. List the service lines, with the department owner, the best landing page and the enquiry label for each.
2. Mark which lines have enough weekly enquiry volume for their own group, and which should be merged.
3. Set brand exclusions before you compare lines.
4. Check each line against Google's personalised advertising policy for health and record the decision.
5. Write assets and themes per line and get clinical sign-off.
6. Launch one group at a time if you are migrating an existing campaign, and keep a change log.
7. Review monthly against booked and completed appointments, not form fills alone.

If you want a second opinion on structure, our Google Ads audit looks at exactly this: how the account is divided, what each division is optimising for and whether the reporting can answer the CMO's questions.