This post is for hospital and clinic-chain marketing teams that run Google Ads on their own name and want to know whether that spend produces bookings or simply re-labels bookings that would have happened anyway. It is a method, not a result. We do not publish a benchmark here, because the right answer depends on your market, your competitors and how well your organic listing performs.
Branded campaigns are the easiest ones to defend. Cost per click is low, conversion rate is high, and the cost per enquiry looks excellent next to a generic "knee replacement" campaign. That is precisely the problem. A person who types a hospital's name into Google has already decided to consider it, and the ad is often the first of two links to the same website. The reporting credits the ad. The question is how many of those patients would have clicked the organic result, the Google Business Profile, or called the number in the knowledge panel if the ad had not been there.
If you manage a wider account, our Google Ads audit covers the rest of the structure. Here we deal with one question and one test.
What "incremental" means for a branded click
An incremental conversion is one that would not have happened without the ad. For branded search there are three outcomes when you remove the ad:
- Recaptured. The patient clicks your organic result or profile instead. The booking happens; you saved the ad cost.
- Lost. The patient does not find another route to you, or a competitor or aggregator ad takes the click. The booking is gone.
- Shifted. The patient still reaches you, but through a slower or less trackable route, for instance by calling the main switchboard instead of a tracked landing page.
The ad's true value is the lost group. Its apparent value in last-click reporting is all three groups together. The gap between the two is what the test measures. This is the same logic behind any attribution question: the credit a report assigns is not the same as the contribution.
Why hospitals differ from other advertisers
Three features of healthcare search change the maths.
Competitors and aggregators bid on hospital names. Doctor-discovery platforms, other hospitals and lead-generation sites often bid on a hospital's name. If you pause, they may take the top of the page. Use the auction insights report on your brand campaign to see who is showing alongside you, and how often.
Brand queries include doctor and department names. Your brand term set is wider than the hospital name. "Dr [surname] cardiology Gurugram" is a branded query. So is a misspelling, a short form used by locals, and "[hospital] appointment". Decide the term list before the test, and keep it fixed.
The SERP is crowded with free assets. The map pack, the knowledge panel, sitelinks and the profile's call button all compete with the ad. If those assets are strong, recapture tends to be higher. If your listing is weak, the ad may be doing work that the organic result cannot.
Test designs, from simplest to most reliable
Option 1: Time-based pause
Pause the brand campaign for three to four weeks and compare with a matched pre-period. This is the easiest design and the weakest. Seasonality, a campaign launch, festival weeks or a news story can all move branded demand during the test. Use it when nothing else is feasible, and say so in the report.
Option 2: Geographic holdout
If you operate in several cities, pause brand ads in some and keep them running in others. Match the pairs on baseline branded demand and population type (for example, two similar tier-2 cities). Compare the change in total branded conversions in paused cities with the change in kept cities over the same weeks. This controls for time-based noise, because both groups experience the same calendar.
A single-site hospital can approximate this by splitting by pincode group where the ad platform allows location exclusions, but the groups are small and overlap through people who travel. Treat the result as directional.
Option 3: Partial pause
If stopping everything is too risky, split the brand term set. Pause the generic-brand terms (the hospital name alone) and keep the competitive ones. This answers a narrower question but it is a safer first step.
Whichever design you use, write the hypothesis, the term list, the duration, the paired locations and the success rule down before starting. The most common failure is to decide the rule after seeing the numbers.
What to measure
Count conversions across all routes, not only the ad's own:
- Search Console: clicks and impressions for a saved brand filter. Create a query filter with a regular expression covering your hospital name, abbreviations, misspellings and doctor names, and check how clicks, impressions and position are defined in Google's performance report documentation. Search Console anonymises some rare queries, so the filtered total will undercount; use the same filter on both sides of the test and the bias cancels out.
- Google Ads: brand campaign clicks, conversions and impression share before the pause.
- Business Profile: calls, direction requests and website clicks from the profile insights.
- Call tracking and CRM: total appointments requested, with source fields, including calls to the main number. If your CRM source field is poor, fix that first.
- Direct and referral sessions in GA4, because some patients will switch to typing the URL.
The headline metric is total branded conversions across all routes. Ad-attributed conversions alone will fall to zero during the pause by definition and tell you nothing.
