Demand Gen Experiments: Testing with a Google Credit Without Risking Search
A clinic had received a credit from Google to spend on ads. The team decided to use it on a Demand Gen campaign for one service, as an experiment. The goal was learning, not direct bookings. Just as important was the second rule: the experiment must not disturb the core Search campaign that already brought patients reliably.
Here is how to run a Demand Gen experiment with a credit, so that you learn something useful and protect what works.
What Demand Gen is
Demand Gen is a campaign type that shows visual ads across Google surfaces such as video and discovery placements, aiming to create interest, not just capture searches. It reaches people while they browse, before they search. It behaves differently from Search, and needs different expectations.
Why a credit makes experimenting easier
A credit reduces the cost of learning. You can test something new without touching your normal budget. But a credit is not free money. Treat it with the same discipline as any test: a clear question, clear rules and a written result.
Demand Gen: decide what you want to learn
Write down the question. Can this format reach people who do not yet search? Does it produce enquiries at all? Which images and messages attract clicks? A clear question keeps the experiment focused, and makes the result useful, whatever happens.
Expect learning, not bookings
For a clinic, Demand Gen is unlikely to produce many direct bookings, since people are not yet searching for care. Set the expectation that success means insight, such as audience signals and creative that works. Treating it as a learning tool avoids disappointment.
Protect the core Search campaign
Keep Search campaigns separate, with their own budget and settings. Do not move money from Search into the experiment, and do not let the experiment change Search targeting. Your reliable performer should carry on undisturbed while you learn.
Choose one service and one message
Test a single service with a clear, honest message and a landing page that suits it. Too many variables make results unreadable. A simple, focused test gives you a clearer answer than a broad one.
Set a budget and a time limit
Decide the amount, the length and the stopping rule. For example, run for four weeks with the credit, and stop when it is used. Write it down. Clear limits keep the experiment from turning into an ongoing habit.
Keep the creative honest and compliant
Visual ads must follow health advertising rules. Avoid before-and-after claims that break the rules, avoid targeting sensitive conditions and keep wording measured. A visually attractive ad that breaches rules wastes time and risks your account.
Track what you can
Set up tracking for clicks, engagement and any enquiries, and separate them from Search results. Look at view-through and assisted effects with caution, since they are easy to over-read. Record what you can measure and what you cannot.
Watch the effect on Search and brand searches
Check whether searches for your clinic name or service rise during the experiment. Any effect is worth noting, but treat it cautiously, since many things influence brand searches. Compare with your baseline, and look at trends, not single days.
Do not judge it too early
Early results with small budgets swing widely. Let the experiment run to its planned end, unless something is clearly wrong or unsafe. Then compare with your goals and the baseline.
Write up the result
Record the spend, the outcomes, what you learned and what you would do differently. Include what you could not measure. A short, honest note turns a one-off experiment into knowledge the team can use.
Decide what happens next
Choose whether to stop, adjust or repeat, and whether any insights should feed into Search. If Demand Gen did not help, stopping is a good result, because it saves future money. If it did, plan a careful next step.
Think about audience signals
One useful output of a Demand Gen test is learning which audiences and images attract attention. Note which ones performed well, and whether they differ from the people your Search campaigns reach. Those insights can shape your messaging, even if the experiment itself produces few bookings.
Decide how you will use what you learn
Before the test begins, agree who will read the results and what decisions they could influence, such as messaging, landing pages or future budgets. A test without a decision attached tends to be forgotten. Planning the follow-up makes the learning count.
Give the creative team clear guidance
If images and video are needed, brief the designer with the service, the audience and the tone, and remind them of advertising rules. Real photographs of the clinic and team usually work better than stock. Good guidance saves rounds of revision and protects compliance.
Avoid over-reading early data
Metrics like views and engagement can look impressive without meaning patients. Focus on what the experiment was designed to learn, and treat the flattering numbers with caution. Honest interpretation keeps the team from mistaking attention for demand.
A quick checklist
- Write the question you want answered.
- Expect learning, not direct bookings.
- Keep Search campaigns separate and untouched.
- Set a budget and time limit.
- Follow health advertising rules.
- Write up the result honestly.
Our companion article on setting a reporting baseline explains how to judge experiments fairly, and our guide to testing a new ad platform gives a wider framework. If you would like help, we are glad at Pulse Digital Health to talk.
Frequently asked questions
What is Demand Gen?
A Google Ads campaign type that shows visual ads across surfaces such as video and discovery placements, aiming to create interest before people search. It works differently from Search and needs different expectations. Write your question and your success measures before you start. A written question keeps the experiment honest.
Should I expect bookings from a Demand Gen experiment?
Probably few. People are not yet searching for care, so treat success as insight into audiences and creative. Direct bookings are a bonus, not the goal. Do not judge it against Search on bookings. Insight is the real prize, and bookings are a bonus.
How do I protect my Search campaign?
Keep it separate, with its own budget and settings, and do not move money into the experiment or change its targeting. Let the reliable performer carry on undisturbed. Keep budgets, settings and targeting for Search separate. A protected Search campaign keeps your steady performer steady.
How long should the experiment run?
Long enough to gather useful data but limited, for example around four weeks or until the credit is used. Write down the stopping rule in advance. Set the end date and the stopping rule in advance. Clear limits stop an experiment becoming a habit.
Are visual ads regulated?
Yes. Health advertising rules apply, including on results images and sensitive targeting. Keep wording measured and check current policies. Check policies for health ads and avoid sensitive targeting. Careful wording protects your account and your reputation.
What should I do with the result?
Record what you learned and decide whether to stop, adjust or repeat. Stopping is a valid result if it saves future spend. Record the outcome and decide the next step in writing. A written record turns an experiment into knowledge for the team.
