Every ad campaign effectively starts with a bet: "we think our message will resonate with our audience." Then the creative gets approved, the media budget gets committed, and the ads go live. Only then do you find out whether the message is actually persuasive. By the time performance data comes back, you've already served your message to a portion of your servable audience. And if you've A/B tested your ad, then 50% of impressions are automatically served a losing message.
Ad testing through market research flips that order. Instead of learning what works in-market, you learn before launch, with evidence of which ad concept or strategy actually changes what people think and do.
With Grow Progress, teams can pre-test ad concepts with randomized controlled trials and get results in as little as 24 hours. See how Rapid Message Testing works.
What is ad testing market research?
Ad testing market research is the practice of exposing a representative sample of your target audience to ad creative before launch, and then measuring how they react. The goal is to find out which version of an ad is most likely to achieve your objective, and which versions could backfire, all without exposing your marketable audience to a live version of the ad.
Ad concept testing is the earliest form of this work. It tests the core idea behind an ad (the frame, the story, the argument) before you invest in full production. Testing concepts early is more cost effective than testing finished spots, and it tells you which direction is worth producing at all.
Later-stage testing can then compare finished executions: different cuts of a video, different headlines, different calls to action.
Engagement vs. persuasion: what to measure in your market research
The most common mistake in ad testing is measuring the wrong metric. Many tests ask people whether they liked an ad, found it believable, or simply clicked on it. Those are reactions to the ad, not evidence that it actually worked.
An ad can be rated poorly in an A/B test and still shift attitudes. An ad can have great engagement but poor persuasion. Engagement metrics like clicks and view-through rates have the same blind spot. Our research on clean energy ads found that the ads people engaged with most were not always the ones that persuaded them.
A stronger approach is to measure the outcome you actually care about, without referencing the ad:
- Favorability toward your brand, candidate, or organization
- Support for a policy or position
- Likelihood to donate, volunteer, vote, or buy
- Agreement with a core belief behind your campaign
- How important the issue feels
Then compare people who saw the ad against a control group of people who didn't.
Common ad testing methods
Each method answers a different question. The right choice depends on whether you need reactions, ideas, or proof of impact.
| Method | What it tells you | Limitation |
|---|---|---|
| Focus groups | How a small group talks about your ads | Small samples, and group dynamics skew responses. Read about why many focus groups fall short |
| Ad rating surveys | Which ads people say they like or find persuasive | Measures stated preference, not actual attitude change |
| In-market A/B tests | Which ad drove more clicks or conversions | Requires live spend, and part of your audience always sees the losing ad |
| Randomized controlled trials | How much each ad changed attitudes and intentions, by audience | Measures effects in a test setting, so results guide decisions rather than guarantee field performance |
Randomized controlled trials (RCTs) are the same method used in medical research. A sample is randomly split into groups. Each group sees one ad, and a control group sees no ad. Because assignment is random, any difference in outcomes is caused by the ad itself. Learn more about the science of RCTs.
How to run an ad concept test in 5 steps
1. Define what success looks like
Pick one or two outcomes the ad needs to move, such as favorability, donation intent, or policy support. Everything else in the test flows from this choice.
2. Start from what's already been tested
Search the Persuasion Library, which holds 8,000+ tested messages searchable by topic and audience. Seeing which frames have moved people on similar issues saves you from starting at zero.
3. Write distinct concepts, not small variations
Draft three to six concepts that differ in their central idea. Testing two shades of the same headline teaches you little. Testing an emotional personal story against a hard economic argument teaches you a lot.
4. Run a randomized test
Put your concepts into a Rapid Message Test. Each concept is measured against a control group on your success outcomes, with results in as little as 24 hours. For video, in-scroll testing shows ads the way people actually encounter them in a feed.
5. Read the subgroups before you pick a winner
The concept that wins overall may not win with the audience you most need to reach. Check results by party, age, gender, or other segments that matter to your campaign, and look for backlash as closely as you look for lift.
When to pre-test your ads
The best time to test is before creative is locked and before media is bought. That usually means three points in a campaign:
- Before production. Test concepts as scripts, storyboards, or static mockups. This is where testing saves the most money, because you only produce what works.
- Before launch. Test finished executions to choose your lead ad and decide where audience-specific versions are worth the cost. Pre-testing for skippable video is especially useful here, since most viewers only see the first few seconds.
- Mid-flight. When news breaks or the conversation shifts, a fast test confirms whether your message still fits the moment.
Testing also isn't a one-time event. The strongest campaigns test continuously, building a body of evidence about what works with their audiences over time.
Common ad testing mistakes
- Asking people if the ad persuaded them. Questions like "Does this ad make you more likely to support X?" measure what people think about their own reactions, which they're bad at judging. Measure the outcome directly instead.
- Optimizing for engagement alone. Clicks and views show attention, not persuasion. An ad can win the click and lose the argument.
- Testing too many near-identical versions. Small wording changes rarely produce detectable differences. Test ideas that are genuinely different.
- Only reading the topline. A concept that performs well overall can backfire with a key audience. Always check subgroups.
- Ignoring null results. An ad that moves nothing is a real finding. It tells you where not to spend, which is often the most valuable result in the test.
- Testing too late. If production and media are already locked, test results can only confirm a decision, not shape it.
Know what works before you spend
Pre-testing turns ad spend from a bet into a decision. World Central Kitchen tripled its return on ad spend after pre-testing its creative, and Save the Children's ad tests revealed trade-offs no performance dashboard would have shown.
Stay ahead of planning with the Grow Progress Pre-Launch Playbook.
Talk to our team to see how ad testing market research can strengthen your next campaign.