Service
Ad creative testing
We run live tests of hooks, formats and messages, so budget follows what actually works. Each test starts with a question worth answering and ends with a written finding your team can use on the next round of ads.
What's included.
- review of past creative performance
- testing roadmap and hypotheses
- creative briefs for test versions
- test setup in your ad accounts
- budget and duration planning
- analysis against the agreed number
- written findings after each test
- running list of what you've learned
- recommendations for where budget should go
What creative testing is.
Creative testing is the practice of putting different versions of an ad in front of real audiences and comparing what they do. It tells you which idea, opening, format or message gets better results from the people you're trying to reach, based on behavior and not on opinion.
Creative testing can happen before launch or during it. Pre-launch methods, such as surveys and panels, ask people for their reactions. Our service is live testing, where the ads run on the platform with real budget, so the result reflects how people respond in their feed with everything else competing for their attention.
When creative testing makes sense.
Testing pays off when you spend enough for the difference between two ads to matter and you can make more than one version. It's especially useful on Meta, TikTok and YouTube, where the creative does much of the work of finding the right people. It also helps when a team keeps arguing about direction and needs evidence to settle it.
A small budget can't support many tests. If a test can't collect enough results in a reasonable time, you'd do better to compare two very different ideas than ten small variations, and to accept a rougher answer. Testing also can't help much when tracking is broken, or when the offer itself is the problem, since no ad can fix an offer people don't want.
What we test.
We test in order of likely impact, and we start with the big differences. The concept or message comes first, meaning the reason someone should care. Two different arguments for the same product usually tell you more than two background colors.
Hooks come next. On a video ad, the opening seconds decide whether anyone sees the rest, so we often test several openings on the same body. Then we look at format, such as short video, static images, carousels, or footage that looks filmed by a person next to polished production. Copy, headlines, calls to action and end cards can follow, once the bigger questions have answers.
How we set up a clean test.
We write down one question per test, what we expect to happen and which number will decide it. Changing only one meaningful thing at a time lets you say what caused the difference.
The setup matters as much as the idea. If you put several ads in the same ad set on Meta, the delivery system will often favor one early, on limited data, and the others never get a fair share. That's fine for running a campaign, and weak as an experiment. When we need a clean answer, we use the platform's own split testing tools, such as Meta's A/B test and TikTok's split test, or Google Ads experiments where they apply, so the audiences don't overlap and budget is divided evenly.
We agree on the budget and the length of the test before it starts. We let it run for at least a full week where the budget allows, in order to cover differences between weekdays and weekends.
Reading the results.
We read a test from the top of the ad down to the business result. Early signals, such as how many people watch past the first few seconds, how long they stay and whether they click, help explain why an ad worked. They shouldn't decide the winner alone, since an ad can earn plenty of cheap clicks from people who never buy. Where tracking allows, the decision rests on cost per qualified lead, cost per sale or return on ad spend.
We also report results that are too close to call. When two ads are within the margin of error, we say that, and you can run whichever one suits you.
A result has limits too. A winner in one audience or season may not hold in another, and ads tend to lose strength as the same people see them repeatedly. So we watch how frequency and results move over time.
How it works.
- 01
Review what's already run
We look at past ads and results in your accounts, using delegated access. The accounts stay with you or your agency.
- 02
Pick the questions
We list what's worth learning and rank it by likely impact. We also work out how many tests your budget can realistically support.
- 03
Brief the versions
We brief the variations, and your team or ours produces them. Creative production is identified separately in the scope.
- 04
Run the test
We set it up, check it and leave it alone until the agreed end, unless something is broken.
- 05
Write it down and act
We report what won, what lost and what we think it means. Budget then moves to what worked, under the approvals we've agreed with you.
How we measure creative testing.
Each test is measured by the number we picked before it started. Across a whole program, we look at whether your cost per qualified result is holding or improving as tests accumulate, and whether what you've learned is changing what gets made. We can't promise that any given test will find a winner, but a clear loser is still a useful answer.
How we price it.
We bill hourly and set the scope in hours. Strategy and advisory are CAD $175/h. Media management and execution are CAD $110/h. We confirm the engineering rate in the scope where technical work is needed. Third-party costs and advertising budgets are identified separately.
You can engage one discipline without buying the others. We can run the tests for your team, or plan them and let your media team run them.
Questions before we start.
What is ad creative testing?
It's running different versions of an ad against each other with real audiences to learn which one performs better on a number you've chosen in advance.
How much budget do we need to test creative?
It depends on what you're paying per result and how large a difference you want to detect. We work this out first and tell you if your budget can only support a simpler test.
How many ads should we test at once?
As few as answer the question. More versions split the same budget into thinner slices, so each one takes longer to produce a reliable result.
How long should a creative test run?
Long enough to reach the amount of data we agreed on, and usually at least a full week so weekdays and weekends both count.
What's the difference between creative testing and A/B testing?
An A/B test is a method that compares two versions of anything, including a landing page. Creative testing applies that kind of comparison to the ads themselves.
Related services.
Start a conversation.
Tell us what you are trying to move. We will tell you which disciplines are worth engaging, and which are not.