Ad Creative Testing
Every ad has a lifespan. A testing system keeps a queue of replacements ready before the current winner tires.
A test changes one thing
If a new ad has a different image, headline, offer and audience, and it does better, you have learned nothing you can repeat. A valid test changes one variable and keeps the rest still. That discipline is slower for a week and much faster over a year.
A sample test plan
| Question | What changes | What is compared | Risk |
|---|---|---|---|
| Does the first three seconds matter? | Opening scene only | Watch-through rate, cost per lead | Different scenes may attract different people |
| Does a price mention help or filter? | Headline with price against without | Lead volume and quality | Quality needs to be measured after the lead |
| Is a person more persuasive than a product? | On-screen person against product only | Click rate, lead quality | Small budgets produce noisy results |
| Which offer works? | Offer wording | Enquiry rate on the same page | Do not change the page too |
What to judge, and what to wait for
- Judge by the outcome closest to the goal, not the metric that looks best. A high click rate on an ad that produces empty leads is a loss.
- Give each test enough time and enough events to separate signal from luck. Very small numbers flip week to week.
- Record every test, the result and the reason you think it happened. The log is worth more than any single ad.
Our rhythm
New creative is tested weekly and weak ads are stopped, which matches our published creative system of weekly tests and a weekly report. We do not promise that every test wins. Many do not, and a log of the losers is useful too.
What can actually be tested
| Variable | Example | What it tells you |
|---|---|---|
| Hook | The first three seconds or first line | What earns attention |
| Format | Video against static image | How people prefer to receive the message |
| Proof | A person’s story against a fact | What builds belief |
| Offer | Free call against a small paid trial | What people will commit to |
| Call to action | Different wording or a different route | How to move people to act |
Keeping a log that can be read later
Name each ad so it tells you what it tests: the offer, the format, the hook and the date. Record the hypothesis in a sentence and the result in another. After a few months, this log is more valuable than any individual ad, because it holds what you have learned about your own audience.
Testing when money is limited
With few conversions a week, formal comparison is unreliable. Do fewer tests, make the differences between versions large, and judge on sturdier signals such as click rate and cost per click first, then on enquiries over a longer window. It is better to say honestly that a test is inconclusive than to declare a winner from two leads.
Signs that a winner is wearing out
- Click rate drops while spend holds steady.
- Frequency climbs and comments turn impatient.
- Cost per lead rises with no change to the page or offer.
- The ad has been the main spender for a long time.
A three-week testing calendar
| Week | Test | Decision rule |
|---|---|---|
| 1 | Two hooks on the same video body | Keep the hook with the stronger enquiry rate, if the gap is clear |
| 2 | Video against a static image with the winning hook | Keep the format that produces better conversations |
| 3 | Two offer wordings on the winning creative | Keep the one that brings better-quality leads |
Ground rules that keep tests honest
- Do not edit an ad while it is being tested.
- Do not add new tests on the same audience while another is running.
- Record the result even when it is a draw.
- Retest a surprising winner before building a strategy on it.
Where new creative comes from
A testing system needs a steady flow of material. Ours draws on reel and video production, short shoots planned around test ideas, and re-cuts of footage that already exists. If your brand cannot produce enough material, tell us early.
Show us your ads
A few of your current ads, plus the ones you believe are working, tell us what to test next.
Show us your ads WhatsApp us