Test Concepts, Not Button Colours
Small tweaks — a different background colour, a reworded headline — rarely change results meaningfully. Big differences in concept do: a founder explaining why the product exists versus a customer unboxing it versus a side-by-side comparison with the alternative. Start by testing genuinely different ideas. Optimise the details of the winner afterwards.
"Three genuinely different concepts will teach you more in two weeks than thirty variations of the same ad will in two months."
Where Testing Happens
Keep testing separate from your main prospecting campaign so experiments do not disrupt proven ads. A dedicated testing campaign with a fixed share of budget — often 10–20% — lets you launch a batch of new concepts every week or two, give each a fair chance to spend, and promote the winners into the main campaign.
Read the Right Metrics, in Order
Judge a video ad as a sequence of moments:
1. Hook rate — did they stop?
The share of impressions that turn into a 3-second view. A weak hook means nothing else in the ad matters, because nobody saw it.
2. Hold rate — did they keep watching?
How many of those viewers stay for the body of the ad (for example, ThruPlays as a share of 3-second views). Low hold with a strong hook usually means the promise in the first seconds was not paid off.
3. Click-through rate — did they act?
Interest translating into a visit. Low CTR on an ad people watch often points to a weak or missing call to action.
4. Cost per purchase or lead — did it work?
The metric that decides. The others tell you why an ad won or lost, which is what makes the next batch better.
Concepts Worth Testing
- Founder or expert on camera explaining the problem the product solves
- Customer-style UGC: unboxing, first use, honest reaction
- Demonstration: the product doing the job, close up, in the first seconds
- Comparison: your product versus the usual alternative
- Objection-handling: answering the question that stops people buying
- Static image with one bold claim and proof — still a strong performer on many accounts
Scale, Iterate or Kill
Agree the rules before the test, not after. A practical version: give each concept enough spend to generate a meaningful number of clicks and at least a few conversions. If it beats your current best on cost per result, move it into the main campaign. If it has a strong hook but weak conversion, iterate — keep the opening, change the offer or the call to action. If it loses on every metric, switch it off and record what you learned.
Spot Creative Fatigue Early
Even great ads wear out. Watch for frequency rising while CTR falls and cost per result creeps up week over week. The answer is rarely to change targeting; it is to have the next set of winners already in testing before the current ones tire. That is why creative needs a steady production rhythm, not a one-off shoot.
The Same Thinking Works on YouTube
YouTube viewers can skip after five seconds, so the hook discipline is the same: show the product, the problem or the payoff immediately, brand early, and design for sound-off viewing with captions. Vertical cuts of the same concept can run on YouTube Shorts and Instagram Reels.
A creative testing system is less about genius ideas and more about volume, honest measurement and a steady rhythm. Brands that commit to it stop asking "why did our ads stop working?" — because they always have the next winner waiting.