The Creative Testing Framework We Use Across Every Paid Media Account

Most brands test the wrong things. They run A/B tests on button colour, headline capitalisation, or emoji placement — micro-variables that will never move the needle enough to matter at realistic ad spend levels. Meaningful creative testing is about angles, not assets. An "angle" is the strategic lens through which you present your product: the problem it solves, the identity it confers, the fear it alleviates, the aspiration it enables. Getting the angle right produces 3–10× performance differences. Getting the emoji right produces noise.

Before producing a single piece of content, map out 6–8 distinct angles. For each angle, write a one-sentence strategic summary: - **Pain-led**: "Our product solves [specific painful problem] that [target audience] experiences [frequency/context]." - **Identity-led**: "People who use our product are the kind of person who [aspirational identity statement]." - **Social proof-led**: "[Specific metric] customers achieved [specific result] in [specific timeframe]." - **Curiosity-led**: Opens with an unexpected claim or counterintuitive statement. - **Authority-led**: Expert endorsement, certification, or industry recognition. - **Value-led**: Price comparison, cost-per-use, or ROI calculation. Each angle should be genuinely distinct. If two angles say essentially the same thing with different words, collapse them into one.

We test on a rolling 4-week cycle: **Week 1**: Launch 2 new angles (3 creative executions per angle = 6 new ads). Minimum budget per ad set: £30/day to generate enough data. **Week 2**: Review performance at 1,000 impressions per creative. Kill anything with CTR below 0.8% (for cold audiences) or CPM that is 40% above account average. Retain and increase budget on top performers. **Week 3**: Scale winning angles. Test creative variables within winning angles (format, hook variation, CTA wording). Begin producing the next batch of 2 angles. **Week 4**: Review month's learning. Document: which angles won, why (hypothesis), what to test next month. Archive all creative with performance data. This cycle means you are always testing, always learning, and never without fresh creative in the pipeline.

Formal statistical significance (95% confidence interval) requires sample sizes that most advertising budgets cannot generate per creative. At £50/day per ad set, you will not reach 95% confidence on a conversion metric within a reasonable timeframe. Practical thresholds we use: - 1,000+ impressions: enough to judge CTR and thumbstop rate - 50+ link clicks: enough to judge landing page quality - 20+ add-to-carts: directional signal on purchase intent - 10+ purchases: actionable conversion data For accounts spending under £5K/month, optimise on CTR and add-to-cart rate as proxy metrics. For accounts spending above £10K/month, wait for purchase data before scaling. Scaling on proxy metrics at low spend is acceptable; scaling on proxy metrics at high spend is expensive guesswork.