How Different Do Iterations Need To Be?

Hey!

Welcome to a new newsletter I’m calling ‘Creative Corner’ where I’ll be pulling back the curtain on exactly how we go about producing winning ads at my agency Adcrate, as well as sharing strategies, tips and examples every week.

This week, I’m sharing the results of a study I did on some of our ad accounts about the effectiveness of iterations.

So let’s get into it 👀

I remember reading a tweet from Ash Melwani saying “if your iteration looks REMOTELY similar to the original, Meta can’t tell the difference.”

So I decided to put it to the test.

I analysed 124 iterations across 17 ad accounts, 11 industries and $12.4m of adspend to see if there is actually a correlation between changes made in an iteration and lift in performance...

The Study

To test the hypothesis, I pulled every iteration we’ve made on client accounts in the last 12 months and ranked it by a ‘Change Score’.

The more different an iteration was from the original, the higher the Change Score.

Simply adding headline bars or applying a social wrap (like a Twitter layout) would score a 2 or a 3, whereas gutting out the entire ad and effectively changing the angle would score an 8 or 9.

I decided to measure performance by spend to make it easy to compare the original vs iterative concepts.

The Findings

After logging everything in a spreadsheet and feeding the data to ChatGPT, it returned a positive correlation of 0.42.

In simple terms, generally the more 'different’ an ad is to the original, the bigger the lift in performance - supporting the hypothesis.

The Analysis

This study isn't perfect by any means - and it’s obviously only results from our ad accounts, but it does provide some interesting insight.

This is a breakdown of the lift by change made.

As expected, the changes that require more time and effort yielded a bigger lift in performance.

The Conclusion

To conclude, I think the amount to which you iterate on an ad depends on the scale it's achieved.

For example, if an ad has already spent $500k, I'd be more inclined to make these smaller adjustments (as well as big ones) because squeezing an extra 10% of efficient spend would still be an extra $50k.

On the flip side, for an ad that's spent $25k, I'd almost exclusively make bigger changes as it needs a bigger relative lift in performance to achieve the same scale.

These numbers are relative and not absolute recommendations of course.

Let me know what you guys think of this.

I’d be happy to conduct more studies if this is useful to you!

Best,

Alex.