Creative Testing’s Output is Learning

. 2 min read

By Tom Murray | Managing Director

When I first started out as a media buyer at large agencies 8 years ago, I am ashamed to say that creative was never really at the forefront of any decisions we made on the buying side. Everything was about the metrics - CPCs, CTRs, and barely even anything about conversions since pixels were pretty new back then. As time has shifted, media buying has become more automated and the main point of differentiation between the best marketers is the creative.

I’ve always said that it doesn’t matter if you are giving away $1,000,000 on a landing page if your ad isn’t good enough to warrant a click, or even grab the user’s attention.

In a world where there are more advertisers joining the auction each day, creative testing is one of the remaining levers media buyers and brands have to lower their costs. If CPMs are going up 20% each year, if you can make better creative to increase your conversion rate by 30%, you’ve mitigated the cost increases of the auction competition (and even came out ahead). Almost all resources should likely be geared towards creative development and experimentation at this point.

However, often times creative testing is not structured in the right way and can be set up for failure before you even begin. So how should you approach it?

Let’s take a look at some examples of things completely unrelated but also quite relevant to the discussion.

Did you know that some fish can lay almost a million eggs, but that nearly 99% can die within days of hatching? They lay that many to ensure they have enough that survive. The same thing can be said of venture capitalist firms, who invest heavily into so many companies, knowing that over half of new businesses typically fail within five years.

How does this relate to creative testing? Well, if you don’t test tens, hundreds, or even thousands of creatives, how will you find that diamond in the rough winner? Softbank invested heavily into Uber and lost, but also tripled their money with their investment into Lemonade. Had they have only invested in 1 company, the risk would be much higher and they might have chose wrong. Creative testing and experimentation allows you to diversify your risk, just like you would your stock portfolio. And you can learn some insightful things along the way about what creatives work, and what your audience responds to.

If you take away one sentence from this entire post, it would be this: Creative Experimentation eliminates performance risk because the output is learning.

Edison failed time and time again to come up with the lightbulb. Had he have stopped on try #500 instead of making it past #1000 we might not have the bulbs we have today. Same can be said for creative testing.

Reduce human bias by removing the creative ego that goes into making ads. Everyone thinks they can judge a good ad from a bad one, but why not run it to find out with true actionable data? Proper creative testing can help you find your winning ads, but you won't know until you try them out. Never assume that an ad will perform a certain way until you try it. You just might be surprised at what you find!