September 20, 2019

The 80/20 Principle of Ad Testing

By Rob Drummond with Jonathan Wilson
Estimated reading time: 4 minutes
Download a print-friendly PDF

Most people miss how profound and pervasive the 80/20 principle is.

We regularly speak to companies who are spending tens of thousands of dollars a month on ads, who don’t have sensible conversion tracking in place. They intuitively have a feel for the results being generated, in that they know how much the phone is ringing. But they don’t know which ads generated those results.

If you spend $100,000 on ads, and $1,000,000 rolls into your credit card processing account, then clearly something is working. But you’re very likely also leaving money on the table.

I (Jonathan) spoke to someone recently who contacted me to help setup his tracking. He was spending almost $400,000 per month on 12 display network banner ads. These ads had been running for a long time, and they knew for sure they were working, because there was no SEO traffic, and no Facebook traffic. They had a funnel, and 12 banner ads pointing towards that funnel.

But get this... He had no idea which of the 12 ads were doing any good. His return on ad spend was about 200% - he would spend a dollar and get two back in revenue.

One of his ads WILL have been more effective than the rest – and probably much more effective. Some of the 12 will have cost him money, and been a complete waste. This comes down to the 80/20 principle.

The 80/20 principle is the idea that a small proportion of inputs generate the majority of outputs. 20% of highways generate 80% of congestion. 20% of cities contain 80% of the world’s population. Those numbers may not be exactly 80/20; it may be 90% and 10%, or 99% and 1%. But the point is it isn’t 50/50, and this also holds true for your ads.

The top 20% of ads generate 80% of your revenue, while the bottom 20%waste 80% of your budget. One fifth of anything you do or try will generate your best results. If you can identify which ads those are, you can quickly improve your results.

If you have ten ads running, you can go into Google Ads and click on any column heading to sort from high to low, and the 80/20 principle will come to life. A small number of ads will generate the best click through rate. A small number of ads will generate the highest conversion rate. A smaller number still will generate the highest amount of overall profit. (You may need to test 100 different ads to find those, rather than 10 ads). But finding them can generate a big uplift in results.

In any aggregate number there is always a hidden 80/20 curve you cannot see. For example if you look at the overall conversion rate of a remarketing campaign, the aggregate figure hides the fact that those results came from a small proportion of the inputs. There’s actually an 80/20 curve nested inside the 80/20 curve, because 80/20 is fractal. If you have 100 campaigns in your Google Ads account, 20 of those will generate your best results. Within your most successful campaign, 20% of your ad groups will generate most of your profit. Within your most successful ad group, one ad or keyword will generate the most profit. Which is why if you only have 2 or 3 ads running in an ad group, you simply don’t have enough ads running to find a high performing ad. You likely haven’t tested a wide enough net of ad creative. If you only run 1 or 2 ads, you’re presuming you know what will work best, when in reality you don’t.

80/20 works in both directions, both in profit and cost. Going back to Jonathan’s client spending $400,000 per month on banner ads, they were likely buying ALL the ad inventory available to them. He couldn’t just increase his spend to improve his results – he was already maxed out. But statistically the likelihood is that of his 12 banner ads, only 2.4 (20% of 12) were doing him any good. The bottom 2.4 were costing him huge amounts of money, and he had no idea which ads those were.

The language of 80/20 has entered the marketing lexicon recently. Perry Marshall has an excellent book on the 80/20 principle. Richard Koch has a number of books. But most people miss that within your ads account 80/20 might actually be more like 99/1. Most people under-estimate how many ads they need to test to produce that winning ad right at the top of the spike. And it’s a challenge. If it were easy, or if there was a proven ad template or hack, everyone would already be using it.

The starting point is to understand and accept that the 80/20 principle is everywhere, and is pervasive.

You then need to test and kill a lot of ads. You need to find the ads that are only breaking even, because they aren’t worth running. And you need to identify the ads that are costing money. The only way to do that is to test a large number of ads, and remarketing lets you do that affordably.

A side effect of building a maze remarketing system is you end up running ads on more networks. You end up with more campaigns and ads, which amplifies the impact of the 80/20 principle. 

Rob Drummond

Rob is a co-host of the Maze Marketing Podcast, and specialises in content creation, ad copy and email storytelling. He has two published books, Magnetic Expertise and Simple Story Selling, affordable on Amazon.

Click Here to Leave a Comment Below

Leave a Reply: