February 10, 2014 Last Updated 1:32 pm

Digitally Marketing: Better conversion rates do not mean more sales

The mistake I see made by consumer marketers is that they don’t really know how to judge the results of their online ad efforts.

Here’s a test – you have two different creative offers you are testing and each is tied to a different landing page. Your average conversion rate of visitors to paid subscription is 0.92% across all tests you’re running. Your staff show you the results in a well-made powerpoint.

020514-1Lunch is served. Everyone agrees that Page B is Great! And, that all of the tests shown in the chart are better than average. You make the decision to replace your default landing page and creative with Page B and then go out to celebrate the fact that your bonus this year is going to be a big one.

Oops. The probabilities are that you just made a mistake because you didn’t know about Standard Error.

You see, those conversion rates you see are “estimates” of the actual conversion rates you could have gotten if you were able to test all ad impressions in the world. Your estimate is not quite the exact conversion rate. How much is it off?

Well, that’s the question you should have been asking in your meeting. The answer is related to averages and standard deviations. Now, math may make your head hurt but so will the concrete when your head hits it as you’re thrown out on the street for making Page B the company’s new landing page and default creative.

Here is the table you should have been looking at:


The column labeled “SE” is the Standard Error. It shows you the accuracy of the estimates you’ve made of the real conversion rates. Clearly, the size of the error is related to the number of conversions and clicks you were able to measure. Seems obvious but I wonder if you were thinking about it when you were eating your soup during the meeting.

The last column is the probability that the conversion rate you got in your test is, in fact, the same as average. Page A’s conversion rate has 0% probability of being the same as the average conversion rate of 0.92%. Page B? The odds are in your favor – it’s conversion rate is 12% likely to be the same as average. But, is it a better bet than Page A? Which decision would you make?

In either case, the decision you make will be better informed because you considered the Standard Error of your estimates.

Troy McConnell is CEO at AudienceFUEL.

Comments are closed.