Introduction from Andrew Joyce | Senior Vice President, BI+Analytics
MeritDirect | email@example.com
The saying “an average describes everyone…and no one…” may be a tired trope but it is one worth remembering. As marketers, we find ourselves returning again and again to these wise words. Consider: the average of 5, 6, 20, 21 and 28 is 16. Check our math—it’s true. But does anyone really think that 16 accurately describes any of the five numbers in this set? This is an admittedly basic example, but it is useful in illustrating the concept that averages can be deceptive.
In another context, anyone who has ever read political analysis has likely felt that the averaging of polls and cross-tabbing their results by demographic information can miss the mark relative to their own personal views—we are all unique after all! Yet every election cycle, political campaigns are spending more and more on data, analytics and polling. Averaging works…until it doesn’t.
This use of averages can be easily “misused” in any channel or tactic. We have put together a few scenarios that show where we are often mislead by the numbers. In this series you will hear stories from MeritDirect’s BI & Analytics group of learning (or re-learning!) this important statistics lesson in the context of direct marketing.
If you’ve found your way to a MeritDirect blog posting, it is a safe bet that you are familiar with marketing metrics related to new customer acquisition. That means you understand getting 3 out of 100 people to buy from your prospecting campaign whether it be a postal, email, or any other channel, is in most cases considered a success. This metric is most commonly referred to as a Response Rate (RR).
Run enough campaigns where 3 out of 100 people respond and that 3% response rate figure will become engrained as the benchmark for acquisition in your company. Unfortunately, we see all too many companies over-focus on this average response rate figure and not understand that any program producing 3% likely contains segments that perform at 5% as well as 1%.
So, what happens when company executives want to improve the response rate? This is when the segments that perform below average get eliminated. This will, of course, provide short-term efficiency improvements but ultimately will limit the scale of campaign results for your business.
How? All data sets contain values that are both above below the data set’s average. When evaluating postal campaigns, removing those below average segments will save a little bit of ad cost, improving response rates. But the decision to cut below average segments (without replacing them with more effective segments) will inevitably reduce the overall number of new customers acquired.
In saying “No, thank you” to segments that produce a 2% response rate, we also lose any new acquisitions that come along with that. It is important to never lose sight of that. Benchmarking segments to overall campaign averages can lead you down a path where metrics such as response rate begin to look better, but the number of new customers acquired overall will decline, and that is a scenario that no business wants.
Director | BI + Analytics
Catch part two from Maria Jaime, Senior Marketing Analyst: The Deception of Averages – In PPC Campaigns