Have you given much thought to your last purchase? Perhaps you ordered a new computer online. Chances are that your interest in the purchase was sparked by an ad and preceded by some research. After seeing an item on TV, you might have started with a simple Google search, browsed through some sites, and clicked on a few ads before you made your final decision. We’ve all gone through the compare and contrast dance when it comes to purchases, but I don’t think many of us realize just how much work goes into measuring the results and success of ads.
CMOs have a tough job when it comes to digital advertising. How exactly can something as abstract as the patterns people follow when making a purchase be measured? Well, while the solution is not an exact science, media mix modeling and conversion attribution can offer great insight. Just think of the possibilities that come from knowing exactly what type of ad works best for your customer.
In media mix modeling, statistics are used to allocate advertising resources for TV, internet, print, and radio ads. Conversion attribution is a further breakdown of the media mix and applies to individual ads. For example, a keyword search on a website is a conversion attribution. Different statistics and different amounts of data are used for the two. The idea of how the two work is simple enough, but how is the success of each method measured?
Media mix modeling measures the ratio between the amount spent versus the revenue generated on a specific channel. The amount of revenue different channels generate varies. But, what happens when several channels come into play? Say that you see an ad on TV that sparks your interest in a certain product and you decide to do some research about it online. Advanced media mix models need to be used to analyze the interaction between channels. For example, statistics show that TV ads and display ads can greatly increase internet searches for a specific brand or product. While a lot can be learned from media mix modeling, conversion attribution approaches data from the individual perspective.
Let’s go back to the example of your online purchase of a computer and look at it through conversion attribution analysis. You looked at computer options through search engines and various ads. Let’s say that you looked at 8 different ads before you made your decision. After you finally make a purchase, how much credit should be given to each one of those ads? Did they all influence you equally?
There are several different approaches for looking at the attribution problem. The easiest way is to simply give full credit to the first ad/click or to the last ad/click. This is the most common way of measuring ad success today. Another way is to split credit between the ads. Since you looked at 8 different computer ads, all 8 might get the same credit. Maybe one of the ads will get more credit. And if you received an email for a computer sale, that email might get part of the credit as well. And finally, ad success can be measured against no ad instead of against other ads. Would you have still bought the computer if there were no ads for it?
Of course, if rules for collecting information exist, there must be rules for properly evaluating the results. Marketers are able to draw a lot of insight about customer behavior from media mix modeling and conversion attribution. All that information can be broken down and applied to different industries. So what does it all mean in the end? What can successful advertising do for you?
The best thing you can do is make sure that data is collected in the most effective way for your business. Can the results from your ad campaigns be compared easily? How do you track various visitors? Do you have strong customer definitions in place? Think about how successful your marketing organization can be if a structured media mix is applied, and managed over time.
What is your attribution strategy? Do you consider the first touch more important, or the last?