Check your digital media for Adstock effects

When we’re doing marketing analytics work, we’re often asked by clients how much they can spend in a certain media.  The implied second half of the question is “…before my ROI tanks.”  It’s a reasonable question, and one that often opens the door to some analytics as well as starting good strategic discussion around the goals of marketing (both financial and non-financial).    Most of the time we’re measuring both offline (TV, Print, Radio, etc.) and online (search, banner, social) tactics.   One key input to the measurement is Adstock.

What is adstock? In traditional media like TV, Radio, and Magazine are subject to two effects: saturation and decay.  Saturation is the idea that twice the media doesn’t always give you twice the impact, because of diminishing marginal returns.  The theory is that the as you continue to advertise more and more, you either hit the same people more and more, or you hit people less attractive or likely to respond.  Decay is that your marketing has a lag effect that decays after the actual execution.

Once these factors are measured, they can help advertisers understand how much to spend by showing where diminishing returns happen.  But few advertisers think this way when it comes to digital advertising tactics, because they rely on the instant response metrics that emerge from simple conversion tracking.  But with a fairly simple analysis using no more than excel, digital adstock can be measured and then incorporated into budgeting and allocation decisions.

If you have a reasonably robust history of your marketing spend, here’s a quick test you can use to see if the adstock effects are emerging from your digital execution.

First, gather your weekly spend and conversion data, and plot weekly spend against conversion rate on a scatterplot.  Sometimes, unfortunately, you’ll see a scatter that looks like this, which is referred to in this business as “a sneezer.”

The implication from a sneezer is that there are either too many other factors interfering with the data, like site redesigns, large price shifts, competitive activity, or that you are not yet anywhere near the point where diminishing marginal returns is impacting response.

However, sometimes we see relationships like this emerge:

This cuts to the chase, and gives you a direct measurement of digital adstock so that you can calculate an ROI and set spend levels exactly where you want them.  Of course, now comes the hard part: what is the “right” ROI for marketing?  Is it only $1.00 payback?  Or are there other factors beyond near term sales lift that are worth investing in?