This week’s guest blog post is contributed by Michael Anderson, Data Scientist, Strategic Analytics, Oracle Data Cloud.
So, let’s say after three months of running an ad campaign it finally comes to a close.
The numbers are in, the ads stopped running, and your targeted audience is no longer seeing your brand as part of that campaign.
Logic would say that an unfortunate side effect of a campaign ending is that revenue for the campaign ends, too.
Well, our data science team is here to show you how the revenue generated by an advertising campaign doesn’t end with that campaign.
In fact, 53 percent of a typical campaign’s value is derived from additional consumer spending up to 12 months after a digital campaign ends.
How can this be? We’ll walk through our research, which includes some stats that might surprise even the most experienced marketer.
In the January 1978 issue of the Harvard Business Review, Nariman K. Dhalla paints advertising as an inherently long-term investment – marketers have known this forever.
Dhalla writes, “Sales revenue is not generated immediately in a lump sum”—rather, it “flows like a stream over time.”
With that in mind, let’s explore why data is the key to unlocking the potential of that statement.
In marketing terms, we define “lift” as the increase in sales in response to an ad campaign.
An ad campaign generates lift in two ways:
Penetration lift during a campaign is derived from new buyers or existing buyers purchasing “out of cycle” – that is, after the campaign has ended.
Often, sales lift is used as a primary key performance indicator (KPI), but this view alone can underestimate the full effect of penetration lift on sales.
Additional value is accrued when the change in behavior driven by the ad continues into the future, long after the campaign has run its course.
Any additional increase in sales during months 2-12 after that campaign ends is referred to as “long-term value.”
Long-term value (LTV) contributes significantly to the value of a campaign. Because LTV is defined by continuous consumer spending over time and is independent of additional ad spend, it can have a substantial impact on Return on Ad Spend (ROAS).
It’s true that brand loyalty decays exponentially over time – that’s a straightforward concept. Here’s what I mean:
For an advertising example, let’s say the rate of decay is 0.7. This means that 70 percent of the buyers from an ad campaign will buy the advertised product again in the future, then 70 percent of those two-time buyers will buy for a third time, and so on.
This process continues until the effect of repeat buyers is negligible (in this example, after 12 months, only 1.3 percent of buyers will purchase again).
To model this behavior, Oracle Data Cloud uses a Markov chain on past-purchase data to estimate two decay curves:
Long-term value is calculated as the difference between those two estimated curves. (Get more in-depth with Markov chains in the below video).
In our study, we selected 107 campaigns since the implementation of long-term value measurement in early 2017 and compared short-term vs. long-term value estimates.
We discovered these three key results:
I hope this helps you to better understand the value of your campaign, even after it ends.
Want learn more about how your ads can drive long-term revenue?
About Michael Anderson
Michael is a data scientist at Oracle Data Cloud, focusing on using data to create actionable insights and best practices for the ad-tech industry.
Michael earned his Bachelor of Science in Statistics from the University of Denver.