A survey of over one thousand marketers asking what is the most important outcome from their marketing activities showed that incrementality was the most popular answer.
Incrementality in marketing stands for the results that are achieved by the marketing activities, results that would not have happened if it was not for the marketing activities.
Like the old saying: “if a tree falls in a forest and no one is around to hear it, does it make a sound?”
Measuring the incrementality of marketing has been a real challenge. Over the past years, the most popular methods to measure the incrementality in marketing have been randomized control test, better known as a holdout group, or creating a blackout and comparing marketing results before, during and after a blackout.
A randomized control group meant that Marketers would attempt to create a holdout group, either by region, city, or by creating an audience targeting list (applicable only to digital media), activating marketing campaigns in one segment, while restricting those in another. This method has been sufficient in providing some sort of an understanding of the incrementality of the overall marketing activities, however, when attempting to measure the incrementality of a specific channel, such as Facebook vs. Google, or Radio vs. TV – a randomized control group test was not able to provide an understanding of incrementality between channels.
Customers “leak” from segment to segment. The users using Google, are most likely the same users using Facebook. The same went for TV vs. Radio, or practically – any channel vs. another channel.
The second, more absurd method of incrementality measurement has been to run a blackout experiment. A blackout, as it sounds, meant that a marketer would completely turn off all advertising activities in a country, to learn how marketing results look like with no advertising in place, only to activate marketing campaigns – channel by channel, campaign by campaign, to get a perspective over the activities that yield positive results.
The opportunity cost in performing a blackout test are enormous, and more importantly – do not take into consideration any seasonality component, or external influencers such as competition.
To put it simple – December is not January, and while performing a blackout, a marketer has no ability to control the weather.
A new way to measure incrementality in marketing has recently emerged. With innovation and advancements in machine learning and algorithmics understanding – marketers can now apply advanced causal data science models, allowing software to create a retrospective prediction providing a clear result to the question: What would have happened if we did not make a change ?
As marketers make changes frequently, starting and pausing campaigns, making changes to bids and budgets, testing new channels and making creative changes – understanding incrementality by using causal data science can provide an applicable use-case without marketers needing to create either holdout groups, or perform any additional experiments such as blackouts.