X

Oracle Data Cloud Blog

Don’t think about frequency until you’ve maximized reach

This week's blog is contributed by Trent Salazar, Data Scientist, Oracle Data Cloud.

Advertising is full of “magic” numbers. Does an ad need to be shown three, five, or even ten times to a consumer before achieving the desired effect?

This question of how many ads to serve per individual is generally called Effective Frequency, and for well over 100 years, marketers have been publishing treatises on the topic.

Back in 1885, British merchant Thomas Smith published “Successful Advertising” in which he proclaimed that the magic number was 20 impressions, and then proceeded to describe the thoughts evoked by every-single-one of the twenty impressions (My personal favorite is impression 18 – “The eighteenth time, they curse their poverty for not allowing them to buy this terrific product”).

Regardless of whether it’s 3 or 20, the conventional wisdom is that an individual needs to see an ad multiple times in order for it to have an impact.

The data, however, tells a more complex story.

Consider the Dose Response model

At Oracle Data Cloud, we have the benefit of measuring campaigns that reach well over 60 million households, including millions of households who only saw one impression, millions who saw 50 impressions, and the millions who saw every frequency in-between. This data allow us to fit a model that measures the relationship between impression frequency and Incremental Sales Lift.

In medicine, Dose Response models are routinely used to show how a population responds to various dosages of medicine. Our Dose Response model, a unique application developed by Oracle Data Cloud’s data science organization, shows how the exposed population responds (relative to the control group) to different impression frequencies.

Two theories of consumer behavior

The Dose Response model is a useful tool for understanding Effective Frequency. Conventional wisdom holds that several exposures to an ad are needed before the desired effect is achieved. This theory of consumer behavior would result in a Dose Response curve that looks like the following:

Theory 1

The chart on the left shows how cumulative incremental sales increases as impression frequency increases, while the chart on the right shows the incremental sales added by each additional impression. The “S” shape of the left-hand curve indicates that the first several impressions have very little impact, but once a “critical mass” is achieved (between impressions 10-15), incremental sales starts increasing rapidly.

An alternative theory of consumer behavior is depicted in the plots below. This version of the Dose Response curve shows that consumers begin responding to an ad as soon as they receive their first impression. Furthermore, the plot on the right shows that the first impression generates the most incremental revenue and each successive impression generates less and less until things completely flatten out around the 50th impression.

Theory 2

 

So, when does an ad become effective?

One of the features of our Dose Response model is that it does not require a convexity constraint. Sorry. That’s Data Speak for when the shape of the curve is solely dictated by the data, rather than any assumptions we apply to the model. The result is that some campaigns have Dose Response curves that align with Theory 1 while others have curves that align with Theory 2. However, a clear trend emerges when we look across 200 recent ROI studies.

Theory 3

In more than 80% of campaigns, the first impression served to each household was the most effective in driving incremental sales. Contrary to conventional wisdom, most individuals exposed during these campaigns did not need to see an ad multiple times before responding to it.

Does this mean I only need to serve each person 1 impression?

While the first impression is the most effective impression, that does not mean you should only serve one impression. Remember that subsequent impressions still generate positive incremental revenue, just not nearly as much as the first one. The implication of the steep drop-off after the first impression is that exposing a new household for the first time is better than serving another impression to someone that has already been exposed. Only after you have exhausted your reach among a specific audience should you worry about your frequency.

Stay up to date with all the latest in data-driven news by following @OracleDataCloud on Twitter and Facebook! Need data-related answers for your next marketing campaign or client partner? Contact The Data Hotline today. (What's The Data Hotline?) 

Photo: Monkey Business Images/Shutterstock

Be the first to comment

Comments ( 0 )
Please enter your name.Please provide a valid email address.Please enter a comment.CAPTCHA challenge response provided was incorrect. Please try again.Captcha