Editor’s note: John Fix is responsible for returning Procter & Gamble into the world of audio advertising. Since his retirement from P&G, John has opened his consultancy and Westwood One is excited to be one of his early clients. At P&G, John was one of the first to be dedicated to media analytics, from planning to attribution. He was responsible for the analysis and selection of media measurement applications and planning tools. At one point, John led U.S. media mix modeling, multi-touch attribution, and market testing. In his years at P&G, they went from non-existent to first in radio over a 5-year period.

Key findings:

• Response curves depict the impact of varying levels of media investment to sales response and ROI: Underspending generates weak ROI.
• ROI varies by investment level. One snapshot of ROI does not tell the whole story: As-run data of actual deliveries have significant variation.
• Response curves tell a richer ROI story of varying levels of investment: As impressions increase, reach increases and users are exposed to multiple impressions. Because reach builds as a curve and frequency is assumed to influence consumers, there is an expectation that sales will correspond.
• Media Mix Modeling can be used to optimize within the AM/FM radio campaign to achieve the best performance.

Over the last year, I have written Media Mix Modelling Does Not Hate Audio, It Just Needs More Granular Data, which described the importance of using weekly, as-run delivery data to correctly inform Media Mix Modeling.

Using planned media weight is problematic. It creates a “smoothing effect,” which makes it difficult to correlate AM/FM radio ads with sales. The upshot? Audio gets little credit for sales.

I also wrote Want to Get Mixed Modelling Right? Start with your Data, with SXM published by the Association of National Advertisers. The piece revealed the first-ever audio response curve from MMM firm In4mation Insights. Both articles stress the importance of accurate data and that MMM is a process that informs optimization.

Response curves depict the impact of varying levels of media investment to sales response and return on investment

Often MMM reports produce something called a “response curve,” which shows the sales response and ROI of a media channel across the range of impressions and GRP weight. Response curves are used to show how the ROI and sales shift as the investment level changes.

The In4mation Insights study revealed ROI could have improved with greater spend. Notice the reported streaming audio ROI is much lower than what greater spend would achieve for maximum ROI. This response curve clearly reveals spend level drives ROI and underspending generates weak ROI.

ROI varies by investment level: One snapshot of ROI does not tell the whole story

Audio media publishers report advertisers say, “We did an MMM and audio/radio/podcasts don’t work.” This usually signifies the end of the investment. In other words, “We did an MMM for a media channel using the data that we had for the average weekly impressions and the ROI that was below our ROI goal.”

The words that I will focus are “the result for the average impressions per week.” MMMs are models that associate sales over time with media. Models tend to examine 104 prior weeks. Ideally the data shows that there was week-to-week variance. This is made possible by using as-run audience delivery data instead of planned weekly weight.

As-run data of actual deliveries have significant variation. A weekly target of 50 gross rating points (GRPs) might produce as many as 80-100 weekly GRPs and as low as 20-40 GRPs. The MMM result likely represents the ROI at the average of 50 GRPs. This singular number does not tell the whole story.

The price of the advertiser’s product or service impacts the ROI. In the table below, a series of Nielsen AM/FM radio sales effect studies conducted by NCSolutions and Nielsen Buyer Insights from 2013 to 2019 reveals a wide range of ROIs.

CPG and food items to the left of the table have ROIs in the low single digits. To the right of the table, retailers with large baskets of purchases and service businesses with large average purchases have bigger ROIs.

Adopting An Optimization Mindset: Did this execution of AM/FM radio work

The discussion of response curves shows that the analysis of MMM results should be more than one snap shot of average response for the average spend. Using the optimization mindset can also move the discussion from, “Did AM/FM radio work?” towards, “Did this execution of AM/FM radio work?”

In the examples above, it can be seen that a different investment level could have yielded a different result. Given sufficient investment and granular data, MMM’s can also help to answer:

• Which format/programming genre performed the best?
• Which campaign (assuming multiple creative) performed the best?
• Which daypart mix performed the best?
• Which media weight level is optimal?

Given the assumption that AM/FM radio works, MMMs can be used to optimize within the AM/FM radio campaign to achieve the best performance. This can be used to inform the brief and allow AM/FM radio to become a set of tools to help a brand meet its media objective of reach and performance.

Key findings:

• Response curves depict the impact of varying levels of media investment to sales response and ROI: Underspending generates weak ROI.
• ROI varies by investment level. One snapshot of ROI does not tell the whole story: As-run data of actual deliveries have significant variation.
• Response curves tell a richer ROI story of varying levels of investment: As impressions increase, reach increases and users are exposed to multiple impressions. Because reach builds as a curve and frequency is assumed to influence consumers, there is an expectation that sales will correspond.
• Media Mix Modeling can be used to optimize within the AM/FM radio campaign to achieve the best performance.

John Fix can be reached at johnfixltd@gmail.com.