Using Media Mix Modeling to Boost Marketing Efficiency
A recent Gartner report indicates that adding four or more channels to an integrated campaign could increase results by 300%. Marketers today use a variety of media channels to drive branding and revenue KPIs. Media mix modeling (MMM), also known as marketing mix modeling (MMM again!), is an analytic technique that allows marketers to measure the impact of multi-channel marketing and advertising campaigns. MMM also powers simulations to aid planning for future campaigns.
What is media mix modeling?
With the growth of marketing channels and media divides, media mix modeling allows marketers to determine how various advertising efforts, including offline channels, contribute to reaching their goals. Those goals could include, brand awareness, consideration, conversion, or sales. MMM also determines when marketing spend in a specific channel has achieved saturation, which means that additional spend in that channel no longer increases returns. Additionally, MMM can capture the lingering effects of marketing efforts (e.g., brand equity) that persist long-term even if the activities themselves were to stop.
With this information, marketers can refine or optimize their marketing or media mix to drive more efficient outcomes. That’s true whether their goal is website conversions, sales, or another KPI. In addition, MMM helps marketers create simulations of budget allocations for the future that will help achieve those critical goals.
How is media mix modeling used?
Traditional media mix modeling uses regression analysis to determine relationships between outcomes, like sales or engagement, with ad spend, reach, or impressions delivered across the various channels used in an advertising or marketing campaign.
According to a recent post by Digiday, media mix modeling offers an “accounting for digital, TV, out-of-home, radio, podcast and social media advertising but the price of a product and various promotions that are being run [as well as] … inventory levels, seasonality, even shifting weather patterns — basically anything and everything that could impact sales.” This data is then compared to relevant outcomes, such as sales data.
There are three key components marketer needs to leverage media mix modeling:
- A breakdown of marketing or media channels used in a campaign
- A breakdown of budget, impressions, or reach per channel
- Campaign results or goals
What are a few examples of media mix modeling?
Media mix modeling was popular in the 1960s when print, radio, and TV dominated the marketing landscape. At that time, brands — especially CPG companies — were early adopters of the technique. They hoped to measure the impact of their marketing mix on diverging media channels. But, in that era, this modeling approach was slow, inflexible, and infrequently updated.
Today, with the greater limitations placed on online tracking, many brands are re-evaluating media mix modeling to obtain:
- Visibility into marketing performance in comparison to media spend
- Guidance into budget optimizations and ad spend
- Which media channels to optimize or spend more on in future campaigns
State-of-the-art MMM can now be much more agile and easily updated to provide current recommendations. Marketers can stay on top of customer behavior and channel trends to refine marketing tactics.
How can predictive analytics support media mix modeling?
Media mix modeling can help allocate budgets among channels and predict outcomes for the unique marketing mix of any size of organization. Predictive analytics supports media mix modeling by using historical data to provide future values of a marketing program down to the channel level. It also includes simulation features to accurately predict the likely outcomes of specific budget strategies.
A brand using predictive modeling for MMM can optimize marketing budgets for an upcoming period. They can anticipate the likely contribution of each channel to their goals. They might also move resources from saturated channels to other channels proving to be effective. MMM can also incorporate non-digital channels for which detailed data may not be available. That capability makes it a comprehensive complement to a marketing measurement toolbox alongside attribution models and experimentation approaches.
An MMM success story
Using MMM through Pecan’s platform, one game maker saw impressive results after implementing an MMM predictive model built in under 3 weeks. The model predicted day 7 revenue from new installs of the game on both network and channel levels. This model allowed the marketing team to determine how much each channel contributed to their revenue. They could also determine the likely return on ad spend for each channel.
Pecan’s highly accurate models helped this team identify underutilized – but very effective – marketing channels. They could also identify channels at or nearing saturation. These results helped the team adjust their per-channel budget to maximize their marketing efficiency.
As a whole, MMM has excellent potential for improving marketing efficiency and execution, maximizing return on marketing spend, and making the most of available resources.