Most of us have heard the famous quote: "Half my advertising spend is wasted; the trouble is, I don't know which half" (there is debate regarding the author). Unfortunately, decades later, many entrepreneurs, even the most analytically minded ones, still do not understand the effectiveness and efficiency of their advertising spending. Whether you are a single-employee company or a larger one, MMM is still an essential part of a marketing plan.
What is marketing mix modeling (MMM)?
Marketing (or media) mix modeling is a statistical analysis technique used to help measure the impact of promotional spending and external factors on sales and revenue. Activities that do not always sit in the marketing plan, like sales calls and press releases, can also be measured. Even the elusive to measure activities like influencers and brand partnerships can be part of the model. The equations will help determine how various expenditures contribute to different marketing objectives, such as traffic or sales. For example, MMM can assess how increased marketing spend on magazine ads affects overall sales. Because many external variables are part of the model, you can learn if the effects were from your efforts (ad spend) or other factors (competitive declines or something like a global health crisis). Marketing mix models often use two to three years' worth of data to factor in items such as seasonality and organic growth.
The insights derived from marketing mix modeling can help you adapt your campaigns based on numerous factors. These insights help create an ideal campaign to drive engagement and sales. With each subsequent model, your advertising campaign can become more and more efficient.
The model assigns numerical values for the efficiency and effectiveness of each media buy. Meaning, the model will be able to say something like: Overall marketing efforts accounted for 20% of company growth over the last three years. Of this, 60% was from paid search marketing (SEM), 20% was from social spending, 10% was pre-roll, and 10% was from TV. Further, the modeler can provide cost per acquisition values for each channel or strategy. Data like this helps create customer lifetime value calculations. These findings can allow you to determine the ROI of each of your strategies, help you allocate future spending and assist in developing sales forecasts. In short, MMM models make marketing accountable.
Kraft first used MMM techniques in the 1960s and 1970s to assess optimal advertising placement for Jell-O. Kraft attempted to determine which parts of the country and which times of the year they should place their advertisements. They also wanted to look at ideal flighting times in each geographic region.
Many marketers did not start using MMM on a large scale until a few decades later. Then, many consultants and companies popped up offering MMM and MTA (multi-touch attribution models). With the proliferation of data and other technological advances, MTA has become a preferred approach, alone or in conjunction with MMM. MTA models use seemingly endless amounts of data to track users through the customer journey to establish where to spend money. However, the bottom-up approaches used in MTA models have significant disadvantages, the enormous cost being one of them. Since these two models are typically marketed together, many entrepreneurs feel that MMM is out of their reach.
Why is MMM advantageous to other approaches?
MMM is privacy friendly. MMM is a top-down approach. Meaning, it does not start at the bottom, granular, customer-level like MTA or other attribution models. With data breaches and scams in the news daily, your customers likely do not want to think about being tracked through the entire buying process. More importantly, Google Chrome will eliminate third-party cookies worldwide by 2022. Safari and Firefox have already eliminated third-party cookies. A cookie-less world will not affect MMM.
MMM can incorporate both offline and online (digital channels). As media has become more fragmented across more channels, MMM can be a tremendous asset for marketers. MTA models track clicks, so they must rely on click data and have weaknesses attributing offline data. Most companies today use a wide range of media tactics. It is essential to achieve the right reach and frequency of ads. A certain degree of layering different ad techniques from various media types is generally the most effective way to gain customers' attention and awareness. MMM is far superior to other approaches in terms of measuring different channels.
MMM allows marketers to factor in external influencers such as seasonality, promotions, brand changes, varying creatives and unprecedented external conditions. If the data exists, it can be incorporated into an MMM model. As we have recently found by the worldwide pandemic and resulting economic conditions, no one is immune to environmental influences. These factors must be included in the modeling, or the results will not be accurate. MTA models cannot address external factors that are hard to quantify, such as brand equity, word-of-mouth, pricing, seasonality and competitor activities.
MMM is also plainly more accessible and does not require extensive database integrations. You can run the model if you can hire the right internal or external resources to run the analyses. Other modeling techniques need data scientists to complete extensive data integrations.
What are the challenges of MMM, and how should you overcome them?
Media mix modeling is not without deficits. One would be how infrequent the results can be. Typically, companies get MMM reports every quarter or once a year because of the significant time needed to gather data and run the models. While this is a legitimate concern, I would argue that this is a disadvantage for MTA. MTA read-outs can be daily.
However, as a marketer, you cannot adjust your promotional strategies frequently due to contracts with the media vendors. Nor should you want to change your plan that frequently. Rapid shifts in media strategy can risk alienating customers. Also, rapid shifts in promotional spending may mean that customers do not achieve the correct frequency level needed for a response. Further, MMM keeps organizations aware of broad trends and patterns over many years, especially in the external environment. I have never been a fan of rapidly shifting strategies. Even with the massive changes that we have lived through recently, changes should be made slowly and strategically.
Another argument against MMM is the data granularity. With click data available via a tool like Google Analytics or MTA, executives can track a specific user's experience through the purchase journey. Therefore, MMM cannot focus on consumer experiences. However, for the entrepreneur with a limited budget (more on that later), an MMM coupled with a close eye on a Google Analytics account can fill in many gaps regarding granular data.
How to use MMM
No tool should be the sole technique to managing improvements in your marketing strategy. As with all measurement techniques, use MMM in conjunction with other techniques such as A/B testing, other forms of experiments and tests, surveys, creative testing and the different methods of click analyses. To be effective, combine MMM with additional tools to provide a unified marketing measurement. Specifically, a simplified sample measurement plan follows:
Continual monitoring of Google Analytics and Google Ad Words, or similar tool that uses tagged tracking.
Qualitative and quantitative new creative testing.
A/B testing for new geographies or products.
Surveys to assess brand health metrics.
MMM once a year.
By using MMM once a year, you can always have a top-down, holistic view of advertising's contribution over each season and many environmental conditions. A plan as discussed above will provide campaign insights regarding the efficiency and effectiveness of your spending. MMM will help you understand how various external factors (consumer trends, economic changes, health crises) and internal factors (PR, sales, creative) affect conversions. MMM will help you understand historical trends and the effects of seasonality. Brand positioning and creative messaging can be assessed, and their effectiveness can be directly measured.
An MMM model produces a 20-40% improvement in spending efficiency and a 10% increase in marketing effectiveness. A large MMM/MTA company will charge you anywhere from $500,000 to upwards of $2 million. Consultants may charge anywhere from $10,000-$100,000 for a one-time model. However, it would help if you considered the outlay of cash against the savings. Assume a 10% annual savings from a model. So, the maximum recommended budget for an MMM model on a $1 million yearly advertising budget would be $100,000 (with $25,000-50,000 being an even more conservative maximum).
You will likely want to continue having models run for you each year. The initial changes are likely to produce better results than subsequent changes. Because, hopefully, you are becoming more and more efficient with your spending. But you are likely to become more efficient at starting the modeling process and can reduce your consulting costs. Using an MMM and other advertising measurement techniques can help save significant money for a marketer with any promotional budget. It is one of those tools that most big companies know they need to have. Small entrepreneurs either think they cannot afford an MMM or are not aware of their value. This is simply not true. Marketing mix modeling is vital for companies of all sizes.
#marketingmixmodeling #MMM #mediamixmodeling #advertisingefficiency #advertisingeffectiveness #ROI #marketingmeasurement #dataanalysis #marketingstrategy #digitalmarketing #offlineadvertising #seasonality #externalfactors #costperacquisition #customerlifetimevalue #marketingaccountability #marketinginsights #advertisingmeasurement #A/Btesting #createtesting #surveys #brandhealth #marketingbudget #smallbusinessmarketing #entrepreneurship
Comments