It is clear to all that business performance is continually impacted by marketing activities (controllable) and marketplace activities (little to no control). But it is notoriously difficult to quantitatively measure and evaluate the impact of such marketing activities. Additionally, it is very difficult to differentiate clearly between the marketing actions that are working and those that are not; making it difficult to learn and build a better marketing strategy. â€œI waste half of the money I spend on advertising. I just donâ€™t know which halfâ€ is a famous quote first attributed to John Wanamaker a US store merchant in 1862. The sad news is that we hear it all too often even today.
Marketing mix modeling, or MMM as it is more commonly called, is a ubiquitous term today. It is a domain in Marketing Analytics that uses predictive modeling to measure the effect of different marketing factors (like price, distribution, advertising, consumer and trade promotions, and new product/variant introductions) by estimating the incremental sales generated due to each individual marketing factor. Using this, companies can measure the impact of different marketing factors in terms of incremental sales generated and the Return on Investment (ROI) which is simply the incremental sales generated per rupee spent on that marketing vehicle. For example you can find out if spending 100 rupees more on TV ads is giving you more or less sales than the same 100 rupees spend on Digital ads. In addition to measuring the effectiveness the same models can be used to optimize overall marketing spends across different marketing vehicles.
In addition to these 2 main questions (i.e. measuring impact of different marketing factor and optimizing overall marketing spends) the same models can also be used to understand other related business questions.
One can simulate impact of changes in the marketing strategy and estimate impact on sales. One can develop different marketing strategies to see how to meet a future growth target.
One can estimate how competitionâ€™s growing presence in store or on television airwaves will impact their own brandâ€™s sales.
Understanding how marketing spend for one brand/variant in the portfolio affects the sales of other brands in the portfolio. How much is the cannibalization or halo?
Estimate price sensitivity of the market. Figure out the optimal frequency and depth of price discounts
Deep dive into media planning. Optimize flighting patterns of TV ads; determine how much GRPs to gain on prime time vs. non-prime time; determine when to use longer duration ads vs. short duration ones
While the concept of marketing mix has been around from the â€˜50s, the market mix modeling fever did not catch on in earnest until the 1980s. It started with the CPG companies (like P&G, Kraft, Coca-cola, Pepsi, etc) trying to make as much of their already slim margins. Over the years Retail, Financial services, Insurance, Telecommunication, Entertainment and pharmaceutical companies have all boarded the wagon in the quest to optimize their marketing spends.
The Marketing Mix Model if done well and its recommendations actually used in the annual marketing planning process, companies can get anything between 15 per cent to 30 per cent additional sales using the same marketing budget. This is a huge impact considering the marketing budgets of companies today. This is why most Fortune 500 companies have made Marketing Mix Modeling a compulsory input to their annual planning process for every brand, every country.
While the term Marketing mix modeling is widely used and applied indiscriminately to evaluate different components of marketing strategy, the most common analytical approach is the multivariate regression model. These models are based on a number of inputs and how these relate to an outcome such as sales or profits or both. Once the model is built and validated, the input variables such as advertising, promotion, etc. can be manipulated to determine the net effect on a companyâ€™s sales or profits.
However, in the last 5 years, several variations of the same have evolved, like non-linear models to capture the diminishing return of spends in a single marketing vehicle, structural equation models to capture the long term effects of marketing on overall Brand Equity and in turn on long term sales, Attribution Models to accurately estimate the elusive impact of the digital and social media marketing efforts.
Most companies today are capturing marketing data and also doing some form of analytics. Some companies have more data than others while others are more analytically advanced. But irrespective of the quality of data and level of sophistication, starting and regularly building marketing mix models can be the key to effective and profitable marketing spends. Companies can expect to unlock 15-30 per cent more value from their marketing budget, incentive enough to include Marketing Mix Models in the Analytics portfolio.