What is marketing mix modelling?

Marketing mix modelling(MMM) is an analytical approach to optimize marketing efforts. It quantifies the impact of various marketing inputs,such as product features, pricing strategies, distribution channels,promotional activities, competition activity, and environmental factors, to maximize return on marketing investment ( ROMI ). It helps marketers allocate marketing resources effectively and efficiently, to achieve their goals.

‍

How does MMM help the marketer?

1. Market Mix Models help the marketer in many ways

2. It quantifies contribution of each marketing input to the sales that is observed

3. ROI calculations inform the marketer of the return she gets for every rupee invest

4. Response curves derived from marketing mix models study thresholds and saturation levels of input

5. Deep dives into the MMM discover how more granular levels of inputs work i.e. creative, duration,channel mix etc.

‍

What is the data that is required for MMM?

Sales :

‍ This database captures the sales of a product in volume, value or units, sold at a daily, weekly or monthly basis, in chronological order. Marketing mix models help in attributing these observed sales to various marketing inputs and other environmental factors.

Placement:

Involves the distribution channels through which products or services are made available to customers. It includes decisions about where and how products should be sold, in order to reach target markets effectively.

Price:

Refers to the cost customers pay for the product or service and measure price elasticities. Pricing strategies aim to discover a price that is a balance between maximizing profits and ensuring customer satisfaction.

Promotion:

These include sales promotional schemes like price discounts, bundle offers, free gifts for purchase and free samples that entice the customer to make an immediate purchase. Schemes could also be offered to the trade (called trade promotions) where retailers and encourage to sell the particular brand due to an incentive.

Advertising:

Encompasses advertising in various media channels aimed at increasing awareness, generating interest in the product, and persuading customers to make the purchase. These channels includeTV, Print, Out of Home, Radio, Cinema and In-store advertising, Digital channels include website, search, social, email and mobile. Marketers also conduct market events, trade fairs, public relations and spend on sponsorship.

Environmental factors:

Sales of a product are affected by seasonality which could be climatic, festival related and weddings and also because of academic and fiscal year endings. Moreover competition activity like promotions and advertising  could have an effect on sales.

Brand Equity:

Finally sales can also occur because customers tend to buy products they are familiar with and have developed a liking for it. This, habitual and repeated purchases are due to the equity of the brand. Customers feel that the brand fulfils a specific and unique need of theirs, in a cost effective way. However Brand Equity is not usually measured – it is a latent variable in the model that needs to be estimated.

‍

What are the techniques used in Market Mix Modelling

‍

Data Gathering: MMM starts with collection of historical data on your sales, marketing activities, and external factors like seasonality and economic trends. This could be for a period of 2 to 5 years on daily,weekly or monthly granularity.

Modeling : Using statistical techniques (regression analysis is a popular one), a mathematical model is built, that considers all these ingredients.  A mathematical model is a set of equations that mimic what is happening in the market place. This model attempts to quantify the unique effect of each of the factors, on sales. This quantification is also sometimes referred to as attribution. Some regressions often used are OLS and Bayesian

Insights & Predictions:The marketer translates these complex equations into easy-to-understand insights. They will infer which activities drive the most sales, which actions don’t deliver, how long impacts last, when they stop working and even how they interact with each other. Finally they will use the model to optimise a given marketing budget most efficiently, and forecast future sales.

Indeed it is said that MMM delivers the highest bang for the marketing buck.

In the next post we will discuss how MMM is different from other methods of marketing measurement like Multi Touch Attribution and Test& Control studies.

‍

‍

‍

‍

‍

‍