The common terms and definitions used in Marketing Mix Modeling
March 16, 2023 | Carmen Bozga, Chris Kervinen
Marketing Mix Modeling (MMM) is a method used to measure the effect of marketing efforts on sales and performance metrics. It's gaining popularity among marketers as they aim to maximize their marketing budgets and comprehend the ROI of various channels and strategies. Plus, with the decline of third-party cookies, MMM has emerged as an ideal tool for marketers. However, it can be tough for beginners to grasp as it involves several technical terms and definitions. This blog post aims to shed light on the most common terms used in MMM and what they mean.
Let’s dive into it:
Marketing Mix Modeling (MMM)
Marketing Mix Modeling (MMM) is a statistical analysis technique that helps businesses measure and optimize the impact of their marketing efforts on sales. It involves analyzing the different components of a marketing campaign, known as the marketing mix, and their impact on sales performance. MMM considers both base sales and incremental sales when analyzing the impact of marketing efforts. Another important metric of MMM is Return on Marketing Investment (ROMI).
Base Sales refers to the level of sales that can be attributed to factors other than marketing activities. These factors may include seasonality, economic conditions, and other external factors that can influence sales. Base sales are usually estimated by analyzing historical sales data and identifying patterns or trends that are unrelated to marketing variables. By separating the effect of base sales from the impact of marketing activities, MMM models can more accurately measure the incremental effect of marketing on sales. This allows marketers to better understand the ROI of their marketing investments and make more informed decisions about future marketing strategies.
Incremental Sales refers to the change in sales that can be attributed to a specific marketing variable or activity. Incremental sales are the additional sales that are generated by a particular marketing initiative or campaign, over and above the base sales.
The math is simple: Total Sales = Base Sales + Incremental Sales
ROMI stands for "Return on Marketing Investment" and is a key performance indicator (KPI) used in Marketing Mix Modeling (MMM) to measure the effectiveness and profitability of a company's marketing campaigns.
ROMI is calculated by dividing the incremental revenue generated by the marketing campaign by the total cost of the campaign, including all marketing expenses. ROMI is a useful metric for evaluating the profitability of marketing campaigns because it measures the return on investment (ROI) generated by each dollar/euro spent on marketing. A positive ROMI indicates that the marketing campaign is profitable and generating a positive return on investment, while a negative ROMI indicates that the campaign is not generating enough incremental revenue to cover its costs.
Data refers to the information that is collected and analyzed to understand the impact of various marketing efforts on sales performance. This data can include a wide range of metrics and variables, such as:
- Sales data: Information on the number of products sold, the revenue generated from those sales, and other key performance indicators (KPIs) related to sales performance.
- Marketing data: Information on the different marketing efforts that were implemented during the period being analyzed, such as advertising spend, promotional activities, pricing changes, and other marketing initiatives.
- External data: Data on external factors that may have impacted sales performance, such as economic indicators, weather patterns, and changes in consumer behavior or preferences.
- Other relevant data: This may include data on market trends, competitor activity, and other factors that may have influenced sales performance during the period being analyzed.
Input and Output
Input essentially means “the data that is used as the basis for the analysis”. This data usually includes information on sales, marketing spend, and other relevant factors that may impact sales performance, such as external market trends and economic indicators (please see the previous definition).
The input data is used to develop a statistical model that identifies the relationships between the different inputs and sales performance. The model is then used to make predictions about how changes in marketing spend or other inputs may impact future sales performance.
Output represents the results of the analysis. They are the insights into the effectiveness of different marketing efforts and recommendations for optimizing the marketing mix to improve sales performance. The output may also include predictions of future sales performance based on different scenarios or changes to the marketing mix.
A data connector is a software tool or application that is used to connect and transfer data between different systems or databases. In the context of MMM, data connectors are used to connect data sources such as sales data, marketing spend data, and external data sources to the modeling tool.
Scenario planning is a strategic planning tool used in Marketing Mix Modeling (MMM) to help businesses anticipate and plan for different future scenarios. It involves creating multiple hypothetical scenarios based on different assumptions about the future such as incremental sales.
API stands for Application Programming Interface. An API is a set of protocols, tools, and standards for building software applications, allowing different software systems to communicate with each other and exchange data.
In the context of MMM, an API can be used to connect different software applications or tools used in the modeling process, such as data management tools, statistical modeling software, or visualization tools. This allows for a more efficient and streamlined modeling process.
What drives base sales?
Seasonality is a regular pattern of fluctuations in sales or other performance metrics that occurs as a result of predictable, recurring factors such as time of year, holidays, or weather patterns. Seasonality can have a significant impact on sales, and it is an important factor to consider when analyzing the impact of marketing activities on sales. By accounting for seasonality in the MMM model, marketers can obtain a more accurate estimate of the impact of marketing variables on sales, and identify patterns or trends that are unrelated to marketing activities.
Brand-driven sales is the portion of sales or revenue that can be attributed to the strength and influence of a company's brand. Brand-driven sales are typically measured by including brand-related variables in the marketing mix model, such as brand awareness, brand preference, or brand loyalty. These variables are used to estimate the impact of the brand on sales or revenue and to differentiate it from other marketing activities, such as advertising or promotions.
Brand-driven sales are important to measure in MMM because they represent the long-term value of a strong brand and its ability to generate sales even in the absence of other marketing activities. By quantifying the contribution of the brand to sales or revenue, companies can make more informed decisions about their branding strategies and investments and allocate their marketing budget more effectively to maximize their ROMI and profitability.
Externally-driven demand is the portion of sales or revenue that is influenced by external factors beyond the control of the company, such as weather conditions or the COVID-19 pandemic. In MMM, externally-driven demand is typically measured by including external variables in the marketing mix model, such as temperature, precipitation, or the number of COVID-19 cases in a region. These variables are used to estimate the impact of external factors on sales or revenue and to differentiate it from other marketing activities, such as advertising or promotions.
Short term and long term effects
Short term and long term effects are used to describe the impact of marketing efforts on sales performance over different time horizons.
Short term effects represent the immediate impact of marketing efforts on sales performance. These effects can be seen within a few days, weeks, or months of implementing a marketing campaign, and may include a temporary increase in sales, brand awareness, or customer engagement. Short term effects are measured using metrics such as sales lift, ROI, or conversion rates.
Long term effects represent the lasting impact of marketing efforts on sales performance. These effects may take longer to materialize, and may be seen over a period of months or years. Long term effects may include changes in brand perception, customer loyalty, or market share, and are typically measured using metrics such as customer lifetime value, brand equity, or market share.
Marketing optimization related terms
Incremental sales is typically divided by individual media channel, and in addition to medium-specific incremental sales and ROMI, the results should also reveal Diminishing return curves and saturation points for each medium that'll help the marketer to optimize the marketing effectiveness.
Diminishing return curve
A Diminishing return curve is a graph that shows how the incremental impact of a marketing effort decreases as the level of that effort increases. You can find out more about this topic here.
Optimization in Marketing Mix Modeling is the process of finding the best combination of marketing tactics or actions that will maximize a specific objective, such as sales, profit, or market share. It involves identifying the optimal allocation of marketing resources across various channels or campaigns. MMM usually has a budget optimization tools which will help you allocate your marketing budget more effectively or reach your established sales targets.
Advertising response curves
Advertising response curves refer to the relationship between advertising spend and its impact on sales or revenue. In MMM, advertising response curves are typically estimated by analyzing historical data on advertising spend and sales or revenue and fitting a mathematical model that describes the relationship between the two variables. The shape of the response curve can vary depending on factors such as the industry, product, or market conditions.
Saturation points represent the point at which additional investment in a particular marketing activity, such as advertising or promotions, will no longer result in a proportional increase in sales or revenue. Simply put, at a certain point all marketing investments reach a saturation stage, making them unprofitable.
Bayesian Marketing Mix Modeling terms
Bayesian Marketing Mix Modeling is the gold standard within the field, as it enables the modeler to utilize hierarchical models and business priors in setting up the model. Moreover, as Bayesian MMM isn't a black box solution, it also enables model validation over time. No model is perfect, however, and so it is a good idea to use Lift-test to calibrate your MMM results every now and then.
Bayesian is a statistical approach that involves the use of Bayesian inference to estimate the parameters of the model. Bayesian analysis is a technique that allows the incorporation of prior information into the analysis, which can be useful in cases where the sample size is small or the data is noisy. By using Bayesian analysis, MMM models can produce more accurate and reliable estimates of the impact of marketing variables on sales or other performance metrics. Additionally, Bayesian methods can help to quantify uncertainty in the estimates, which can be useful in making informed decisions about marketing strategy and budget allocation.
Bayesian inference is a statistical technique used in Marketing Mix Modeling (MMM) to estimate the posterior distribution of the parameters of the model. Essentially, prior knowledge or beliefs about the parameters are combined with the observed data to obtain a posterior distribution of the parameters, which represents the updated beliefs about the parameters after analyzing the data.
Simply put, we start with some prior knowledge or beliefs about how certain marketing activities might affect sales, like "I think that increasing advertising spending will lead to more sales."
Then we look at data and see how much sales actually increased when we increased advertising spending, as well as other marketing activities. We use this data to update our beliefs and make better predictions about how different marketing activities might affect sales in the future.
By using Bayesian inference, we can make more accurate predictions about how our marketing activities will impact sales, and make better decisions about how to allocate our marketing budget to get the best results.
Hierarchical modeling is a statistical approach used to analyze complex data sets that have a hierarchical structure, such as sales data at different levels of aggregation, such as by region, product, or channel.
Business priors are the preconceived beliefs or assumptions that a company has about the impact of different marketing activities on its sales or revenue.
In MMM, business priors are often used as a starting point for estimating the impact of marketing activities on sales or revenue. For example, a company may believe that increasing its advertising spend by 10% will result in a corresponding increase in sales by 5%. This belief can be used as a business prior in the MMM model, and the model can be refined and adjusted based on the data to arrive at more accurate estimates of the impact of advertising on sales.
The process of evaluating the accuracy and reliability of the marketing mix model by comparing its predictions with actual sales or revenue data.
Results calibration is the process of adjusting the estimates produced by the model to better match the actual results observed in the marketplace.
Lift tests are statistical tests used to measure the impact of a specific marketing activity, such as an advertising campaign or a promotion, on sales or revenue. You can read more about them here.
Other relevant terms in Marketing Mix Modeling
The Adstock effect is the phenomenon where the impact of advertising on sales or revenue persists over time, even after the advertising has ended. You can read more about it here.
The Spill-over effect is the impact of one marketing activity on the performance of another, unrelated activity, either positively or negatively.
Promotion uplift is the incremental increase in sales or revenue that is generated by a promotional activity, such as a discount or a coupon, over and above the baseline level of sales or revenue that would have occurred without the promotion.
In MMM, the promotion uplift is estimated by comparing the sales or revenue during the promotion period to a similar period when the promotion was not in effect, and attributing the difference to the impact of the promotion. The promotion uplift can vary depending on factors such as the size and duration of the promotion, the type of product or service being promoted, and the target audience for the promotion.
SISO is a term used to describe a situation where low-quality input data is used to generate equally low-quality output. The term emphasizes the idea that the quality of the output produced by a system is only as good as the quality of the input data it receives.
In other words, if the input data is inaccurate, incomplete, or irrelevant, the output generated by the system will also be inaccurate, incomplete, or irrelevant. This can lead to erroneous conclusions, flawed predictions, and ineffective decision-making.
Curious to learn more? Book a demo.
Related articlesRead more posts
No items found!