Econometrics, also known as marketing mix modelling, has become a popular method of measurement in marketing. This chapter outlines the key benefits and challenges of this approach, and how marketers can implement it effectively.
- Econometrics/marketing mix modelling (MMM)
- Benefits and challenges of econometrics/marketing mix modelling
Econometrics/marketing mix modelling (MMM)
Econometrics, or marketing mix modelling (MMM), is an approach that brings together data from a variety of sources such as on and offline media, sales and promotions, as illustrated in Figure 1. It can also factor in external influences such as seasonality and the economy.[1] It then applies statistical models to help marketers better understand the impacts of each of these variables on marketing KPIs such as sales or awareness.
Figure 1: Econometrics/marketing mix modelling
Source: Econsultancy
This measurement approach is growing in popularity, as it one of the methods that depends the least on data collected by third-party cookies, instead pulling in a wide range of data sources. It remains a viable way for marketers to measure the effectiveness of digital marketing, as well as gaining insight into marketing incrementality. The support for this approach is evident from responses to Econsultancy’s Future of Marketing report, where close to two-thirds of respondents (63%) indicate they are using or considering econometrics as an alternative form of measurement.
By bringing together a companies’ offline and online channels in a comparable way to provide an overall view of what is affecting sales, MMM enables companies to make those bigger strategic decisions and provides good media planning input. The importance of this was highlighted by Dr Grace Kite, Economist and Founder of Magic Numbers: “Everything that you get from all the advertising platforms is not comparable because of the different methodologies.” Marketing mix modelling is also an approach that respects privacy regulations.
Benefits and challenges of econometrics/marketing mix modelling
When evaluating econometrics/MMM, key benefits include:
High-level strategic insights. Econometrics by definition includes a broad range of data sources, giving marketers an overview of all the elements of a marketing campaign and providing some top-level indications of what is working most effectively.
Improved trust from a more comprehensive view. A layer of mistrust can often exist around attribution models and how the algorithm has been calculated by a particular platform or publisher. MMM can provide a full view of a company’s budget and the impact, as long as all stakeholders are engaged and provide all the data that could impact the model.
“MMM is not an if but when you do it, since it has so many benefits. MMM is one of those tools that gives you a read on incremental benefits you get from the market, seasonality, weather, and the impact of these on your sales. It is about using signals to inform the model and about making small correctional changes to your strategy rather than pulling a handbrake on ad spend.”
Gary Danks, General Manager, AIM, Kochava
Improved efficiency from marketing spend. A core output of MMM is a cost-benefit/ROI analysis providing insight into which media channels are the most cost-effective and which should be used more in the future.
Estimate marketing incrementality. MMM can help estimate incrementality, the additional value that marketing brings, since it includes non-marketing business drivers in the models in order to best estimate and forecast total sales.
“Marketing mix modelling shows the marketer where the influence actually came from, providing a different perspective on the market, rather than relying on last-touch attribution, for example, which doesn’t take account of all the touchpoints a consumer may have interacted with. Touch attribution will also treat a sale the same whether it happened on a Monday or a Saturday, and it will not take account of seasonality, which marketing mix modelling can do.
“Marketing mix modelling gives a marketer that top-down view. It will attribute activity across all of their different online and offline media and take account of external factors that affect sales, whether they’re positive or negative. Marketers can then strategically adjust their marketing investment to allocate it where there’s more influence than maybe they thought they were getting, as well as put the brake on networks where there’s less influence.
“Marketing mix modelling delivers efficiency from a company’s marketing spend, where clients see an improvement in sales of around 17% for the same amount of budget. It’s giving them the insights to make better investment strategies and increase their return on investment.”
Gary Danks, General Manager, AIM, Kochava
However, some of the challenges often raised around traditional econometrics include:
Limited granularity. While econometrics modelling provides high-level strategic insights, it can create more challenges for marketers who are seeking to understand the specific detail of what did or did not contribute to the success of a campaign, by channel, media placement or creative, for example.
Not available to everyone. To implement econometrics effectively, much data, both historic and current, is needed to input into the statistical models. Not every marketer has this data, and even if they do, they may lack the budget to invest in the approach.
Lack of wider understanding. Econometrics models are complex tools that depend on high-level statistical analysis. There may not be the internal skills available to carry out the analysis, which then requires external agency resources.
Using econometrics/marketing mix modelling
It is important at the outset for companies considering this approach to have clarity on the scope of what they want to explore and the key questions to be answered. MMM requires historical data for the analysis and there will be resources required in the data collection stage and for the hypothesis development. In terms of carrying out the modelling, companies will need to determine whether they have skilled data analysts in-house who have experience of MMM, or whether they need to engage the services of an external agency.
“We promote a people-first approach to measurement. We bring everybody from across the business together to understand what drives sales, what their hunches are, what their knowledge is, and what research have they got already. Then we build that into our models along with marketing and everything else to help explain what is happening.
“We can look at our data and even with a holistic picture get quite granular, which enables a company to get an understanding of performance across their digital channels. For example, this could be around keywords and different bidding strategies for PPC. For social media channels we can provide feedback on the objectives people chose, whether impressions, views or reach.
“We are able to connect the data and look at a time series of data to build a model of sales over time. This enables us to map the trends and variables to understand what’s driving sales at any different point in time and what is working. I see the best value from this type of modelling is for making those bigger picture decisions.”
Dr Grace Kite, Economist and Founder, Magic Numbers
In terms of the frequency of reporting that a company may want, this will again depend on their marketing activity, business priorities, product portfolio, budget and ability to respond to the outputs from the analysis. Interviewees for this report that were using MMM were likely to carry out a minimum of two pieces of analysis a year, with others conducting analysis more frequently.
As a result of technological advances, solutions have emerged for marketing mix modelling which can also provide results on a more regular basis, whether that be daily, weekly or monthly.
“Marketing mix modelling is built on historical data so will need at least a year’s worth of data, but it’s a data science methodology so the more data, the better. This includes whatever a company’s KPIs are that they are measuring and optimising against at that moment, as well as sales, cost and ad spend data.
“We then have access to econometrics such as seasonal data, elections, sporting events which can all have an impact on a brand’s sales activity. The data is then injected into the model. Sometimes our clients also have market intelligence data as well as competitive data that they’ve got access to, which can be used in their model as well.
“With our product, we pull in the data every day via APIs, and then we refresh the model and publish the results every 24 hours, so our clients always get the most up-to-date data and insights. Whilst we pull data from a company’s attribution platform, we only work at country level. We look at their total sales for the UK – for example, their sales funnel could be something like installs, first-time purchase, total revenue, second purchase etc. – and we model all of the sales funnel.
“If the data is good, we can get our model to an average of 95% accuracy over a two-week forecast. In comparison, when we modelled last-touch attribution, the same system could only get to a 30% accuracy over a two-week forecast, meaning for a company using last-touch attribution data for measurement, their model had a 70% error rate over a two-week forecast.”
Gary Danks, General Manager, AIM, Kochava
A key reason many companies are choosing to use marketing mix modelling is its ability to provide context and help answer the ‘why’ behind changes in sales, particularly if campaigns run across a number of channels. Dr Grace Kite, Economist and Founder of Magic Numbers, provided some example case studies across different sectors where MMM had helped answer some key questions.
Table 1: Use cases for marketing mix modelling
| Sector and challenge | Analysis and solutions |
| Online retailer wanted to understand how to combat slowing growth and falling margins. |
|
| Financial services client with a multiple product portfolio wanted to understand how to balance their budget across the portfolio to drive overall profitability. They wanted to test whether there was an opportunity to test stronger weights on above-the-line activity. |
|
| Subscription services company recognised their category was changing and their traditional tactics needed to evolve. |
|
Source: Magic Numbers
It is important for companies considering this approach to be clear on the questions to be answered by the analysis. Working with agencies can help to formulate these. As well as deciding whether this measurement approach is the right one, companies should carefully select the agency/consultancy they work with if they choose to collaborate with an external partner. In a letter published by 17 senior and well-respected specialists with deep expertise in marketing mix modelling, including Les Binet and Dr Grace Kite, the authors provided advice for companies when evaluating suppliers to ensure the solutions being considered are not too simplistic and are properly adapted to the specific business being analysed. They outlined a series of questions to ask, and recommended seeking three quotes.[2]
- Does the model include factors like price, economy, seasonality, covid? Will the supplier report on how these things affect the business?
- Does the model cover at least two years of data, preferably three?
- Does the supplier measure how upper-funnel ads like TV and YouTube affect outcomes in lower-funnel ads like PPC?
- Will the supplier share advertising response curves with the company’s media planners?
- What happens if the results come back and the company doesn’t believe them, say, because they don’t line up with something else they know?
- Will the supplier be able to explain the model to a company’s finance people? Will their numbers line up with those from finance?
- Could analysts who understand regression look under the bonnet at the model and all the tests and carry out statistical due diligence?
- Can the supplier demonstrate to the company that the models are good at forecasting?
Magic Numbers | Beware Misleading MMM[3]
- Econometrics/MMM combines offline and online channels to provide an overall view of what is influencing sales, from online media data to seasonal and promotional sales data.
- Econometrics models are complex tools, so when introducing this measurement approach to the business, it is essential to ensure the right skills and capabilities are available internally, or that there is budget to invest in external agency resources.
- Aim to have access to at least a year’s worth of data, as the effectiveness of econometrics/MMM is dependent on the amount of data available.
- Prepare a list of questions when evaluating external suppliers to ensure solutions are adapted to specific business needs and objectives.
- Econometric modelling can provide useful top-level strategic insights, but it also has limited granularity, which can present marketers with more challenges.
This guide is based on primary research which involved exploring findings from two reports:
- Econsultancy’s 2023 Future of Marketing report, which was based on a survey of 835 client, vendor and agency-side marketers. The survey was fielded to Econsultancy and Marketing Week’s audiences between 9 June and 3 July 2023.
- The Language of Effectiveness 2023 report has been produced using responses to an online survey of 1,369 qualifying marketers conducted by Econsultancy’s sister brand Marketing Week between 27 March and 28 April 2023.
In-depth interviews were carried out with industry experts. Econsultancy would like to thank the following interviewees for their invaluable contribution of time and expertise to this guide:
- Kumar Amrendra, Head of Digital Marketing, Sky UK Ltd
- Amy Blasco, Partner, Enterprise Data, Experience and Marketing Lead, IBM
- Laura Chaibi, Director, International Ad Marketing and Insights, Roku Inc
- Sebastian Cruz, Regional Digital Marketing and Media Director, Shiseido, Asia Pacific
- Gary Danks, General Manager, AIM, Kochava
- Mauricio Ferreira, Marketing Effectiveness Lead, Confused.com
- James Hurman, Founding Partner, Previously Unavailable
- Gabriel Hughes, CEO and Founder, Metageni
- Dr Grace Kite, Economist and Founder, Magic Numbers
- Chloe Nicholls, Head of Ad Tech, IAB UK
- Roxane Panopoulos, Group Manager, Regional Measurement & Insights – Netherlands and Nordics, Snap Inc
- Marina Peluffo, Head of Business Intelligence, Prima (speaking as industry expert)
- James Sharman, Northern Europe Digital Acceleration Lead, Haleon
- Steven Silvers, EVP, Global Creative and Media Solutions, Kantar
Lynette Saunders is a Senior Analyst at Econsultancy, where
she works on delivering industry-leading research, briefings and
reports for the digital marketing industry and speaks at a number
of external conferences.
Lynette’s previous experience includes delivering web analytics, measurements and insights, as well as leading usability and
customer experience programmes focusing on improving the
overall online customer experience for Cancer Research UK
and the Royal Mail Group.
