FAQs - Holmes & Cook - award winning experts in evaluating marketing effectiveness  

 

How does econometrics differ from marketing mix modelling? 

Marketing mix modelling is originally an American term and is now being freely inter-changed with econometrics. This may not always be appropriate though because econometrics has a very specific approach and philosophy not shared by all statistical disciplines. Econometricians are particularly interested in ensuring their models have a correct theoretical structure i.e. that they do not simply find correlations between sales and the various marketing levers but capture the appropriate short and long-run relationships between them. Econometricians are also concerned with the technical validity of models. A wide range of econometric validation tests have been developed in recent years which specifically acknowledge the properties of marketing and economic data and which econometricians employ. 
 
 

I have heard other terms in association with econometrics but am not clear about their role? 

Econometrics is often discussed in the context of marketing or advertising efficiency, ROI, brand value management or customer value management. By analysing the patterns in historic consumer response, and quantifying what happens to e.g. sales when a particular marketing lever is pulled, its results help to devise strategies which will improve efficiency and generate value. For a detailed explanation and report please click here. 
 
 

Do I need econometrics?  

The three most powerful attributes of econometrics are its ability to identify, separate and quantify the influence of key factors on sales. If you are struggling to understand why your sales are going in a particular direction, if you can’t tell what is doing what or if your boss or the Finance Director are pushing you to show what return you got on that media or promotional investment, then econometrics probably can help. 
 

How long does it take? 

It depends on the scope of the project. Once you have collected and handed over all the data (probably allow yourself a month for that) the econometricians will need anything from 8-12 weeks to final debrief. Particularly difficult projects can take longer as extra “thinking” time as well as modelling time may be needed. The right model won’t always drop out of the data straight away. 
 
Please allow extra time for data queries. Source data isn’t always correct or clear and the econometricians may need to send selected data back. 
 
 

How much data do I really need? 

This does depend on the questions being asked. There are a number of factors to take into consideration. Firstly, the more influences there are on e.g. sales, then the longer history will be required. Where there is any sort of seasonality in the market, a minimum of three year’s monthly (or more frequently reported data) is recommended to ensure underlying seasonality is correctly separated from seasonal communication or promotional activity. Data will be required for sales and all noteworthy potential influences. 
 
 

How can I tell whether a model is any good? 

There are two ways of assessing a model – both which should be used. Firstly, do its findings make (common) sense? That is, is the model theoretically sound? If you are struggling to believe the results, there is probably an issue. Second, does the model have appropriate statistical properties? The person doing the modelling should subject the model to a range of statistical tests (which the final model should pass) and the intuition of which the modeller should be able to impart to you. 
 
 

Never been involved in an econometrics project before – any tips? 

Don’t be afraid to ask questions about stuff you don’t understand or about what you might get out of the project. And keep asking until you are clear. 
 
Nominate a “data Tsar”. All data from you goes through this person to the econometrics agency and is pre-checked to ensure that it is complete and clear. Ideally, it should be someone who has been with the company for a while and has a good knowledge of the various data sources. 
 
Hold interim meetings where the client is willing to discuss the interpretation of the results and the reasons for any anomalies. It is your knowledge of your market here which is vital, not your understanding of statistics. Just talking about what has happened at particular periods of time can help the econometrician to formulate the model or correctly identify anomalies. 
 
 

How do you develop a succesful project?  

The most successful projects share a number of features. They typically kick-off with a “brainstorm” where main stakeholders (client and their agencies) can share their intuitions regarding important events in the brand’s recent history and the way the brand responds to key stimuli. This then provides the “theoretical background” against which the models are developed and against which everyone can jointly assess whether the modelling results are sensible. 
 
 

How do we share data with Holmes & Cook? 

There is someone at the client end nominated as the “data Tsar”. All data goes through this person to the econometrics agency and is pre-checked to ensure that it is complete and clear. Ideally, it should be someone who has been with the company for a while and has a good knowledge of the various data sources. 
 
 

How do we track the project?  

Interim meetings where the client is willing to discuss the interpretation of the results and the reasons for any anomalies. It is your knowledge of your market here which is vital, not your understanding of statistics. Just talking about what has happened at particular periods of time can help the econometrician to formulate the model or correctly identify anomalies. 
 
For more information or an initial discussion about how we can help your business please call us on 01908 745904. Alternatively, please click here to complete our short contact form >> 
 
Holmes & Cook - experts in evaluating marketing effectiveness