Insight, Analytics and the New Marketing
Who’s good at econometric modeling?
Last week Ogilvy announced a partnership with Marketshare Partners, a marketing science company based in LA with offices in NY, San Francisco, Chicago and London. I had been looking for a while for a partner who could help us with econometric modeling, the type of analysis that can help you tie overall marketing spend to business outcomes and can help you determine how much to spend on marketing and how to allocate it by product, segment, geography and medium.
I won’t go into the details of why Ogilvy chose Marketshare Partners or what the joint offering exactly entails. This would not be the right place for it. However I did speak to a lot of different companies in this space over the last couple of months and I thought I’d share my thought on how they differentiate from each other. Hopefully this could be useful when you are looking for a partner - if you need one just give me a call 
I started with laying out the various tasks involved with running an econometric modeling project that would result in a marketing investment strategy. These steps are listed below.
Data Gathering : you need to collect data on marketing spend and performance from a wide range of sources.
Data Aggregation : you need to put all that data in one place where it can be accessed easily by analysts.
Statistical Modeling : you need to build the statistical models that link marketing spend to business outcome.
Optimization : you need to use those models to determine what the optimal marketing spend and allocation is.
Scenario Planning : you need to build several investment scenarios that demonstrate the impact of different investment strategies on business performance.
Dashboards & Management Systems : you need to give non technical users access to the data and scenarios through decision support tools.
Strategic Consulting and Planning : you need to help decision makers use these tools for decision making.
These are generally the tasks at hand when you run an investment optimization project that uses some science to get to the answer. These tasks map perfectly to the market differentiators that started to emerge throughout my conversations with the various vendors. This is illustrated in the diagram below.
Data Access : Some vendors had access to special data sources through alliances they had with 3rd part data vendors or search engines.
Automated Data Preparation : Data preparation can use up a lot of your resources on projects like these. Some vendors had automated a lot of these steps so that they could do the data gathering and aggregation described above faster, better and more cost effectively.
Unique Algorithms : The algorithms have been around for decades so you would think all econometric modeling skills are pretty level. However, you would be surprised how much variance there is in the quality of the statistical modeling between different vendors. Make sure you look underneath the hood and talk to who is actually doing the modeling.
Decision Support Tools : Some vendors have packaged their tools in dashboards and scenario planning and decision support tools. These tools can be extremely useful for helping the end users adopt some of the scientific evidence into their decision making.
Value Added Services : This is probably the most important area. I have seen so many modeling projects collect dust on the shelf of the marketing intelligence managers. Building models is easy. Using them for decision making always seems to be the hardest. You should always look for a partner who can help you with both.
Hopefully some of these learning can be useful for you at some point. The reality is that far too few marketers are using econometric modeling to really understand the impact of their marketing efforts on their business. As a result, they end up making multi-million dollar decisions on pure gut feel. If there is anything this recession has taught it’s that that will ultimately get you into trouble!
[...] Who’s good at econometric modeling? [...]