Analytics & Testing, Ecommerce and Retail, Technology

Predictive Marketing Analytics with ThinkVine

What would the Return on Investment be if you could change your marketing mix?

This is a question that large customers with complex marketing strategies (that are balanced between a multitude of mediums) ask themselves every day. Should we drop radio for online? Should I shift marketing from television to search? What will the impact on my business be if I started marketing online?

Typically, the answer comes through a myriad of testing and lost marketing dollars. Until now. Marketers have been utilizing past performance to predict future marketing performance. There are huge risks associated with this as new mediums are added over time. The shift of classifieds from newspaper to online is just one small example. If you continued your classified spends without shifting them online, you wouldn’t be reaching the maximum potential. In fact, you could be simply wasting your money.

ThinkVine has been working on “What if” scenarios for almost a decade. Their customers are pretty impressive… Sunny Delight, SC Johnson, LegalZoom, Del Monte, Hershey, and Citrix Online.
agent-based-modeling.png

ThinkVine is able to do this through a proven agent-based modeling system that was actually developed in the 1940’s. By understanding the market segments who have purchased from you through each medium and applying the model to the segments in other mediums, ThinkVine is able to build a predictive model of how your marketing will work in those other mediums. It’s quite a system.
marketing-trend.png

The scenarios that ThinkVine develops can be applied long-term, short-term for occasion-based marketing, and segment-based marketing efforts. ThinkVine can even predict the ultimate scenario… what if you stopped marketing altogether!
no-media.png
Learn more by taking a product tour of ThinkVine’s Marketing Simulation and Planning Software.

Full disclosure: CEO Damon Ragusa and I worked with Bruce Taylor of Praesage many years ago to apply similar methodologies to direct mail marketing. Damon built dynamic statistical models from customer profiles and, using Bruce’s automation, we could automate applying those models to prospect databases. The application was called Prospector and worked brilliantly. Bruce has fine-tuned the application over the years and still utilizes it for a number of large direct marketing clients.

2 Comments

  1. 1
    • 2

      Adam,

      It definitely requires historical data. I suppose if they had enough clients, aggregating profiles could be possible. Doubtful that their clients would appreciate that, though! I think they use at minimum 1 year of data – I think 2 is recommended.

      Doug

Leave a Reply