Marketers that effectively use analytics have a lot in common with fortune-tellers: They can “see” into the future. Unlike the traditional carnival clairvoyant gazing into a crystal ball, however, marketers who apply analytical techniques to their marketing programs can gain powerful, useful insight to apply to their future marketing campaigns.
Ask marketers how they use analytics, and you’ll likely hear a myriad of answers. Most marketers are accustomed to using simple reporting and traditional online analytical processing (OLAP) tools that are good at capturing data from past campaigns and customer activity. Others take their analytics a step further, using automated analytics and “push button” software tools that add retrospective information to lend insight into campaigns. The organizations that really stand apart use advanced, customized analytical techniques. By using their data and past results, they gain greater insight based on logical inferences. This kind of insight gives companies the proverbial leg up on the competition. To gain this advantage, most marketers will need to step outside their traditional comfort zones and enter the realm of analytics. But analytics doesn’t have to be intimidating. In fact, most marketers can be successful with just a basic understanding of analytics, provided that they partner with someone who carries demonstrated knowledge and experience in this area.
When done right, analytics will help you maximize all aspects of your marketing programs–audience, offer, contact, media, channel, message, creative. Each requires a different analytical approach, or modeling technique, that combines customer information from a variety of sources.
For example, you can find the best audience, or customers, using lifetime value (LTV) and response models, while you can determine effective offers by using segmentation and offer models. You can evaluate your best contacts by using promotional history analysis. And you can take a look at your message and creative with customer profiling and experimental design models.
No matter what model you use, a best-practices approach to modeling includes three major components:
- Model scoping: In this step, the context of the models–that is, what you hope to learn–is defined, and key metrics are established. A project plan and timeline is created.
- Model estimation: Samples are identified, data are prepared and validated, and an exploratory data analysis (EDA) is performed.
- Model rollout: Scoring process and models scores are defined, and a quality assurance check is performed to ensure that all elements are in place.
Using models will help you determine the best use of your marketing dollars and will allow you to focus your efforts on the best customers. Modeling will also lend insight into your most effective messages, creative, offers, and prices so that you can more effectively market to these audiences in the future–no crystal ball required.
Scott Cone is vice president/client leader for Merkle, a Lanham, MD-based database marketing services provider.