How do meteorologists determine future weather conditions? Among other resources, they rely on predictive modeling – a system utilizing statistics to create scenarios of possible outcomes. Probability and other mathematical disciplines enter the picture in a big way.
Increasingly, marketers are recognizing the value of this meteorological magic. Supporters of the system are using predictive modeling to forecast possible consumer behaviors. ‘Will they buy or fly?’ is among the many unknowns whose probability is calculated by the modeling technique.
An adequate data supply is critical to accurate predictive modeling. You can’t envision what’s down the road without a solid informational basis. Therefore, thorough data collection is par for the course. Essentially, you’re on a local info quest.
What types of data should you seek? Basically, the desired goods fall into three distinct categories. They are as follows.
Purchase Info. This high value info covers variables such as average purchase value, order size, product choice, and reams of other sales data . Likely, the data is tightly held by the financial gatekeepers. Pry it loose at the earliest opportunity.
Whudja Think? Info. This is the slew of feedback carefully extracted via probing survey questions. The body of info will be answers to questions such as ‘How satisfied were you with the product or service?’ and ‘Would you recommend us to those in your social circle?’.
Exit Info. Like it or not, customer abandonment is a fact of life for most businesses, large or small. Currently, these abandoners are quantified by a measurement known as churn rate. The numbers gleaned from this rate speak volumes about your customer loyalty standing. Not always easy to face up to this form of rejection. Nevertheless, it must be dealt with.
Once the data in your possession, architectural excellence is on the horizon. You now have the raw material with which to build an effective predictive model. Here’s to sunny skies!
If you have any questions or comments about predictive modeling, or about any other brand-related topic, feel free to send them our way.