Dennis Hecht, Chief Intelligence Officer, Rooster Strategic Solutions

If the data actions in your organization were bound into a single volume, what kind of book would it be? For many marketers, unfortunately, it’s a dusty resource book left unopened for long periods of time; to others, a useful but mostly incomprehensible tome, as friendly as a Russian existentialist novel; others might see it as an atlas, a map to reach a specific location or objective. But I prefer to think of it as a build-your-own-adventure series where the data tells you where you need to go.

In previous installments of this series, we examined how to create a Customer Data Platform (CDP), how to turn that data into shoppers, and ways to identify where prospects are in the purchase lifecycle. Now it’s time to create predictive equations from your available data.

For instance, one of the first times I used data science to help predict customer behavior was to help an equipment manufacturer that wanted to sell more of a current product. Looking at their past sales history – which is always a great place to start – the data showed that a cluster of farmers who purchased this product had also purchased another piece of equipment from an adjacent product line or had purchased a competitive product in a younger age group. From there it was easy to create pools of people who were likely to buy. Pretty basic? Yes, and it just proves that data science doesn’t have to be complicated, and it’s considerably more effective than throwing money at a blanket demographic, such as 500-plus-acre growers.

This was an example of a purely data-driven approach, the library equivalent of using your data as a map. You look at the data you have – all the attributes and all the values within these attributes – and identify the driver or drivers that predict sales. It’s a direct, objective approach that lets you make an educated guess about customer behavior (better than random chance). Where data science gets a little more complicated and a lot more effective is when you start with the educated guess or hypothesis and use your data to test and refine it.

Here’s another example. The subject matter experts at an insurance company I worked with authored several articles about risk management for their website and wondered if people who read these articles would be more likely to purchase a new policy. By targeting people around their behavioral propensity to read risk management articles, we were able to generate quality leads. Next, by further examining these leads and identifying who went on to purchase policies, we were able to pinpoint other behavioral attributes that led to even higher adoption rates with bigger quotes.

Obviously, data science by itself won’t sell anything. You still need a solid marketing plan.  But once you demystify the assumptions you have on why people buy your products you can use the data you’ve collected to help you create a smarter, more effective marketing plan. As mentioned previously, a great place to start is to look at people who have already purchased your or a competitor’s product or service, and drill down to identify the real motivation for the purchase. Was it a relationship? Price? Size? A different emotional driver? You’re going to find out that a lot of your assumptions are wrong, and that’s ok. You’ll also want to ensure that you have a clear plan of the outcome that you’re predicting, or what you’re trying to do with the data you have, lest you fall into the trap of “analysis paralysis” where you spin your wheels looking at data but with no clear plan on how to make it actionable.

In the end, data science can help you understand where a particular person is in the customer lifecycle so you can set up natural trigger points that encourage them to engage and purchase. It’s all about the journey they’re already on. It shouldn’t feel intrusive or awkward, it should feel like a natural progression of interest. Good stewards of data science are simply trying to enrich a tailored experience that encourages a prospect to make the right decision on whether or not to move further down the customer lifecycle.

If the science seems mysterious or “black box,” you may be dealing with the wrong partners. It should be transparent, repeatable, explainable, and most important, trusted.

If you want to refine – or start – a data science strategy, or if you’re not even sure where to begin, I’d love to have a conversation. Otherwise, look here for the next article that discusses how best to share your successful data science stories with others in your organization, particularly those in areas outside of marketing.