Five natural enemies of predictive data

Years of experience helping media and marketing companies turn into performance-based data businesses has shown me that CIOs, CMOs and sales managers have high expectations that predictive analysis will be key to future success.

However, many operate with an incredibly oversimplified perception of what it takes to get truly predictive data insights. At the same time, enterprise IT and digital operations have a steep learning curve with the age-old need to “learn the business” that predictive data strategies are being designed for.

I’ve described these obstacles to predictive data — and data strategy in general for that matter — as the “natural enemies” of successful implementation. Some are related specifically to the media and marketing business, but I know from my work in other industries that the principles are the same regardless of the vertical.

Natural Enemy #1 – Operating in a Business Culture

Executives go to business school, not data school, right? So why are we surprised when there is no data culture in an organization? I’m not at all referring to using data for business decisions. What 21st century companies need to instill is an awareness that interactions at any level, in any format are opportunities to gather data exhaust. This means that the enterprise must create data that generates experiences, rather than simply gathering atomized pieces of information with no cohesive, connective tissue.

This requires a corporate culture where employees are rewarded financially for the data exhaust and insight they create. Where in the past it was called a risk-reward structure, it’s now known as a data risk-reward structure. Just like the citizen-journalist movement we’ve seen as a result of easily accessible publishing platforms, we now must now embrace the “citizen-quant” in business enterprises.

Natural Enemy #2 – Confusing Engagement with Behavior

I’ve worked with many companies that have tremendous amounts of data about engagements of many kinds, ranging from clicks on website content to call-center data. While this data should by no means be trivialized, the simple act of engaging is but a very small piece of behavioral analysis that could lead to predictive behavior. Some of the more sophisticated readers of this piece will be saying, “This is obvious!” But I can assure you that many firms try to make what they think are predictive assumptions based on single data points.

I’m not saying this can’t be done, especially with highly commoditized, price-sensitive products. But I do know that this killer data point might be unearthed far too along in the decision process to make a deeper brand relationship with the customer possible.

Natural Enemy #3 – Traditional Taxonomies

Most taxonomies can be characterized as developing buckets equivalent to “animals, vegetables or minerals”. An example in the media business would be content types such as whitepaper, webcast, news article and presentation. Another set of taxonomies might include whether the reader engaged with content on a particular subject area, such as cloud, virtualization or security.

By definition, predictive and behavioral analysis imply a connection between the engagement history and the propensity to purchase or re-purchase a product in the future. For this reason, emotional elements must be baked into the taxonomy architecture so as to capture the neuroscientific triggers critical in determining buying intention. Think about the sentiment and emotional indices that would go into predicting whether a wealth management customer would be likely to purchase certain stocks. Emotional indices such as doubt, optimism, panic, greed and safety become foundational in a predictive sense.

Natural Enemy #4 – Random Acts of Content

In the same vein, content used for engagement must possess the emotion that reflects the taxonomy. Most companies operate in a world of “Random Acts of Content,” where there is no science behind the creation of content and how it relates to future actions. In the new world of content marketing, each piece of content is developed with consideration for tone, theme, buying triggers, business elements and emotional indices that create a picture of exactly what engagement means. Taxonomies that include only format and content topic (ie. webcast on data security) provide only limited insight on propensity to buy.

Natural Enemy #5 – Inside Out Thinking

Many predictive models are based on what marketers think triggers purchases as opposed to what buyers tell us. Again, this sound obvious, but at minimum, the field sales staff needs to be brought into the predictive data strategy to assure the closest proximity to buyer. Smart enterprises will look closely at the customers that have bought the product, and then analyze what journey the buyer took on the way to purchase.

Only by eliminating these natural enemies can enterprise IT and data organizations assure their stakeholders that they will be able to target the best customers and provide accurate reporting on account based pipelines.

RELATED LINKS

The connected healthcare ecosystem: Integrating medicine, data and IT

From lakes to watersheds: A better approach to data management

Why big data is such a big deal now

Trackbacks

  1. […] Five natural enemies of predictive data […]

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: