Predictive analytics: A return to fundamentals

This article is an excerpt from Experian Marketing Services’ 2016 Digital Marketer Report. Download the full report to discover more insights and trends for the upcoming year!

These days, people are always looking for the latest buzzword. For a while it was big data, which introduced a level of scale and speed in terms of processing and making sense of data that no one had yet experienced. Today’s buzzword is predictive analytics, but many organizations don’t know where to begin to make it a reality or to translate it into business value. They look to my team for guidance on what we’re doing and best practices we can share.

Ironically, the spotlight on predictive analytics has actually emphasized the need for marketers to return to fundamentals. Brands know their customers the best, so predictive analytics must essentially pass the smell test: do these insights align with the intangibles you know about your customer? My best advice for creating a predictive strategy is not to aim for the next shiny object. Don’t focus on the algorithms and the technology, but lay a solid foundation of questions you are trying to answer on which you can build a successful analytics program. Specifically:

  • Re-structure your organization so that you have a clear understanding of your goals and what you want to accomplish
  • Make sure you have clean enough data and the technology to make it actionable
  • Forge the right partnerships to succeed

Restructure your marketing organization

Marketers are constantly trying to understand and improve the customer experience. The problem is that marketing organizations are not structured in a way that allows them to effectively engage the channel-agnostic consumers of today. It’s nearly impossible to deliver a cohesive experience when the email team is fighting with the social or mobile team for budgets.

Advanced analytics, like cross-channel attribution, can actually help to solve this problem. Proper attribution allows an organization to look at the customer journey regardless of channel and start to understand the best ways to engage customers, negating the basis for internal feuds.

Specify your marketing goals

Before you write your first algorithm, however, you need to document what the organization is trying to accomplish and make sure the data is ready to produce actionable insights. Talk with relevant business users to identify the questions you’re trying to answer: What don’t you know about your customers? What challenges are you having? Where are you losing engagement? What do you wish you knew? These questions will become the foundation for your predictive analytics initiative.

Check your data inventory

The exercise invariably leads to analysis of your data inventory, in which you’re likely to see data quality issues arise. Executing advanced analytics programs requires data quality and identity management. No amount of predictive analytics and insight can replace what you know about your customer.

Attacking the flaws within the data is the first step. Aligning the data is next in line. But remember, customers are more than just a transaction. Your level of understanding needs to extend above and beyond their purchases. That means being able to marry transaction data with behavioral data, lifestyle data, campaign history and so on, to create a cohesive picture of your customer.

Find the right technology and partners

When you have a platform that’s a one-stop shop, like the Experian Marketing Suite, you can look at performance and levels of engagement across channels all on one screen. This gets teams on the same page, talking about the same customer experience. With a clear layout, you can easily pinpoint organizational silos or areas where the customer experience is disjointed. If you focus on the goals set upfront, it becomes hard to justify why those silos exist in the first place.

Predictive analytics is a buzzword for a reason, and it’s one you can’t ignore. According to our Digital Marketer survey, two-thirds of marketers plan to incorporate predictive modeling programs into their strategies this year. If you’re part of the 33 percent who are not, be forewarned: you will be left behind.

As you build out your predictive strategy, don’t just go hiring data scientists to chase the next big trend. Make sure you have a solid foundation and a clear understanding of what you are trying to achieve. From there, you can start to build really cool predictive analytics to intelligently engage your customers.

Blog CTA banner_DM16 report