Behavioural and Engagement Data Key to Success for Predictive Analytics

Matthew Kelleher | Chief Commercial Officer
At the Internet Retailing Expo in Birmingham in April I presented six case studies about the successful (i.e. they made money) application of Predictive Analytics. All the case studies were based on the application of Predictive Analytics to help target customers or prospects at various points along the customer journey, for instance, identifying the single buyers most likely to make a second purchase or which VIP customers were at greatest risk of lapsing.

If you prefer to sit back and listen, instead of read, you can watch my interview with the Internet Retailing Editor-in-Chief:

At the end of the presentation I was asked ‘What is your definition of Behavioural Data?’. I had repeatedly talked about the importance of accurate and complete data to drive Predictive Analytics and described 1st party customer data as falling in 3 types… transactional, engagement and behavioural. But I had fallen into the trap of failing to explain what I meant by each of the definitions I was using. So, very briefly:
 
  1. Transactional Data. Or RFM. Or RFV. What a customer has bought, how much, when, where… Data that is used to build RFM models (i.e. identifying prospects, single or multi buyers, VIPs etc.)
  2. Engagement. Points of interaction with the brand, from opening an email to receiving a catalogue to visiting the website.
  3. Behavioural Data. Usually related to website activity, what the prospect or customer has browsed, their recency frequency across devices, what they’ve clicked, do they visit the website direct from social, have they clicked on a discount offer? Etc.
Behavioural data has always been critical. It is the core of data driven personalisation. By building up, and allowing marketers to react to ‘behaviour’, from marketing the products someone is interested in through to identifying if the consumer is an offer junky or full price buyer, this is what underpins RedEye’s approach to Predictive and AI. Using this information to work in combination with engagement and transactional data identifies prospects and customers in terms of what someone will do next and when.

But despite presenting six case studies all showing conversion and revenue improvements, we can’t escape the fact that there is a little bit of market weariness to the subjects of Predictive Analytics and AI! Gartner are now stating in their Hype Cycle that Predictive Analytics has fallen into the trough of disillusionment! And with so many marketing tech businesses out there are talking about it, but not many are able to demonstrate the real value it can bring for retailers.

Back in 2016, Forbes research showed that 89% of marketers had Predictive Analytics on their roadmap. Fast forward to 2018 and 93% of consumer-facing businesses are unable to use Predictive Analytics. This really shows the disparity between the desire to implement Predictive Analytics vs. the actual implementation.

The aim of the presentation was to try to reinforce the potential value of these tools to the market,  but the key is to start with data. Customers are getting more and more difficult to understand with the proliferation of marketing channels. Just recently Whatsapp was added into the fold, yet another channel that marketers can target their consumers through.

Brands are losing touch as they struggle to track all their customer’s moves, which in turn leads to a decline in customer loyalty. Consumers can feel their favourite brand just doesn’t understand them. As humans we just can’t keep up! This is where AI comes in!

Start by understanding your data. Where is it coming from? I recently joined a panel with the Head of CRM at Domino’s, he told me their journey began by creating a Single Customer view by collating their offline data and combining it with their online channels.

He was right, by putting in the leg work at the beginning and creating a true Single Customer View was key. A Single Customer View means you can tie together transactional, engagement and behavioural data, allowing you to paint the full picture of your customers.

Finally, it is key to apply AI and Predictive Analytics to something tangible. At RedEye, our predictive models are based on the customer lifecycle. By making incremental improvements at each of the key customer moments, you can see substantial increases in overall customer value.
 

"Making sense of large datasets is AI’s strength"


You can find more about how we’ve driven revenue increases for our clients using Predictive Analytics here.
Share:

RedEyeUK @RedEye UK