Rob Walker kicked off day 2 of PegaWorld talking about Decision Management and the role of analytics in customer engagement. He began by talking about John Boyd and his theory of engagement – the OODA loop – Observe, Orient, Decide and Act (something I referenced in Decision Management Systems: A Practical Guide to Using Business Rules and Predictive Analytics). The OODA loop is ideal for next best action:
- Observe customer behavior
- Orient to the channel and context
- Decide on the best action
- Act to fulfill this action.
Good next best action approaches, he says, have multiple dimensions of customer engagement. The action must be:
- Customer Desirable
- Brand identifiable
- Economically viable
- Operationally feasible
When customers are engaging with an organization in real-time it is only possible to deliver on these dimensions if the OODA loop is being executed very quickly – responding to changes as customers interact with the user interface, tell customer service staff something, exhibit a mood etc. Decisions must be contextual.
At the same time they must be differentiated, treating customers differently. A small number of customers typically contribute the majority of your profits while others are loss making. Analytics allow you to identify which customer is which, focusing on the value of each specifically. This might require Big Data but often only requires “little data” – standard, structured data about products, transactions and customers.
To drive these analytics, which are highly automated in the Pega Decision Management product, Pega creates what it calls the Customer Movie. This pulls together all the strands of data, little or big, to create a high resolution picture of the customer. A series of snapshots create the movie, capturing everything a customer does or that is known about the customer as they do. The customer movie can be used to see what happened at a moment in time and “fast forwarded” to see how different scenarios play out for the customer.
All of these decisions, each of these customer engagements, are connected. As channels have multiplied, engagements have become faster and the level of noise has increased it becomes both more difficult and more important to tie customer decisions together. The decisions made about how to engage with prospects, onboard customers, nurture customers and more should all leverage the whole customer movie and be connected and managed using the OODA loop.
Rob introduced the way decision strategies are defined in the Pega platform (in Pega Decision Management) to show how different customers with the same interests but different customer value and churn risk can be treated differently across channels and over time. As customers accept offers this changes their context, adding to the customer movie and allowing new differentiated decisions to be made in that new context.
A customer, EE (the merger of Orange and T-Mobile in the UK) came up next to talk Decision Management. EE has 27M customers in the UK and uses the OODA loop to manage customer interactions. This is easy to say but hard to do – differently skilled and distributed customer agents have to handle interactions with customers, working with vast numbers of product combinations and those customers are highly differentiated. EE has been applying this approach as it moves from a batch, campaign orientation to an outbound, next best action mindset.
EE takes the customer lifecycle information, the reason for the call, the products the customer has, their history, rules for which offers can be fulfilled right now and uses all this, as well as analytics and business objectives, to drive a short list of actions – a top 3. This top list if recalibrated as the conversation progresses. The conversation is relevant to the customer and the system means that agents are consistent and able to focus on the conversation not the rules about offers etc. It allows for private offers to be included in public channels and focuses and has driven tremendous results.
EE has seen great results from this. 90% of the offers accepted come from the top three lists that the system generates, they generate 4x the accept rates, and EE gets extra value and better retention rates at the same time.
In addition analytics help them identify customers with the potential for extra value and for these customers they generate highly personalized recommendations based on an individual investment budget. This too is dynamic, recalibrated as agents interact with customers. These personalized offers have improved retention by 10% and these customers are worth 14% more.
This whole environment is multi-channel. For EE most customers go online first when considering upgrades before calling and visiting a store to see the phone. Making sure this conversation is connected ensures an agent can pick up an upgrade process and move it forward even as customers change channels.
EE’s use of Decision Management is a classic illustration of using analytics to drive differentiated, contextual and connected decisions across the lifecycle.