How Vodafone uses data to guess customer’s satisfaction rating

Vodafone has figured out a way to calculate its customers’ net promoter scores (NPS) without having to survey them directly.

The telco’s marketing communications chief, Nilanjan Sarkar, told SAS Forum delegates in Sydney last week that only five to seven percent of customers typically respond to surveys gauging their satisfaction with customer care.

So halfway through a project to enhance its data analytics, Vodafone’s marketing team decided to trained its SAS customer profiling engine to re-populate customer sentiment for every one of its account holders on a daily basis.

Using survey and customer interaction data it has built up over time, Vodafone is “training” its algorithms and prediction models to project how a customer is likely to respond to a particular event - such as breaching their data cap - based on where they are in a specially mapped user lifecycle, Sarkar said.

Vodafone has sorted its customers into four groups: those still enjoying shiny new phone feeling are at the ‘welcome’ stage.

Just beyond that, at the 'consolidate' stage, Vodafone will try to solidify and expand a customer's business. If anything goes wrong they might reach the ‘at risk’ stage. The telco will then try to thank and nurture those who make it to the ‘loyalty’ stage.

“These categories allow to develop broad response rules for certain events, because a similar event in a welcome stage or a loyalty stage could generate a very different response from us,” Sarkar said.

Vodafone is using past data on how customers have responded to experiences on the Vodafone network to automatically re-calculate NPS across all accounts.

“A particular set of experiences may have led to a particular net promoter score for that interaction,” Sarkar said.

“You can take those scores and project them on to a larger set of customers who have had similar experiences, and you can predict their happiness and unhappiness. Over time your models can be trained to predict a certain customer experience score based on a certain set of events."

The results are fed into the SAS real-time decisioning engine which populates each account profile in Vodafone’s Salesforce CRM with guidelines for the most appropriate way to communicate with each customer according to their specific circumstances.

“We need to make this simple for the people who are front of shop,” Sarkar said.

This customer database at Vodafone is now re-populated on a daily basis. The telco is currently three years into its data analytics upgrade, which Sarkar describes as “one of the largest and the longest customer experience enhancement projects that we have undertaken in the last few years”.

He said the journey so far has allowed the company to halve the number of communications it sends out to account holders, and spend more time listening and profiling their needs.

“We have been able to organise ourselves into proactive and reactive teams that are able to take that information, which is based on a prediction of what a customer will go through in terms of his or her experience, and act on that in a very agile manner.”