How Telstra is applying data analytics to customer experience
- 05 May, 2015 07:38
Building a clear action plan around customer analytics, along with cross-function collaboration, have been vital in making Telstra’s customer engagement efforts a success, according to its director of research, insights and analytics, Liz Moore.
“Analytics and insight for its own sake is just an academic exercise: It’s about everything we do driving to an action and outcome,” Moore told CMO in an interview.
“When we talk with internal stakeholders and work out how to allocate revenues, it’s about being clear on the pathway to action – what we will do differently as a result of knowing that information? It’s a powerful question for analysts and researchers.
“We use it as a lens over our program of work so we strip away any ‘nice to knows’ and focus on things that will drive to an outcome organisationally.”
Moore will be speaking about Telstra’s recent analytics efforts around improving customer services using sentiment analysis at this week’s SAS Forum in Melbourne and Sydney.
The research, insights and analytics team operates within Telstra’s CMO function and is divided into four parts: Market research; customer analytics and modelling; competitive intelligence; and data and compliance.
Over the past four years, Telstra has been building analytics assets and a program of work to support revenue-driven business outcomes, Moore said. But in the last 18 months, the telco has started extending these capabilities into customer experience and optimising services outcomes.
“In the same way we have built assets to predict customers who now have needs and requirements – our suite of cross sell, churn predicting models, which are typical of large telcos – we realised we can build models to predict how our customers currently feel about our brand,” she said.
To support these objectives, Moore’s group has built strong engagement links with IT as well as Telstra’s one-to-one marketing division, led by Nick Adams.
“This helps us move more effectively from insight generation to insight action within our one-to-one marketing team,” she said. “We also have a good roadmap and relationship with IT, making these projects much easier to run.”
NPS: The backbone of customer engagement
To kick off these customer sentiment efforts, Moore’s team developed a Net Promoter Score (NPS) attribution modelling approach. NPS is the flagship metric used right throughout Telstra to gauge customer engagement efforts and the effectiveness of its marketing activities and services.
As a result of market research undertaken each year, Telstra is able to record NPS across 30,000 customers on average. According to Moore, about 90 per cent are willing to have that score attributed to their customer record at Telstra.
What Moore’s team has now done is applied that NPS information to all customer analytics records to create an imputed NPS score for wider user base.
“Just like statistical modelling for cross-sell or customer churn, we now have NPS attribution. We can bring that into our ecosystem of targeting, so we know which customers - with about 80 per cent accuracy - are promoters, passives, detractors, and gradients of those,” she said.
“We then work with marketers to help adjust messaging in the market to align with those customer sentiments, as well as look at how to use that sort of information to craft experiences for our customers.”
To achieve this, Moore said the analytics and insights team is increasingly tapping into a wealth of customer service channel data, such as contact centre staff notes.
“We have always collected a range of customer service metrics around predicting and churn, but we’re extending that further, applying some sentiment analysis over the call centre records and expanding that capability,” she explained. “That analytics has helped us to move more into contextual customer insights. Then we’ve looked at building engagement programs using these insights to personalise services.”
Actioning customer analytics: Thanks a million
The first high- example of how sentiment analytics has been used by Telstra to improve customer experience was its highly publicised ‘Thanks a million program’, which ran last year. Using data insights, the objective of the six-month program was to empower staff to interact with customers on a personalised level and say thanks to customers for sticking with Telstra.
To ensure the program’s success, Moore’s team combined 79,000 data variables around customer service and marketing, producing an imputed NPS score. These insights were used to produce direct marketing and EDM content, as well as to inform how staff could engage directly with customers through outbound calls.
For example, Moore said detractors with an NPS score of between zero and four received communication crafted by the marketing team that “aligned with those customers currently having a less than happy time with Telstra”. Potential passives with a 5-8 score received a different EDM and direct marketing strategy, while advocates were targeted with a different EDM and direct mail again.
“It was very much a collaborative effort where we had marketing analytics, one-to-one marketing and customer experience teams from our channels area involved in shaping and driving this program,” Moore said.
‘Thanks a million’ grew into a whole-of-company exercise, with all staff involved in making outbound calls to customers. This was backed by an internal portal, developed by the marketing team, where employees could download details of customers to call, as well as detailed call scripts, all based on data rules and the value-adds available to each customer.
Key to the success of getting staff engaging with customers in the right way was how the data and analytics was presented, she continued.
“There was a range of data provided to staff including products a customer had, how long they had been a customer, and an imputed NPS score,” Moore said.
Data was essential in ensuring there was a follow up point for staff to use with customers, Moore said. For instance, in the case of Telstra’s movie tickets offer, employees could inform customers if they lived nearby a cinema that would accept the voucher.
“We did a lot of piloting to ensure what we needed to have on the portal to help staff talk to customers, leveraging off best practices we have in our call centres for outbound calls,” Moore said.
“We wanted to give enough and clarity and confidence with that call, but don’t want people overanalysing information. Getting that balance is really important. And it was about living one of the core value of Telstra’s, which is making the complex simple, which is what we tried to do.
“We weren’t exposing the 79,000 variables, just the imputed NPS score.”
Like any business-led program of work, Moore said Telstra measured every aspect of the campaign. In terms of activity, the team sent 1.9 million emails and letters.
During these interactions, employees also used forms to record any issues a customer identified, which was then passed on to an activation teams handling these issues to follow up.
“We also identified 50,000 specific customer issues that were solved, which is a very important metric for us,” she said. “We now have 50,000 customers who had an issue identified and fixed, which may or may not have been solved elsewhere. We also measured staff engagement in the program, then there was the NPS outcome on customers and the shift in perception.”
Moore’s team did all the data work to fuel the program, but were also consumers of it. “It was really exciting to provide analysts with opportunity to connect with customers on a one-to-one basis,” she said.
While Telstra anticipated some growth in terms of core business outcomes, Moore said the ‘Thanks a million’ program was all about customer service and engagement. That was also why NPS was front and centre of the corporate plan.
“NPS is the proxy for future value for the organisation: We know customers with higher NPS tend to have more products and stay longer,” she said. “In some ways, it’s a measure of future value, and how we are going in our transformation into a customer-centric organisation.”
Case study two: Checking in with customers
The second and current example of how Telstra is using data analytics to fuel customer engagement is its ‘Check-in’ initiative. The program invites all consumer and SMB customers to engage with the telco around ways to optimise their current products and services usage.
The six-month, big data-driven project stretches across in-store interactions, above-the-line radio and local advertising, SMS push notifications and EDM campaigns. The program went live in January and is expected to generate 8.7 million conversations as a result of these activities, Moore said.
Going above-the-line at a local level meant the analytics team had to ensure data on every customer’s product and service usage was on-tap to support these conversations, she said.
As was the case with the ‘Thanks a million’ campaign, decisions around what data insights to serve up to Telstra’s frontline staff and customers wasn’t something the analytics and insights team could do on its own, Moore said.
“We have to work in cross-functional teams, and work with front-of-house needs as well as marketing teams on how we’ll connect with customers,” she said. “It’s complex to make this happen. We are an important part of the process but we’re the fuel in the Ferrari; we didn’t design the Ferrari or pick the paint colour.”
To be able to action the ‘Thanks a million’ and ‘Check-in’ customer programs, Moore said Telstra had to first gain the maturity in its analytics and CRM environments.
“We had been generating these assets that then allowed us to play in non-traditional areas for the marketing team,” she said. “But we needed the capability built first, then the organisational focus around service and experience. Those two pieces coming together has allowed these programs to come to life. It was making this an internal engagement program too for staff.”
All analytics work is underpinned by Telstra’s overarching customer contact strategy and customer framework, managed by the one-to-one marketing team, she said. It is this operating model that ensures the right programs take precedence, keeping track of which campaigns are currently running, what data work is done, and what content is appropriate for which customer.
“Content needs to fit into the overall framework and strategy and how deeply we go on customer attributes – revenue target, work customers are up to in the product lifecycle and so on,” Moore said. “There is a hierarchy. We don’t want to over contact customers.”
Follow CMO on Twitter: @CMOAustralia, take part in the CMO Australia conversation on LinkedIn: CMO Australia, join us on Facebook: https://www.facebook.com/CMOAustralia, or check us out on Google+: google.com/+CmoAu