HKT partnership with Lynx Analytics helped them identify ‘silent complainers’ and be alerted on moments of network experience that trigger a potential impact on churn and ARPU.

Predicting Silent Network Complainers

About HKT

Hong Kong Telecom (HKT) is the largest mobile network provider in Hong Kong, with 3.24 million mobile subscribers relying on its 3G and 4G networks for daily communications and mobile entertainment services including the Group’s very successful NowTV and ViuTV, Pay TV and Free TV services.

Challenges of HKT

Since 2016, HKT has been partnering Lynx Analytics to develop a new solution to not only identify ‘silent complainers’ but be alerted on moments of network experience that trigger a potential impact on churn, ARPU or other customer outcomes.

4% of their customer base had a Network Experience sub-index score of 2 or less out of 10. A low score leads to a very high chance of these customers complaining and eventually churning – with a risk factor of 3.5x higher than the rest of the base.

How did Lynx CHI solve the problem?

Affectionately named the Customer Happiness Index, this analytics tool aggregates data sets from across the business: product, network, billing, customer service, point-of sales, etc. It allows them to map out their entire customer base on a predictive scale of 0 to 10, where customers getting close to 10 are customers that are the least likely to churn and most likely to increase their ARPU.

Network Experience is a key sub-index and is actively managed by HKT’s network teams and their objective is to keep the score at a healthy level above 5. The Network Experience shows the impact an individual’s network experience – data speed, wifi usage, etc – has on his or her ARPU, NPS score or propensity to churn.

Network teams used the Customer Happiness Index solution to identify that one ingredient – ‘high round- trip time for data users’ as the main root cause for the low score and required immediate attention.

Traditionally, the network team would look at complaint data received from Customer Service. However these are few and far in between (<0.33% of those estimated to be affected) and do not provide a sufficient sample size to represent the true causes affecting these customers.

By looking at the distribution of the network sub-index scores (pictured below) the network team identified a threshold score of 1.6, below which the subscribers are observed to be experiencing dissatisfaction with the network. This meant that, despite relatively few network complaints, nearly 4% of HKT’s customer base were at risk of churning due to poor network experiences. What’s more: the Customer Happiness Index also allows the teams to create mirco-segments by location, demographics and spend, and drill down to segment-of-one thus enabling them to pinpoint individual customers, who require pre-emptive care.


Screen Shot 2019-08-28 at 2.12.55 PM

They initiated a pro-active targeted campaign that offered these customers a device upgrade, so they can make the most out of the network or inform them how to optimise their device for a better network experience.

As a result of the proactive campaign, they were able to increase the score for 35% of these silent complainer subscriber segments from less than 2 to more than 6 in just three months, thereby eliminating their need to complain and significantly reducing their need to churn in the long run. This delivered an immediate and tangible saving of 14% of their customer service operating costs within 3 months Now they are applying the same approach across other ingredients to realise significantly higher savings as well as long term revenue uplift.

This is just one example of the use of the Happiness Index within the organisation.


Screen Shot 2019-08-28 at 2.14.51 PM


The Outcome

The Customer Happiness Index has become the de facto decision-making index for customer experience optimization across the HKT organization.

Powered by Lynx Analytics’ proprietary analytics engine, it crunches daily data and aggregates different data points across time. This allows the operator to create a limited but actionable set of ‘experience ingredients’ – such as dropped calls, mobile data entitlement, roaming experience, data throughput experience etc.

The solution then aggregates these experience ingredients into operational components: Network Experience, Customer Service Experience, Omnichannel Experience and more, so the internal functions at the MNO can own each component.

These components are combined into an individual customer score that predicts their propensity to Churn, increase ARPU and NPS score.

Armed with this information, the operator can take action to address which network issues truly impact customers before it is too late.

Beyond network experiences HKT has deployed the Customer Happiness Index throughout the entire customer experience journey. Across the organization diverse teams from customer ops, loyalty programs, to retail and digital use the Customer Happiness Index to solve underlying customer challenges that impact top-line results.

Applied across its 3.24M mobile subscribers we observe a steep decline in customer churn rates until a happiness score of 5. In fact by migrating customers with an average score of 4 to 5, we can retain roughly half of the churners with the original score.

Our experience indicates that once a customer is successfully retained, they tend to spend more as well. By moving subscribers at a ‘happiness score’ of 4 to 5, HKT estimates a potential doubling in ARPU for these subscribers (see illustration right). Even better the upswing continues as the customer’s score improves over time, creating higher lifetime customer value.


Screen Shot 2019-08-28 at 2.17.35 PM


Screen Shot 2019-08-28 at 2.17.40 PM

Interested in exploring the Customer Happiness Index for your business?

Request a Demo