The recent “text to switch” regulation, introduced by Ofcom in the UK, is another factor telecommunications companies need to take into account when planning their customer strategy. Customers can now leave their mobile provider by simply sending a text message – avoiding the hassle of calling a call centre – and eliminating the chances for last minute ‘rescue’.
On a market threatened by value erosion and defined by blurring boundaries, telecoms are struggling to stay relevant and find new sources of growth. One of the big questions to answer is how to keep existing customers and increase their lifetime value when prices are dropping, services are a commodity and consumers have an expanding pool of offerings to choose from.
The obvious answer would be – become even more attuned to customers‘ needs.
The problem is, however, that not a lot of customers will answer your surveys. Customer satisfaction prompts have a traditionally low response rate – between 5% and 20% for different industries. In our experience with telco, around 10% of customers bother replying when prompted. If you apply this rate to your whole customer base, that means that out of every 100 customers there are:
- 10 to 20 customers that are either very dissatisfied and in high risk of leaving
- 10 to 20 customers that are very happy and will be willing to recommend you to friends and bring you additional business
The big problem however? You have no way of knowing who exactly those customers are, so you can tailor your communication efforts to their needs. And when switching your operator is easier than ever in the UK, you are left without your last chance to make amends and decrease churn.
There is one way to anticipate better customer needs – predictive analytics.
It’s accurate because it’s based not on what customers tell you when asked, but what they do and the experiences they’ve had with your brand and product. Based on internal data about usage patterns, interactions with customer service, price sensitivity, we can build a machine learning model that creates a profile for each customer and predicts their satisfaction levels.
In practice, this means that you will know where up to 90 out of every 100 of your customers stand. Opportunities for high-impact individual-level action from there are boundless.
A good start, however, would be to call up dissatisfied customers before they’ve had a chance to type up “PAC” and press “Send”.
Check out how a telco reduced churn by 34% using our predictive model. Learn more about GemSeek’s predictive analytics for customers or if we have gotten you thinking, drop a message to firstname.lastname@example.org.