Mobile Operators Finally Get Segmentation Right
By Steve Wallage, Tue Dec 07 08:30:00 GMT 2004

You might expect mobile operators to be on the vanguard in customer analysis. In fact, though, other industries, such as retail, have far better insight into why their customers behave in certain ways and how they may behave in the future.


Mobile operators should surely be good — in fact, very good — at customer segmentation and analytics. Why? Let's add it all up: First, they have an incredible amount of highly detailed information on their customers — namely, with whom, when and how they communicate. It's like supermarkets knowing not just what people buy, but when and how they cook. Second, they have a pressing need for a strong understanding of their customers. Not only is reducing churn a key issue, but think of the importance of knowing the identity of the most profitable and influential customers. Furthermore, mobile operators need to understand how particular customers will respond to a whole raft of new services they are introducing. Third, they have very few products to worry about (particularly when compared to industries like retail).

So Much For Theory, What's the Reality?

Stefan Tobiasson, CRM Strategic Manager at TeliaSonera, believes that mobile operators are often satisfied if they can understand what a customer has historically done, not tending to care about the reasoning behind this data, nor about understanding their future behavior.

Many mobile operators still rely heavily on the marketing department to drive customer segmentation and use just four to six customer segments, based only on such data as demographics, usage and spending, according to Jouko Ahvenainen, chairman of Xtract, a company specialized in customer analytics.

Michael Hulme, founder and non-executive chairman of Teleconomy, a research-based consultancy specialized in understanding consumer behavior, is even more scathing. In his view, supermarkets have developed the expertise to target services based on a user at a particular place, at a particular time with a particular background. This well understood and tested approach uses "spider's webs" charts to map the customer's likely behavior. It can be seen in all the thought that goes into a supermarket — the placing of products, the music, the ambience, the store layout and even the smell. Real-time customer analytics allow supermarkets to print, at the checkout, promotional coupons based on what the customer has just bought. By comparison, Hulme says, the mobile industry has hardly started on the ladder towards understanding their customers.

Why this reluctance among mobile operators to use customer segmentation and analysis? The answer seems twofold: no sense of urgency and no market precedence. First, incredible mobile market growth has meant there is less pressure than a mature market to utilize this type of information. Second, because there has been less pressure, there has not been a culture of using such information. Mobile marketing departments have been driven by tactical issues and advertising campaigns. The idea of having a group of information analysts dedicated to crunching modelling customer data has just not taken off.

What Are the Best Operators Doing?

Change is in the air, however, as market growth slows and competition intensifies. Xtract, which also serves the retail and finance sectors, is finding greater demand for its services among leading-edge mobile operators.

Behavioral-based analytics help these operators understand their customer's calling and messaging usage, according to Xtract's Ahvenainen. Predictive analysis aids in understanding which customers will buy which products, what influences them and how much influence they have over others. Call center agents can then use this tool to sell customized offers to customers. The Holy Grail in all this is to discover which individuals are at the center of their social networks, and thus exert the most influence on others. These key individuals can then be offered the latest services at a discounted price, as others in their social networks will then likely follow their lead.

A good example is one Xtract customer, Finland's Elisa Mobile (formerly Radiolinja), which uses behavior-based segmentation to tailor marketing campaigns and target upgrade offers. Behavioral analysis allows it to better predict which "Best Next Offerξ will be most attractive to particular users.

This is all a far cry from the sorry state of many operators' customer segmentation. Even so, Elisa typically only typecasts users in four to six headline customer segments — such as "pioneers", "explorers" and "emergencies". The Xtract system, however, encourages the creation of potentially hundreds of sub-segments based on increasingly granular data on customer behavior — not just in their mobile environment, but also in their lives.

Frontrunner TeliaSonera uses customer segmentation and analytics to understand the person behind the device and details including their age, home life, lifestyle and usage of IT, as well as views, attitudes and behavior across a wide range of areas. Using data provided by groups such as Experian, as well as statistical modeling tools from vendors like Alterian, TeliaSonera aims to collect and use up to 400-500 pieces of information per subscriber.

With this kind of knowledge at its fingertips, TeliaSonera has been able to reduce churn by predicting when and why customers may move; target specific groups of prospects, including those using a competing mobile provider; and anticipate the likely uptake of MMS among users. It has also been able to assess the likely buying triggers — for example, is it price, functionality or simply being able to do something before anyone else.

TeliaSonera's highly detailed customer databases also include important financial information such as a Net Present Value (NPV — future spending flows) per customer. Using this combination of data enables the company to develop very sophisticated telemarketing and advertising campaigns.

What Are the Lessons To Be Learned?

It works! That's the feedback from mobile operators who are the early users of detailed customer segmentation and analytics. Even so, the most advanced mobile operator can still go a long way in maximizing customer data.

First, look at the big picture. Ahvenainen believes mobile operators need to know how their customers behave in similar markets such as IT and fixed telecoms. TeliaSonera's Tobiasson also says that consumer spending on mobiles can't be viewed in isolation, but rather in the context of their entire disposable income. He cites the music industry as an example: everyone talks about how that industry's woes were brought about by the Internet and free music downloads, when sales were already down significantly before Napster came on the scene, as people moved their disposable income to — ironically enough — mobile phones and services.

A persistent criticism of customer data analysis is that it is historical data and therefore cannot be used to predict future behavior with any accuracy. This is where knowing the big picture comes in. TeliaSonera has found little value in identifying, say, high-spending customers — unless there is a similar effort to understand this segment's lifestyles and wider spending patterns. Otherwise, the information could prove to be totally misleading for example, is it their company paying for the spending or it is abnormal spending that will soon return to average levels?

Second, focus on the business model and return on investment. Too many mobile operators gather the customer data, but are not sure what to do with it, or else have only historical data rather than real-time information. In fact, Gartner Group research found that up to 75% of Customer Relationship Management (CRM) implementations fail to meet expectations. However, TeliaSonera has proven that its customer segmentation work can reduce churn. Xtract has mobile customers who have clearly found a revenue boost from their investment.

Third, make the most of what information you do get. Tobiasson has found, perhaps to no cynic's surprise, that there can be great difference between what people say they will do and what they actually do in practice. Additionally, finding out their views on a wide range of subjects is not an easy, or cheap thing to do. However, the value of such data makes the effort worthwhile.

Fourth, learn from other industries. The mobile industry should be leading the way in customer segmentation and analytics, not lagging behind. Be that as it may, mobile operators can benefit from being fast followers and take the best ideas from other industries. They can borrow winning concepts from the retail sector and experiment with loyalty cards or highly personalized offers. Or operators could simply adopt the best statistical models for predicting customer behavior and implement the most effective ways to sub-segment the market. One more contentious idea is that the operators could sell the valuable raw data they have on customers to other organizations to use f or their own segmentation schemes.

Whatever their choices, mobile operators can only get better at customer segmentation and analytics. The danger is that it is easy to waste money in this area by developing models and databases that have little real impact on revenues. However, if used correctly, such analysis can have a major impact on future growth. The real measure of success will be if, in two to three years' time, retailers are looking at the mobile industry to steal customer segmentation strategies.