Industry is slow to make best use of data

By:
Published on:

Telecoms operators fail to build customer profiles and take action on the basis of the results

Telecoms operators are have been analysing call data records for years, but they fail to build customer profiles and take action on the basis of the results. But at the same time they need to ensure the profiles remain private

Tom Savigar

Tom Savigar: telcos don’t use the data they have about
customers and still think in silos

Pieter Knook

Pieter Knook: Vodafone 360 is incorporating popular
social networks into its services

Stéphanie Hospital

Stéphanie Hospital of Orange: we monitor what people
say about us in the blogosphere



By Jane Bird

Telecoms network operators collect large amounts of data about their customers, but are failing to exploit it fully to increase revenue and boost margins.

Telcos know where their customers are at any given time, precisely when their calls are made, and who is talking to whom. Now, with digital call data analysis, they can identify connections, spot which individuals are at the hubs of networks, and pinpoint key influencers in social groups.

Assimilating this information and acting on it fast enables customised services and smart billing to suit individual customer profiles and budgets, such as micropayments or subscriptions.

Even more opportunity to give customers what they really want can be gained by combining call data with intelligence gathered from trawling blogs, forums, and social networks such as MySpace, Facebook and Twitter.

But while some telcos are beginning to explore the opportunities, overall the industry is proving slow to wake up to the idea. “Telcos don’t use the data they have about customers and still think in silos,” says Tom Savigar, strategy and insight director at The Future Laboratory, a consultancy. “Many tariffs continue to be confusing, complex and not based on lifestyle or life stage.”

Steven Georgiadis, head of customer intelligence practice at SAS, the software company, agrees. Telcos have been analysing call data records for years, he says, but they still fail to build customer profiles and attribute values on the basis of this.

“They treat customers in isolation, which is limited,” he says. “Clearly communication involves at least two people.”

For example, customers who don’t make many calls may not appear particularly important based on the amount they spend. But they might receive a large number of calls and be at the centre of a group that is worth a lot of money overall.

Key influencer

Such people can also be relatively big influencers. A SAS study found that six times more people than average, or 30%, adopted a new handset if a key influencer in their social group had done so. Similarly, take-up for a new tariff plan was 17% more effective.

Dr Carlos Pinheiro, a telecoms researcher formerly with Brasil Telecom, has been conducting research in the area of customer relationships on behalf of Eircom in Ireland.

He too has found that people who sign up to new bundled offerings often recommend them to their friends. Follow-on subscriptions in groups where a key influencer has signed up are 130% higher than average in his research, while churn rates are reduced by 80%.

It makes sense for telcos to target individual customers who they have identified as high influencers in their social network, Pinheiro says.

“You might decide to give away handsets, or offer the first three months of a service such as international calls free, to key individuals because you know they are likely to talk to their friends and trigger diffusion.”

This is clearly the way ahead, and very different from the past, when telcos used to try to understand customer behaviour but not the relationships, says Pinheiro. “If you can protect one customer from churn, you can protect his or her whole social network.”

By examining call characteristics, telcos can also make deductions about people’s lifestyles. For example, they can look at whether an individual’s calls are mostly text or voice, the most active times of day, and whether the traffic is mostly inbound, mostly outbound, or a mixture.

This helps determine even something as basic as whether a customer is using the phone mostly for personal calls — with heavier traffic in the evening — or for running a small business, where you can expect a relatively large number of calls during the day. Telcos have long wanted to know this.

New customers

Eircom is also identifying the telecoms network operators to which its key influencers make the most calls, so that it can spot potential new customers or those it is at risk of losing.

“It’s all about working out how much money to spend keeping individual customers and which to subsidise for maximum advantage,” says Pinheiro.

Examining calling patterns and social networks has the additional benefit of helping spot fraud, he says. Brasil Telecom has achieved a huge reduction in its losses to fraudulent customers by monitoring social networks, often spotting fraudsters at the moment of the event or even before it occurs.

Its detective work exploits the fact that criminals tend to communicate by phone, collaborate in familiar groups and repeat the same pattern of activity in setting up their frauds.

Linking internal call data with external internet-based services such as email, blogs, forums and social networks offers further potential.

Mobile operators are beginning to integrate voice, text and multimedia with social networks so that customers can combine all their contacts in one place, as with Vodafone 360. Orange is also incorporating social networks into its services, such as the Dolphin package, which is targeted at young people.

Customers are ready for this type of service and are signing up fast, according to Pieter Knook, Vodafone’s director for internet services. There may be some resistance in places such as Scandinavia, where the telecoms market is mature and customers may see personalised services as intrusive.

But in South Africa, where mobile operators have leapfrogged fixed networks to offer tailored to individual customer needs, the idea is expected to prove popular.

Recommend products

It’s early days for Vodafone with 360 in Europe, “but in all our research this proposition tested extremely positively,” says Knook. Subscribers choose which social networks they want included. The hope is that they will recommend products they have bought from Vodafone, such as ringtones, on-screen wallpaper, music, games and videos, to people in their networks.

If people in your circle know that you like the latest recording by Coldplay, they might be considerably more persuaded to buy it for themselves than if it was recommended by someone they didn’t know on a public website — at least, that’s the theory. And Vodafone is confident people will choose to share their recommendations with their contacts, whether individual tracks, entire albums, or subscription services.

Stressing that their activities are restricted to anonymised data, European mobile network operators have signed an agreement to track customer browsing patterns on the internet. They are also starting to use data-mining software to evaluate people’s comments posted on social networks. They look for patterns in text, assess whether sentiment is positive or negative, and categorise conversations.

This is tricky because the word “cheap” might be good if referring to price but bad if referring to quality. “Quite nice” is positive in the US, but in the UK can often mean just OK, being a qualifier rather than an intensifier. “Wicked” is another word that can be easily misunderstood by automated software.

But such analysis gives an idea of overall brand perception and lets companies see what feedback they’re getting compared with competitors.

Orange uses these techniques to provide a monthly “barometer” of how it is perceived. “We monitor what people say about us in the blogosphere and use it internally to see how we can improve,” says Stéphanie Hospital, Orange’s vice-president for audience marketing and development.

Social media analysis is providing a qualitative layer that fits in the overall picture or looks at specific issues, says SAS’s Georgiadis. It can be useful for assessing responses to a service outage in an area or people who have had bad experiences calling a contact centre.

“Telcos can identify the one-off events that are having an impact and the overall effect of this on a large group,” says Georgiadis. “So they can decide to show contrition, or perhaps change the way they do things internally.”

Privacy for sale

To make the most of the new opportunities, telcos need customers to trust them. This is no small task. By signing up for some of the new services, customers are effectively putting “their privacy up for sale,” says Gordon Rawling, Oracle’s senior marketing director for communications in Europe, the Middle East and Africa.

People can afford to be choosy about who they let into their lives. Telcos will have to “bid for trust and a relationship” with consumers, says Rawling.

Telcos argue that they have an advantage over companies “higher in the ecosystem” such as Google or Yahoo, because they already have a relationship with customers, involving billing, data, payment handling and dispute resolution. But Simon Davies, director of Privacy International, a global watchdog, is less sure.

“The Holy Grail for telcos is to establish full control of all personal information generated through their networks,” says Davies. “But the downside of personalisation of information is if it falls into other hands, then there can be huge reputational damage.”

Fears were raised at the launch of Google Latitude, the location detector system, that people might accidentally activate it, unaware that they would be telling the world where they were. The leaking of customer records by someone in T-Mobile’s UK staff did little to encourage confidence.

Telcos need to understand the extent of their responsibilities, and the potential “Chernobyl” to their businesses if security is breached, says Davies.

Many services automatically ask for full permission from customers to collect and process all behavioural and transactional information.

The trouble is that data can fall into the wrong hands. Apart from internal fraud, it is accessed by hundreds of thousands of people from investigating authorities each year — police, local authorities, government departments and so on.

Telcos would do themselves a big favour by taking this issue more seriously, Davies says. “They should be prepared to walk the extra mile for stronger protection. It’s pure common sense for risk mitigation. If you have gold ingots of data, you need the equivalent of Fort Knox for security.” GTB