Connect the Dots With CRM

A CRM data mart will allow a view of each customer's experience.

Wireless operators already have the basic ingredients of the customer-relationship-management (CRM)-driven business, but without a single view of each customer's experience and behavior available across the organization, the benefits of CRM will stay out of reach. In a traditionally network-centric business, where do you start? It's something that service providers figured out years ago: Network-usage data contains useful information about each customer. Fraud-detection systems have searched a long time for patterns of network usage that would point to a stolen or cloned handset to report to their various departments and systems in order to take the necessary actions.

But what about all of the other possible uses of the network data generated every time a customer uses the network? What does this data contain? Could you use it to understand customers better, to improve customer care and to market new services? What could you learn about your networks if you could match customer and business data back to your call-detail records (CDRs)? To start, you could map customer profiles and network usage; map site performance against high-value subscribers; and contact customers when there are network changes, problems or improvements.

Data Volumes
There are millions or even tens of millions of CDRs generated daily within service providers' networks. They are at a very low level and, although useful in their own right, are unmanageable at the level of a specific customer. CDR data needs to be processed and aggregated to gain a meaningful level even before you begin to look for any real information. This often is done on an ad-hoc basic and is a time-consuming activity prone to inconsistent aggregation of different data sources.

This puts a heavy load on the IT/IS departments as they deal with multiple requests for similar files to be produced for different people. Worse, it allows different views of the same truth to exist in different parts of the business.

This means that two people in the organization can attempt to find the same information, but, given the vast array of data stored at different levels, in different systems, they often may use different source data for their analyses. Their results likely will not agree, even though they are trying to find out the same thing. This is bad news for the business where actions are based upon this information.

Disparate Data Sources
All of the related information that would give you more of the full customer picture (network, customer, business, value, contact history, demographic) lives in disparate systems across the business. Many of these will have been developed separately and will hold duplicate information, often in a number of different formats. Often, core systems such as billing will have been developed in isolation when the business was starting out, whereas a major warehousing project may be under way, but is rarely considered finished.

The view you really need is not available. For each subscriber, you need a single view of his customer and business data matched to aggregated network usage data, so that, for example, an information specialist can mine it, a database marketer can profile it, a marketing manager can trend it, a call-center operative can see it, and a campaign manager can segment it.

Architecture
What hugely complex architectural upheaval must a wireless company suffer to enable this? First, you need a number of sets of data:

• CDR data (network usage)

• Billing data (revenue and usage)

• Customer data (name, address, start of service)

• Product data (handset, services)

• Any additional data available (market research, demographics).

Service providers have this data already available, so what next? Take a look at the simple architecture in Figure 1. This shows the missing enabler. What you need is a single, analysis-ready view of each customer's experience, a CRM data mart. This can address each of the issues in turn.

First, the aggregation of data entering the data mart brings the data to an information level where it is analysis-ready for CRM applications.

Second, it contains a near real-time view that encompasses network, customer and business information. This is the central foundation of your CRM architecture — feeding the customer-experience information to the front-office systems, and, in turn, being fed with the updated information via those same front-office systems, back into the source systems and back into our updated customer-experience records. This is true continuous learning throughout the organization. Better than that, because it starts with the network data you already collect, it's not as difficult as you think.

Network Performance & Loyalty
The current frenzy for all things 3G is understandable, but in all of the noise, it's easy to think that customer loyalty comes from offering the latest gadget or the latest WAP game service. Over the years, these new services will have a huge influence, but what do customers rate as the most important now?

A recent MORI Poll¹ asked wireless customers what was the most important factor in their level of satisfaction with their wireless providers. The top four were:

• 88%: Better network coverage

• 84%: Better call connections

• 81%: Better information on network coverage

• 76%: Fewer dropped calls.

So, it's network, network, network. Customers want better performance from the network (less drops, less blocks, high quality with maximum coverage), not free gifts, loyalty clubs, redesigned bills. They just want a phone that works well when they use it.

On top of network performance, customers actually want to hear about network coverage:

• "Yes, there is a problem near your home. We are working on it."

• "Sorry about the problem you had last week. It's now resolved."

• "We think your handset is faulty. Pop in and we'll check it for you."

• "Good news. We have substantially increased capacity in your area."

• "Good news. Coverage in the area you ring most often has been improved."

If this is what customers want to be more satisfied, then this is what we must provide. Without usable network-usage data matched to the rest of the customer's information, we will not be able to deliver any of this.

Network Data in Action
Consider using a CRM data mart to support a major initiative to manage faulty handsets. (See Figure 2.)Research shows that up to 20% of handsets may develop a fault in any one year — worse than that, subscribers automatically blame the network.

By looking across the range of data generated through your network, you can see many things that help identify likely handset faults, for example:

• Customer-service call activity

• Fault codes captured within CDR data

• Background performance of the cell-site

• Mobility patterns

• Dropped-call levels, which are measured against other customers' experience.

There are many pro-active steps to take from phoning the customer or writing to him to sending him an SMS: "Pop in and see us; we think there is something wrong with your phone."

Maximum Knowledge From Combined Data
If you then map value across those same customers, you can tailor your actions depending on the worth of the customers. For low-value customers, you might just offer to check the phone at any shop, whereas for your top-value customers, you might want to offer a free handset-replacement service.

You are now relating to your customers and their needs using a sophisticated combination of information about the customer, derived information such as value-segmentation models and actual network performance as experienced by the customer.

---

¹ A poll conducted by MORI on a representative sample of high-using business customers is the first of a series of five international surveys sponsored by MSI.

---

Data-Mart Attributes

A data mart should:

• Deliver a single-experience view for each customer
• Be embedded into front-office applications
• Deliver a common picture to all customer-contact points
• Be available across the entire organization
• Be aggregated to a level suitable for CRM analysis
• Contain near-real-time data
• Be built from customer, business and network data
• Accurately reflect the actual network traffic for that customer
• Be available from a single system.

If all of the above can be applied to your operation, you will lead the wireless market in managing customer relationships, retaining customers and maximizing immediate and sustainable revenue streams.

O'Brien (sobrien@msi-world.com) is Metapath Software International senior CRM consultant.