August 2003

By Bipin K Paracha and Anupama Bulusu, Integrhythm Inc.

The Problem
Consider a very common case that customer service centers in companies often confront. A customer purchases an item from a company as a result of a telephone marketing campaign and the item is shipped. The customer, however, feels dissatisfied with the product and would like a replacement. The marketing department sees this customer as someone to whom they successfully sold a product. The order processing and shipping department sees this customer as someone who should be satisfied because the order was filled, shipped and delivered on time. The customer service department sees a dissatisfied customer asking for a replacement, and a customer service request that has not yet been resolved. If the three departments have disparate systems with discrete data points, each department is unaware of the other department's "snapshot" of the customer. In such as case, a week after the first time it made the sale to this customer, the marketing department may do some profiling, and find out that in the past, customers who bought this product are also likely to buy a service that the company offers. They call the customer with a new promotion or offer, but the customer is unhappy with the company for calling him with other offers when his complaint with the product he has already purchased is pending. The company projects an image of chaos and disorganization, and the customer does not receive satisfaction in his or her dealings with the company. The company may potentially lose a customer who could have proved to have a valuable lifetime value because of a lack of judgment. Had the marketing department known that the customer had a pending customer service request, they may have waited until it was resolved satisfactorily and then tried to pitch the service to him or her, thereby increasing the possibility of a successful reception to their sales efforts. In this case, the fact that the marketing department did not have data from the customer service department caused the sales effort to fail, when better communication between the departments could have prevented this.

The Solution
How can a company solve such problems? Frequently, departments within a company need to "trade" data about customers, and not having the capability to easily do so can seriously limit them in performing efficiently. A unified view of customer relationships allows the company to provide each customer with a consistent approach to his or her problems, and the ability to offer solutions no matter which department in the company is contacted by the customer and through which medium. This increases cross-sell and upsell opportunities for the company and increases organizational efficiency. It also helps the company with better business intelligence possibilities, in terms of segmenting clients more competently and recognizing trends and profiles more efficiently. Not only does this allow the company to have a unified view of the customer, but it also helps the customer perceive the company as a single entity as opposed to an assortment of several departments. This increases the customer's satisfaction with the company and the probability of a future relationship.

Strategies
Companies have long realized that this organization-wide view of customer relationships is necessary and lacking, and there are several industry giants trying to solve this problem. Several strategies have been proposed, but few have gained prominence or garnered respect as authoritative solutions. What are leaders in the CRM industry doing about this?

One solution may be to use a single system for all the applications. A single application for all the systems in the office is not feasible, as no single application can manage everything from human resources to customer service management. Additionally, this solution does not leverage investments in existing systems and applications. It is impractical to throw away all the existing systems and launch a mammoth project to replace all of them with a single system.

Logically, integrating the data in all these applications without replacing the systems (or at least the ones that contain data essential to provide a unified view of the customers) has emerged as a practical solution to this problem.

So where can all the integrated data be stored? Can one application be chosen as the "master" application and all other data merged into it? Maybe'but this approach can lead to performance, scalability and security issues and can cause huge maintenance problems. Is it a better approach to create a common data repository where all data are integrated? This would create a central storage area for the integrated data in the entire enterprise, and performance and security issues can be managed with more ease. Scalability would depend on how well the data model for the repository is planned, designed and constructed.

Shortcomings
Some common questions that come up in discussions of a unified view of the customer are:

• Can the data be integrated real time?
• Does there have to be a lag time?
• If there is a lag time, what is the freshness of data and how can accuracy be ensured?
• What is the time needed for incremental loads?
• If data is integrated in real time, what happens to errant data?

In the last question, certain records may be rejected by the system due to missing information or certain unsatisfied rules. In such cases, there has to be a robust error-handling mechanism and a way to temporarily store these rejected records so they can be rectified at a later stage. It is also important to ask in a real-time system how the updates and inserts to the system are handled and what the response times are to such requests. The response times need to be in the order of microseconds. It may be counter-productive if the end users must spend a lot of time waiting for the records to be updated while entering data. What kind of middleware solution is used by the system to perform the real-time integration becomes important when answering this question.

In both cases, whether the data integration happens real-time or during incremental loads at off-peak times, an important issue to be addressed is: how are the records from various systems correlated and how is data duplication avoided? Each system may have its own method of identifying a customer, and correlating the records belonging to that particular customer across all the applications may be a difficult and daunting task. All the integrated data exists as discrete data in the individual applications, so data are already duplicated. But if there are several different versions of data regarding the same customer, then unifying the data has produced negative results and compounded the problem instead of resolving it! Therefore, within the integrated data, duplication needs to be avoided.

How reliable will the software used to integrate the data be, not just in terms of the expected or unexpected "down time," but also in terms of the accuracy of data? How are scalability and security issues addressed? If the solution is indeed long-term, it needs to be able to handle increasing volumes of data per year and also needs to be able to handle demands on functionality and efficiency. An internal study conducted by the company to project reasonable data volume surges and its relation to enterprise growth might be useful in making this decision. For example, the software may boast 90 percent reliability, but if your company deals with 10 million customer records a year, this translates to one million inaccurate records. If the company projects that every year its growth will cause the data volume to increase by one million records a year, the problem is compounded.

Most companies wish to provide customer service for 12 hours or more per day, and usually also provide service on the weekends. This raises an important question: what will the hardware requirements be for a system that is needed for 24/7 operations and what is the cost of implementation? Most decisions are made with the bottom line in view. If it costs a lot to implement the solution and then maintain and use it, and the reliability is less than ideal, does it make sense for the company to still implement it? Would it be better for the company to perform without a unified view of the customer and hope to maintain profitability by dealing with customers as best as they can? Time cost and scope accounting is very valuable in the implementation of huge projects, and whether there is a return on investment on such a capital outlay should be weighed when making the decision.

There is very little doubt that a truly unified view of the customer can have a lot of advantages to a company, and make contacts more satisfactory for its customers. But achieving a truly unified view is a complex and protracted problem. For the unified view to be accurate and reliable, several things have to be in place. The data in the individual applications need to be reliable. Business needs and strategies must be clearly defined and mandated. Also, it is very important to find the strategy/product that solves the company's individual requirements and issues instead of tailoring the needs and issues as an excuse to buy the leading product. Company executives need to have a realistic understanding of the expenses and issues involved. End users need to be trained and trained often.

The ultimate reward is that the company that achieves a completely unified view of the customer can be a market leader in terms of quality of service and organizational efficiency.

Bipin Paracha and Anupama Bulusu are senior consultants at Integrhythm Inc. Integrhythm provides consulting services for CRM implementation and integration.

A Unified View: The Holy Grail
By Andrew Shepard, Getronics

Capturing a 360-degree view of the customer is considered the Holy Grail in banking. Today's financial institutions understand that future business will come from nurturing and maximizing existing relationships. Yet, while financial institutions have a huge amount of customer data, this information is often segmented and stored within various divisions that don't communicate. A host of software has been developed to capture a unified customer view, yet institutions should spend more time determining what they're looking for before the IT department gets involved in the crusade.

Financial institutions lag behind many other industries in understanding their customers. Ironically, they have more information at their fingertips than any other industry -- they just don't maximize it. When a customer applies for a mortgage, think about all of the valuable data the bank captures: assets, debts, net worth, credit history. The bank knows the make, model and age of the customer's car, the lien holder and the monthly payment amount. It knows the customer's investments, and even more important, potential investment-worthy assets. In short, the bank knows almost everything about the customer from a financial perspective.

Why aren't financial institutions using these data to boost customer satisfaction levels and anticipate customer needs to sell their full portfolio of services, from insurance to brokerage and investment services in addition to traditional banking services? The reasons are common: it's expensive and time-consuming to figure out how to share information among autonomous business units, and the financial world's culture can be resistant to change. However, if managed properly, most financial institutions are sitting on a treasure trove of underused data. Getting answers to the right questions can help financial firms to achieve the holy grail of a unified customer view.

1. Who is the customer, anyway?

The first step in maximizing customer data is to understand not only who the customer is, but also whose customer it is. In the financial arena, the customer might be the account holder, the head of the household, the business executive or any number of stakeholders.

Determining whose customer it is gets more complex: If a customer has a brokerage account with one bank division and a mortgage with another, who owns that customer? It's an easy question to answer when the units operate autonomously -- both divisions own the customer. But when data are shared among the institution's different silos, political issues and competing interests can get in the way. Imagine a loan officer advising a customer to use money from his brokerage account as a down payment for his desired car or yacht -- and imagine the brokerage manager's reaction when he finds the account, and subsequent commissions, are reduced. Competing interests are driven by the highly evolved realities of bottom-line commission.

That is the risk of having a unified view of the customer. Financial institutions must avoid shortsighted gains to achieve long-term profitability. In order to avoid tripping over each other, it's important to insure that one person has accountability for the customer and can manage the relationship to the benefit of both.

2. What do we want to know?

In the banking world, there is a big discrepancy on how much information is needed to develop a customer profile - a 360-degree view means different things to different people. It is all too easy to overwhelm too many people with too much data. Before the IS team can capture data, each department must determine exactly what it wants to know about the customer.

For example, tellers achieve high levels of customer satisfaction by making sure their lines move fast. They don't want to wade through multiple screens of data about customers before they can find the information they need to conduct a transaction. Some banks that tout a higher level of customer service may want to provide tellers with information that enables them to personalize the customer experience, such as total loans, total deposits, attrition risk, market segmentation and history with the institution.

Yet, capturing a full 360-degree view of the customer means going beyond these basics. Data needed for a full customer profile includes demographic information on income, family members and related business accounts, contact history, competitive products the customer is using at alternate institutions and information on debt and investment-worthy assets.

The key to managing the varying levels of information is data integration. Most financial institutions' data is separated into silos: traditional banking areas like lending and cash management don't share information with extended service areas like brokerage and insurance.

In addition to determining who has access to data, banks must decide how often they will refresh it. Real-time information is ideal, but very costly. Firms must balance the need for data quality and validity with the cost of delivering it.

3) Can we share what we gather?

There are a lot of unsettled dynamics in terms of privacy of financial data. The Financial Modernization Act of 1999 broke down the rules of banking and enabled financial institutions to offer integrated banking, insurance, brokerage and other financial services. The Act also required all financial service providers to allow customers to opt-out of enabling them to share data with external firms. However, pending legislation, including the renewal of the Fair Crediting Reporting Act, has the potential to change the rules.

Today, the rules are fairly flexible, as long as the customer hasn't specifically informed the financial institution otherwise, it may share customer data among its own divisions and with external firms. However, it's wise to document and segment customer data so that if the rules change, the institution is prepared. Capture data by division ' mortgage, brokerage, banking, etc. -- and document where each piece originated. It's acceptable to combine the pieces into one integrated database to share among divisions, but make sure that should the rules change, you can pull the pieces back out.

More to the point, it's a good time to start marketing the value of integrated data-sharing to customers. Most countries use the opt-in system, in which consumers must agree before financial institutions share their data. The United States most likely won't move to an opt-in system anytime soon, but it makes sense to start communicating the upside benefits: the ability to develop and offer customized financial services and help customers become more financially savvy.

4. How will we use the information?

Capturing customer data and making the information available to the right people at the right time is just the first step. The driving force behind capturing a unified view of the customer is to increase customers' satisfaction with the financial institution and, therefore, increase their profitability to the bank. In other words, if the customer has a good relationship with the bank, it should make sure it understands and acts on the value of that.

The goal is to use the captured data so the appropriate people within the institution know the customer well enough to offer a logical and tailored package of financial services that meet the customer's particular needs. The bank must use the data to maintain customer satisfaction levels and anticipate their needs. This will drive customer loyalty and, therefore, continue to deliver ongoing revenue to the bank.

After all, those are the true benefits behind the promise of the unified customer Holy Grail: customized, actionable services for clients, and continued profitable relationships for the financial institution. It's time to start the crusade!

Andrew Shepard is a consultant with Getronics Business Solutions & Consulting. With an emphasis in financial services, he assists Getronics' clients with addressing business issues through critical analysis and practical recommendations. His areas of expertise within the financial services arena include business process reengineering, strategic planning and customer relationship management. Shepard has 12 years of experience in the financial services industry.