Operationalizing Customer Intelligence In The Contact CenterOperationalizing Customer Intelligence In The Contact Center

This article originally appeared in the December 2007 issue of Business Communications Review magazine.

Most customer service executives spend a considerable amount of time thinking about the need to change their contact center from a cost/service-oriented operation to one oriented toward profit/loyalty. A key capability in this migration will be the center’s ability to act on customer intelligence.

Computer Sciences Corporation recently surveyed more than 750 Fortune 1000 executives regarding their customer intelligence (CI) efforts and found operationalizing customer insights to consistently be the weakest of the three CI capabilities (customer information integration, customer insights, and customer insight operationalization.):

* 69 percent of respondents have not yet made customer insights available to their customer facing personnel.

* 54 percent do not differentiate service to their “best” customers.

* 70 percent do not have defined business rules in place when customers present a supportable need outside traditional sales and service situations.

Most contact center managers still repeat the same tired metrics to their executives: service level, average handle time (AHT), number of interactions and abandon rate. No wonder senior management yawns and acts like customer service is a cost center.

This article describes the actions that should be taken to prepare for CI operationalization, and the initiatives customer service executives should champion to ensure success in CI execution. This will enable the contact center manager to discuss metrics like customer defection saves, service-to-sales conversions, and cross- and up-sell ratios with management, and talk about what the center is doing to increase the loyalty of the company’s most valuable customers. Preliminaries

Before plunging directly into the operationalization of CI, it will be useful to review a few basics of the subject, starting with a definition of customer intelligence:

Customer intelligence is the ability to use information about a customer to foster deeper relationships that generate greater profitability. It is also the basis for targeting new customers based on matching key attributes of best customers.

Customer intelligence is a multi-dimensional business problem:

1. Customer information integration —Is your data clean and trusted by users? —Have you compiled a 360-degree view of the customer?

2. Customer insights: Segmentation/modeling —Do you know the preferences, needs and value of each customer? —Can you predict the likelihood of a customer defecting, buying or responding?

3. Customer insight operationalization —Have you integrated analytic insights into front-office applications? —Are you able to customize treatments to the individual customer?

Company capabilities generally fall into one of four progressively higher rankings within each of these CI dimensions.

* Basic—No 360-degree view of the customer, excessive manual analysis

* Foundational—Common customer ID, customer segmentation

* Advanced—Complete customer view, treatments driven by value segments

* Distinctive—Differentiated products and services by segment, insight-driven interactions

Together, the CI dimensions and capabilities rankings form a useful diagnostic model a manager can use to assess where they are and where they want to be in developing meaningful relationships, relevant offers and proactive services.

This is shown in the Figure below:

CUSTOMER INTELLIGENCE MATURITY MODEL

Operationalization Of CI

Operationalizing customer intelligence occurs at many points in the corporation: The CEO needs it to make strategic decisions; product development needs it for new product design; marketing needs it for data mining and modeling; finance needs it to understand customer profitability; and IT needs it for data provisioning and report production.

Similarly, the contact center needs customer insights in order to present a customer interaction experience that is consistent with corporate marketing and communication messages and business revenue and profitability goals. So what does the contact center manager need to do to take advantage of corporate customer knowledge? Key activities are:

* Identify who is interacting with the corporation

* Get the interaction to the resource best able to handle it

* Execute specialized treatment on each individual interaction

* Capture the center’s results relating to increased sales and loyalty

* Leverage the contact center’s close connection with the customer to provide input into the CI analytical process

Who’s Calling?: Operational

The first issue is identifying who is calling, chatting, emailing or faxing. Before any special treatments can be applied, the contact center has to identify the customer. Optimally, the contact center wants to know the individual, and if applicable, the business/company they represent.

The point of customer identification is to begin applying segmented treatment by routing the interaction to the most appropriate resource the first time. For example, a catalog shopping contact center identifies customers with known credit risk and sends them to credit and collections before they can order new products. The success a company has in identifying its customers is only as good as the accuracy of its customer databases.

If there are data integrity questions, the contact center can add significant value by offering contact information validation services. Most people are familiar with a customer service representative (CSR) reviewing name, address and phone number information before proceeding with the interaction. What they are doing is improving the accuracy of their customer database. Operationally, these front-line employees should be empowered to make changes in the customer record that facilitates identification. CSRs should have administrative rights to make changes, and average handle time expectations need to be adjusted to accommodate the increased time required to support the contact information validation process.

Customer Identification: Technical

From a technical view, most contact centers use automated attendant technology (e.g., a drop-down box in a chat session or formatted email) to identify customer need. For example, operating system questions go to one group and application questions go to another. Sales goes to one group and service to another.

The technical difference between need-based routing and customer identification routing is the intelligence of the routing engine. Need-based routing uses a simple automated attendant (for voice) or drop-down boxes (for email or chat) with no connectivity to systems outside the routing environment. Customer identification routing requires interactive routing devices that can interact with host databases in a sophisticated, rules-based manner.

In the contact center, customer identification is most often accomplished via:

* Automatic Number Identification (ANI)—which provides the phone number the customer is calling from.

* Caller Entered Digits (CED)—Asking for identifying information in the interactive voice response (IVR) unit, formatted email or chat session.

* From: e-mail address—The originating address in an e-mail.

Most contact centers use all of these tools to accurately identify who is interacting with them. Some industries and business transactions can successfully identify a majority of their customers using ANI. Many cannot because customers may call from their office or cell phones not listed in their customer record. Using an IVR to request that the customer enter account or other identifying information via touch tone or speech recognition (or similarly requesting identification information in a chat session or email) significantly improves the recognition rate.

These customer recognition techniques require computer-telephony integration (CTI) technology, which will allow the customer information collected via the IVR/chat/email to be mapped to customer databases and then returned to the routing platform for intelligent handling. It is strongly recommended that any contact center anticipating CI initiatives have CTI in place. CTI services can be used for many other applications, such as coordinating the quality monitoring sessions of a QMS system or tracking of phone time in a workforce management system (WFM), so the investment in CTI is worthwhile even outside the CI requirements.

A big problem can occur when trying to operationalize CI before significant progress on customer information integration has been achieved. Often, in this stage of CI maturity, customer insight analysis is performed on data collected in an ad hoc manner, based on databases not readily available to the contact center.

The resulting customer insights are, while interesting, not easily actionable by the contact center if a targeted customer cannot be identified when they interact with the company. Very often, the results of a customer insight analysis will be descriptive parameters: age, income, buying behavior and position in customer lifecycle.

With a 360-degree view of the customer, all the databases used to develop this description would be available in a contact center business rules engine to classify an incoming caller. In the absence of the 360-degree view, the manager should request a mapping of these descriptive statistics into data points available in contact center databases.

Having the CI analysts hand the contact center a list of targeted accounts is not an optimal approach to customer identification. At worst, it requires manual identification and at best, it requires the maintenance of a static customer list, often outside the main customer database. It is much better to have target descriptor data points and let the business rules engine manage the selection of targeted customers.

Contact Routing: Operational

Once the identity of the customer has been established, customer insights drive the routing of some of the segments identified as requiring specialized handling. Credit risks (also known as deadbeats) are sent to credit and collections, regardless of what they might really want. The company will not sell to them until they pay on their bill. On the other end, very high-value customers are sent to a “high touch” agent group skilled to handle all the unique needs of their segment. In both cases, questions in the IVR/chat relating to needed resource skill can be bypassed, saving toll-free charges from deadbeats’ calls, and avoiding any hassle for the high-value customers.

An example of routing to credit and collections has already been provided. On the high-value customer side, an electric utility, through customer intelligence, discovered that while their very large industrial customers did drive revenue and kilowatt hours, and residential customers drove call volume into the service centers, it was the middle market customers that drove profit—the entrepreneur that owns five fast food restaurants, the mom and pop grocery stores, the car dealerships, for example.

These customers had very similar need profiles: quick turnaround for service turn-on and turn-offs, security lighting, how to reduce their bill and quick response to outages. There was very little in the way of credit and collections or bill dispute interactions.

These customers were identified and their interactions went to a specialized group of highly trained agents, some of whom were newly-minted electrical engineers, trained to handle all issues of that segment. This special treatment, coincidentally, also had the impact of reducing regulatory complaints since these mid-market business people are often well connected to Public Utility Commission commissioners and their staff.

Contact Routing: Technical

The generic term for the ability to intelligently route interactions is “skills based routing” (SBR). SBR is available on almost all ACD, email, chat, etc., systems as part of the standard packages. (It is best when all interaction channels are integrated and managed by a single routing rules engine).

Robust routing includes a capability called “variable routing.” This means that rather than providing a specific set of routing parameters, almost any data point can be identified as a routing variable. For example, customer intelligence can use demographic data, past purchase history, or any other defining variable to map to a specific treatment. Data points can be within any piece of the interaction-handling platform or from host connections.

Skills based routing is typically associated with ANI, CED from the IVR, from: e-mail address or chat identifier; SBR attempts to match customer need with appropriate skill. Host connectivity is most often used to simply “pop” a screen of customer information to the agent upon delivery of the interaction. Data points from the customer record are rarely used to specify what data to pop, and therefore are not often used as a routing variable, primarily because no customer intelligence exists to make these decisions. However, if host connectivity is already part of the routing platform, any customer data point can be used for pop and routing purposes.

Scripting And The CRM Desktop: Operational

With rare exceptions, the front-line CSRs working in the contact center are not professional salespeople. They will require specific training and desk aids to successfully convert CI into increased sales or customer satisfaction. The best approach to converting customer intelligence into an actionable script for the contact center CSRs is a team approach. The team should be made up of front-line CSRs, contact center analysts, customer intelligence analysts, and, later in the process, IT folks to execute the results. This team needs to develop approaches to:

* Weak signal recognition * Discovery triage/leading questions * Offer mapping * Benefits statement/value proposition * Closing tools

Cellular carriers are masters of this. The minute a customer complains about a high bill, their CSRs roll out higher monthly commitment plans. The beauty of this is both the cell company and the customer wins. The cell company gets a higher monthly commitment that helps their planning and balance sheets. The customer gets a smaller bill.

While this is an obvious and not overly discrete example, it presents many of the objectives of operationalized CI. The CSR has a list of customer statements that trigger specific scripts (weak signals). The high bill-weak signal triggers a prescribed set of leading questions intended to triage the customer’s problem. Specific offers tailored to the customer’s unique situation are calculated and presented along with objective benefits statements. Finally, and most importantly, the CSR knows exactly how to capture the sale in their desktop.

Customers are very good at telling companies what they want. The problem is that companies haven’t taken the time to listen, and map offers to customer needs.

Taking this example one step further, a well-executed CI campaign will identify the customer, analyze their bill, identify the high-bill problem, and present the CSR with a tailored offer at the same time the call (or email or chat session) is delivered. In this scenario, the company has a tremendous opportunity to build customer loyalty. Maybe the customer has a question on how to use their phone. If the CSR presents an offer to lower their bill as part of the conversation, an unanticipated and greatly appreciated proactive service has been provided to the customer.

Scripting And The CRM Desktop: Technical

A key ingredient in the successful execution of a customer intelligence campaign in the contact center is a scripting engine, which will help CSRs react properly to subtle sales opportunities. The ability to walk a CSR through customer discovery and follow-on execution of a CI campaign requires such a scripting capability.

This can be difficult. Some CRM platforms are not easily or cheaply customizable to provide the robust scripting capabilities required to operationalize CI. Companies should do an honest evaluation of their CRM systems and compare the cost of adapting their desktops to accommodate scripting against buying an alternative tool.

An emerging segment in the agent desktop market are “wrappers” like Jacada, portals like BEA, and other approaches to automating customer sales and service functionality. These tools provide many additional benefits such as cross-platform reporting and simplified user interfaces including scripting.

Documenting Results: Operational

In addition to traditional contact center measures, managers should also report CI results up to senior management: number of customer database corrections, customer defection saves, survey work and, of course, sales. This reporting is enabled by the case management module of the CRM system and justifies the costs associated with operationalizing CI in the contact center.

So far this article has discussed the contact center in the context of supporting data integrity, CI execution in cross- and up-sell and proactive service, and as a source for CI. This effort requires training, tools and time interacting with the customer. In short, operationalizing CI in the contact center is not free. The contact center has to carry its own weight. If average handle time goes up and training activity increases, the contact center manager must be able to show incremental value to pay for it.

A baseline against which to calculate incremental AHT, training and other CI activity is needed. Center management should have a mechanism to track spikes in handle time, volume and other key metrics. Spikes due to legitimate customer service issues (e.g., wrong bills, defective products) should not be included in the incremental increases associated with CI. Interactions should be monitored, and time spent on CI activities accurately categorized. It should not be overly taxing for managers to track spikes and monitor calls to make sure CI activities are quantified. For CI activities that are not revenue related (e.g., data integrity validation), this quantification can provide senior management with the true cost of rectifying inaccurate customer data.

Management should include success metrics in every tool and process they build. Sales generated in the contact center need to be recognized as such. Customer defection saves should be tracked and the lifetime value of that customer included in the benefits generated by the operationalization of CI in the contact center.

This may all sound very ominous and risky, but in fact, most high-quality CI initiatives easily justify their operational costs and are very low risk. Most often they are very successful, but management cannot articulate that success objectively to senior executives. Therefore, reporting is a critical success factor.

Documenting Results: Technical

Optimally, contact center customer intelligence performance tracking should be an integrated part of the CRM tool. The sales module should be configured to characterize contact center CSRs as valid sales people, and appropriate screens made available to them to capture contact center sales. This is fairly straightforward in the service setting where any sale from the contact center is an incremental sale.

In an order-taking situation, CI up-sell and cross-sell scripts will map products and services that the customer originally requests to enhanced products and services that have higher margin or other corporate value. Before the sales initiative, a statistically solid understanding of sales patterns of the enhanced products or services can define the baseline, and incremental sales attributable to the contact center can be identified as a “high end” of the CI success rate.

The case management component of the CRM tool should be configured to capture other CI benefits such as customer defection saves and customer database corrections. The baseline exercise can be used to quantify intangible CI initiatives.

Finally, while in theory all of this reporting against AHT (volumes, sales, retention, etc.) should be automated, there is a lot of value that a good statistical analyst, positioned in the contact center, can bring to the table. Most contact centers have some statistical reporting function as part of their workforce management and quality teams. Expanding their roles and skills to include CI can leverage their knowledge of contact center data into CI reporting.

Contact Center As Input To CI: Operational

Almost everyone has experienced the frustration of dealing with a contact center and having the absolute sense that a complaint, comment or suggestion will go no further than the CSR’s ear. It seems that your words are bouncing off the CSR’s eardrum and the mumbled apology and denial of responsibility is the sound of your concerns falling on the floor.

It doesn’t have to be that way, and the solution is the easiest and cheapest idea in this paper. The solution will also provide the CI team with a deeply insightful source of customer concerns and behavior.

Provide CSRs with complaint management training and simple complaint capture tools. CSRs can defuse angry customers by using some variation of the following script:

“Although we have explored all approved options, the company does care when a customer has a comment or suggestion. (Company name) has a customer comment and suggestion tracking system and I will be glad to report your concern (suggestion). These are reviewed and if trends are identified, we will develop new policies or even technologies to address the situation.”

This is powerful. It demonstrates that although the company does have policies and procedures, it is willing to listen to new ideas that make working with them easier. In many cases, this can avoid a supervisor escalation or a complaint callback. And most importantly, CI analysts can comb the input for customer needs and behavior. The CSR has already identified the customer, so analysis by customer segment is possible. High-value customers can be called back if appropriate.

A strong complaint and suggestion management program has many benefits and is inexpensive. Complaint management should be part of every contact center’s basic core capability.

This idea of engaging in the CI process early in the analysis cycle is important. Customer analytics can become esoteric without pragmatic input from contact center managers. Contact center staff can not only provide valuable input into the analysis, but help guide it to outputs that are actionable. Thus, the role of the contact center manager is to keep the CI analytical team focused on operational realities: descriptor data points available in customer identification databases and to routing engines, mapping of descriptor data points to proper treatments and identification of success metrics.

Contact Center As Input To CI: Technical

Optimally, a complaint database should be implemented as part of the CRM package. The collection form is quite simple—a customer identifier field and a free form text field. The addition of drop-down boxes might be helpful, but it is important not to overcomplicate the process. The emerging technology of text mining can allow free-form complaints to be easily grouped together to show trends.

Absent an easily developed CRM solution, the “wrapper” alternatives discussed above can be configured to support a customer complaint database. Most email systems, including Outlook and Notes, can be easily adapted to capture such data. If necessary, a shared directory where CSRs can store text documents can be used.

Conclusion

Operationalizing customer intelligence is a key factor in migrating the customer contact center from a cost/service-oriented operation to a profit/loyalty-oriented operation. Contact center managers should be able to discuss customer defection saves, service-to-sales conversions, cross- and up-sell ratios and what the center is doing to increase customer satisfaction and loyalty with senior management.

To leverage CI, contact center managers need to do the following:

* Identify who is interacting with the corporation. * Get the interaction to the resource best able to handle it. * Execute specialized treatment on each individual interaction. * Capture the center’s results relating to increased sales and loyalty. * Leverage the contact center’s close connection with the customer to provide input into the CI analytical process.

Operationalizing CI in the contact center is a relatively low-risk endeavor, as long as success metrics are captured and communicated.

Ike Mitchell is a principal at Computer Sciences Corporation and a contact center specialist, focusing on both the technical and human sides of contact center management.