The BI Formula

By Brendan B. Read, Senior Contributing Editor  |  April 01, 2011

This article originally appeared in the April 2011 issue of Customer Interaction Solutions

Accurate information, created from data, is the building of block of effective decision-making, such as whether to increase or reduce staffing levels, boost training, increase first contact resolution (FCR), adjust offers or launch new mobile, social media or telemarketing campaigns. And as any manager can tell you, you can never have enough high quality information in your hands fast enough, especially in today’s economic climate where there is little room for errors or missed opportunities to grow business and cut costs.

Contact centers provide data in abundance, and more so as customers increasingly interact with organizations over a multiplying array of channels. Every call, chat, e-mail, site visit, text and Tweet and their contents contain that raw material. Yet there is more than centers can process quickly enough into actionable information. And too often it is dumped into and stored in silos.

Business intelligence (BI) methods and software provides an effective formula to process data into useful information and present it in the hands of managers. It draws and mixes the contents from the silos, providing a complete view of what is happening in businesses including operations.

And as contact centers become more central in organizations’ marketing and sales strategies by retaining existing customers and attracting new ones, it is more imperative than ever for them to learn about and employ BI.

BI Interest and Drivers

There is increasing interest and demand growth for BI applications that touch on contact centers for those reasons. Mark Flaherty, chief marketing officer of InetSoft says that he is hearing this from his firm’s sales team. 

“The driving force for BI is the need to squeeze more knowledge out of the large amounts of data that are pouring through contact centers,” he reports.

Richard McElroy, president and COO of Symmetrics is seeing his clientele becoming interested in the big picture customer interaction and operations assessments that BI provides. They are also seeking to extend BI analysis to pre- and post-call activity and data such as case management from help desk or CRM systems and to transaction (sale/no sale) results from CRM. They are interested in leveraging BI on the back end to aggregate, integrate and optimize data by using it to automate data collection and optimization.

McElroy’s customers also want to understand the details of both “within call” experience i.e. ACD/IVR information and the complete interaction experience. And one of the most critical applications and needs to those ends that are pushing BI use is bridging the contact center systems: ACDs, IVR, quality monitoring and workforce management to haul out the data they contain for processing and analysis.

Each of these systems has their own internal reporting systems that can report on only their particular “islands” or silos of information, reports the Symmetrics COO. His clientele commonly uses spreadmarts, Access databases or running of the canned reports and cutting/pasting info into spreadsheets and documents to get what they want, or a combination. Yet they are too slow and cumbersome to process and assemble the information.

“Many of our customers tell us that they can get a basic report with metrics from the ACD that give results of what happened,” reports McElroy. “What they can’t do is gain insight as to why the result is what it is such as, ‘How is my service level percentage being affected by schedule adherence? What is average time of agent call sessions versus IVR self-serve sessions?’”

“The problem is that supporting data is not always in the same system or reports, and each may be saying different things,” he adds. “With BI, contact centers can get their data, metrics, KPIs and reports in a single common environment to obtain a holistic and singular version of what is happening.”

Another key driver is greater business agility to respond to rapidly changing needs. Many Pegasystems (News - Alert) customers in this space adjust their customer interaction strategies multiple times each day, responding to competition, the flow of the day or simply to experiment and try out new things.

This kind of ‘hyper agility’ enabled by BI has another benefit–making it less critical to get the marketing and sales strategies, such as pricing, incentives and priorities, right from the start “because the strategies can be course-corrected on the fly,” points out Rob Walker, Pegasystems’ vice president of decision management and analytics.

BI, predictive analytics integration?

BI can be regarded as looking back to where one has been to help see where one is going. Predictive analytics can be seen as looking ahead to plan the next moves and anticipate what is going to happen next. Both BI and predictive analytics working in tandem can therefore help organizations reach their destinations.

Symmetrics’ McElroy sees a role for predictive analytics in BI, one that will grow as data mining and predictive modeling technologies around it evolve, become easier to use and become more affordable. Yet contact centers need to get traditional BI right before they can effectively take full advantage of predictive analytics’ capabilities.

“We believe that predictive analytics will become a more prevalent BI technology used,” says McElroy. “However, it will not overshadow or diminish the focus on or importance of traditional BI in contact centers as it allows companies to ensure contact centers are delivering value in context of the overall business objectives. Yet as effective BI adoption increases in the contact center market, we’ll then see “advanced BI solutions” like predictive analytics increase in adoption.”

Erick Brethenoux, executive program director at IBM SPSS (News - Alert) argues that predictive analytics should be installed and used alongside and at same time as BI, as it enables contact centers to become more efficient. For example, when a customer calls a firm on a service issue, predictive analytics can generate, as the agent enters the information, customer-tailored sales offers to the agent that they then can present on the spot. This avoids lengthier conversations, and more costly telemarketing and direct mail or e-mail campaigns.

“Business is going faster and faster; if you can see only one meter ahead of your windshield when the wall comes it is too late,” says Brethenoux. “You have to augment that vision to see what’s coming at you and that’s what predictive analytics does.”

The BI and predictive analytics integration is happening. The IBM (News - Alert) SPSS unit, which provides predictive analytics solutions, now uses IBM-owned Cognos BI tools for dashboarding and reporting. In turn, Cognos 10, which came out in October 2010, uses some of the SPSS statistics tools.

Cognos 10 has other new key features, among them an integration with Lotus Connections that enables social collaboration analytics with intuitive navigation capabilities and simplified user experience. It also provides users with an integrated view of historical information with real-time updates to give users a complete picture of their businesses. Employees can now interact with each other in real time in communities, wikis and blogs.

“Once someone is on the line you can use predictive analytics to cross-sell, upsell and to find out whether or not they are going to churn, and then you can use BI to report on what happened, all of the offers made and which ones were accepted,” says Brethenoux. “It goes full circle to use that data to feed into the predictive analytics.”

There are also solutions like SAP (News - Alert) BusinessObjects BI and EIM 4.0 that work with existing data for the efficient discovery of interesting and predictive findings. SAP BusinessObjects Event Insight helps customers discover and understand the business impact of events across business networks. These include events affecting customer service and contact centers as varied as product recalls and late deliveries. It processes large volumes of event data in real time, enabling firms to discover complex patterns in seemingly unrelated events, analyze the information within business contexts, and identify emerging opportunities or threats via alerts, dashboards, or reports.  

“What-if analysis and predictive analytics play a critical role to uncover trends and patterns to solve business problems, anticipate business changes, and gain insight into what the future will hold,” says Nic Smith, group product marketing manager of BI solution marketing at SAP BusinessObjects.

InetSoft's Flaherty points out, though, that completing such connections between BI and predictive analytics are not automatic. Predictive analytics applications still requires statistical modeling and segmentation analysis using specialized software and highly skilled analysts. But once these models are built, a flexible BI application can use them and apply up to the minute data to generate the desired alarms.“You have to be careful in setting expectations for predictive analytics,” advises Flaherty. “I think some vendors are marketing it as artificial intelligence that will solve problems for you. But predictive analytics still means using past performance to project future performance, not predict it. A BI application can help point out deviations from normal expectations and make decision-making by people more timely.”

BI affordability

As essential as BI may be in uncovering opportunities to generate more sales, improve service and with it customer retention and reduce costs, it is not inexpensive. The price tags run from $25,000 to $500,000 or more. Therefore, real need must be there or the investment will not be approved.

Symmetrics McElroy recommends addressing BI affordability and budget constraints by following this mantra: “think big, start small, go fast.” Here’s how:

*          Establish a BI strategy that tries to encompass needs for a three- to-five year outlook that will lay the technology foundation to get you there

*          Start in phases by addressing the most critical elements first, and then lay in additional elements within the foundation based on priorities and budget availability

*          Once a phase one is established or proven and accepted, move quickly to the next phases to take advantage of internal momentum and executive buy in

“This approach maximizes the possibility of long term success by avoiding project “implosion” often associated with trying to achieve too much at once, and enables shorter term issues to be addressed in priority that reflects the current business environment and challenges,” says McElroy. “It also provides a budget-friendly approach that allows investment to take place over time. And it maximizes user acceptance by controlling the amount and speed of change to manageable levels.”

Up-front costs are an issue though and Symmetrics is taking different routes to managing it. The firm uses a modular architecture that permits the phased approach via several layers. These are data integration/consolidation/optimization, content presentation/visualization and access/dissemination that are priced accordingly. It also offers a full-featured software-as-a-service (SaaS (News - Alert)) hosted application as a lower-initial-and-support-cost option to customer-premises software.

“Our approach allows us to deliver a ‘right-sized’ offering that can either deliver a turnkey, end-to-end solution, or only specific modules that can be integrated with, or leverage, current BI infrastructures in place” explains McElroy. “It also enables those organizations that are comfortable with hosting to reap the benefits of BI while taking up a smaller infrastructure and IT staff budgets without the complexity of installations.”

Symmetrics is now working on a software appliance that combines its BI application and a streamlined version of system software which can be readily installed on industry standard hardware and deployed as a turnkey. This method will save IT costs by avoiding server and integration expenses and eliminates the need to manage multiple maintenance streams, licenses, and service contracts.

“Our clients have told us they would prefer to avoid having to deploy and manage multiple IT layers like operating systems, database environments and web servers just to run one application if they don’t have to,” says McElroy.

Making BI more usable, and useful

With the tasks BI can accomplish and the value it can create, it is not surprising that the software is often complex and can be challenging to learn and use. Yet to achieve today’s demanding C-suite goals of “strong results yesterday, not today or tomorrow”, this expensive software must be highly usable in the design and in the setup.

Suppliers are working on just that. For example, InetSoft has enabled users to individually customize dashboards using its Style Intelligence BI solution without any IT assistance. Both agents and managers can filter, sort and drill down into Style Intelligence to give them access to a huge amount of information that previously was only accessible to analysts.

Symmetrics has come out with a SaaS offering that delivers a new natural language interface and extended metadata layers within data models for its solution. It can answer simple queries and enable report generation directly by managers, supervisors and team leads in minutes, instead of taking days to have analysts do the same thing.

The interface allows users to ask questions in three parts: “what am I measuring, how do I want to organize and group the information and what filters do I want to apply?” This tool permits analysts to be used for more productive endeavors like diving into the data at a deeper level to better understand the “whys” – why are the results what they are and how can they be improved if they are not satisfactory.

Equally, if not more importantly, firms should obtain manager/user buy-in for BI and provide adequate training to avoid it from becoming costly shelfware. One step to take is to select metrics and KPIs that are aligned with business goals, recommends McElroy. In that fashion the results measured by contact centers are more readily translatable into actions that support overall business goals.

For example, if a software company’s primary goal is “to lead the industry in technical support quality/satisfaction”, then having a contact center KPI that promotes agents to have the shortest possible average call duration as a top priority is not in alignment with this objective. An FCR KPI would likely have been a better choice.

“Often our clients would say that reporting and analysis does not work or it is too difficult or hard to use, but once we began educating them, the lights went on and they would say ‘this is now a major competitive advantage for me,’” says McElroy.

Social Media and BI

The rapid rise and increasing customer decision-making influence of social media has made integrating and analyzing this data, converting it immediately into actionable information via BI imperative. The challenge is that myriad comments, like e-mails and SMS/text messages are unstructured. They need to be integrated with data from traditional sources, like transactions.

Nic Smith, group product marketing manager of BI solution marketing at SAP BusinessObjects, reports that there is shift towards sentiment analysis, which permits text data integration to analyze unstructured data sources like documents, e-mail, blogs and Twitter. It then summarizes the information into key sentiment that is surfaced throughout the content.

Data governance and quality visibility is important to empower business and IT users with a single environment to discover, assess, define, monitor and improve the quality of their enterprise data assets, he points out. And bringing information together from multiple sources is critical, whether it’s one source or hundreds, making report authoring tremendously easier.

To permit such analysis and integration, SAP recently released SAP BusinessObjects 4.0. It provides deeper text analysis integration that complements traditional business insight. With it organizations can monitor, analyze, explore, report and act with confidence in the accuracy of not only their transactional data but also with full insight into the trends and sentiments expressed in the unstructured content of blogs, e-mail and social media streams.

SAP BusinessObjects 4.0 also features a new multi-source multi-dimensional semantic layer and a new common authoring experience to simplify ad-hoc analyses and content creation. This allows users to obtain a more holistic meaning from the information in less time.

“[There have been] mind-blowing increases in the amount of data captured, the growth of consumer use of social media to discuss, laud and discredit brands and the move toward mobility on a global level,” says Smith. “In order for companies to retain and grow their customer base, they need to be able to capture, monitor and analyze enormous amounts of data from a growing number of sources. Organizations need insight with confidence across business and social data.”

Going Mobile

Contact center supervisors are becoming less tethered to their desks. Thanks to smartphones and tablets, they can leave their desks and stay in touch with what is going on.

InetSoft is working on mobile apps of its BI solutions that are optimized for smaller screen spaces and different input methods scheduled to be released by mid-2011.

“The mobile channel influences the usage demands of contact center managers,” says Mark Flaherty, chief marketing officer at InetSoft. “Now they expect to be able to stay in touch with the BI application that they are using to manage their operations when they walk away from their desk.”

The following companies participated in the preparation of this article:






Brendan B. Read is TMCnet’s Senior Contributing Editor. To read more of Brendan’s articles, please visit his columnist page.

Edited by Stefania Viscusi