Six variations on the theme of analytics

InsightaaS: A conference focused on analytics featuring more than 40 formal user presentations and case studies and thousands of informal discussions is an ideal environment to gather real-world BI examples that illustrate both the core capabilities and the diverse applications of BI technology. These six examples from the Information Builders Summit 2014 provide excellent insights into how BI tools are used today — for reporting, to support management decision-making, and as a means of attracting and engaging customers.

Cambridge Memorial Hospital

Business context: Cambridge Memorial Hospital provides acute care services to residents of Cambridge, Ontario, the township of North Dumfries and the Region of Waterloo. It employs 1,100 healthcare professionals and staff, 280 physicians and 400 volunteers.

BI activity within Cambridge Memorial

WebFOCUS is used to present business-critical information throughout the hospital. The facility has an ER tracker that flags variances from expectations — for example, alerting staff if a patient has been in the ER longer than the target time. It has a number of corporate-level systems “geared towards senior administration and the board” that connect executive analyses with action plans, so that the hosptial’s leadership can register awareness of and improvement plans to address operational areas that are off target. The hospital has other systems geared towards specific administrative functions, such as an HR attendance management system. And Mike Meyette, director, IMT and corporate services for Cambridge Memorial, has built a data warehouse in WebFOCUS that is used to populate management dashboards that connect funding with the various budget categories used to justify government payments to the facility. These systems generally replace manual or Excel-based systems that were inconsistent and labour-intensive to maintain. Using WebFOCUS, it is possible for Mayette and his small team (a data modeller and two analysts) to respond effectively to the hospital’s diverse requirements.

Prioritizing initiatives

Meyette’s team doesn’t have the cycles that would be required to develop formal cost/benefit analyses for each BI initiative. Instead, it is driven primarily by available capacity: requirements that consume substantial team resources are the first targets for automation. For example, the scorecarding required to support management business analysis “was really a workload issue,” with potential for reducing time spent finding, cutting and pasting information. Cost is also a driver; for example, the HR system replaced one full-time employee, easily cost-justifying its development effort.

Driving adoption

Cambridge Memorial HospitalThere are environments where employee ‘sponsors’ are cited as important to new system adoption, but Cambridge Memorial isn’t one — “we haven’t done that much,” Meyette said. “I think the product will champion itself” as new information access systems are rolled out to an ever-wider user community within the hospital. “People want the information; they know that they need it. They’re being benchmarked…and people really do want to understand.” And in Cambridge Memorial, it behooves staff to stay current with metrics — Meyette notes that “our CEO and VP are both very information-driven.”

At an institutional level, Meyette observed that Information Builders has been building a substantial presence within the Ontario healthcare community. He mentioned several hospitals — Grand River, Quinte, Guelph, Brantford — which have adopted Information Builders as a platform, and noted that the company is “picking up four or five or six hospitals every year.” Meyette believes that it is very helpful to have local organizations to talk with about issues and challenges, and that with Information Builders’ growth in the segment, there is potential for an Ontario hospital users group, which would help spread BI expertise further.

Lessons learned

Pressed for advice that he would give new users, Meyette stressed investment in human deliverables — training and consulting. Information Builders instructed Meyette to “keep some dollars set aside for training,” and this helped his staff to develop “the starting point” needed to usi the product. When it was time to develop more sophisticated capabilities, “we brought in consultants from Information Builders. They were fabulous — you could sit alongside them…and get a huge amount of knowledge transfer. It would have taken us a couple of months to do something they could help us get through in a week.”


Scotiabank is one of Canada’s largest financial institutions. It is a global organization, operating in 55 countries. Scotiabank has 83,000 employees and serves 21 million customers.

BI business objectives

Scotiabank’s need to accommodate multiple and changing reporting requirements with overlapping inputs and outputs, drawn from a wide range of data that is also continuously evolving, argued for a loosely-coupled approach to BI development. In this way, the bank could extend key core capabilities (such as security, monitoring, pre-set alerts, etc.) across multiple reporting initiatives, while remaining flexible in its approach to assembling data to meet specific business and regulatory objectives, and allowing the BI team to stay responsive to changes in requirements and to obtain leverage from the development of core modules.

Key deployment factors

To create its BI infrastructure, Scotiabank “started from the ground up.” It first built the hardware infrastructure, observing bank-standard architectural policies for redundancy in business-critical systems. It then established the software stack: a dedicated database platform capable of providing a basis for integrating requirements from multiple analytics initiatives, and most of the components of IB’s iWay suite, including iWay service manager (plus Enable, a service manager tracking add-on), the iWay data migrator, and iWay Data Quality Center, plus IB adaptors to connect multiple data sources. The third step was the software development lifecycle, which was described by Scotiabank as “one of the biggest challenges” in the process, as the bank has rigid rules around the products and processes used to automate code movement from the development environment to the integration testing environment to the quality assurance testing environment to the functional testing environment and finally to the production environment. The fourth level was data integration; to address this, Scotialbank employed “two main industry standard best practices” for moving files and for sharing data.

Once this platform was in place, Scotiabank defined what it refers to as a “logical unit of work.” These units have several different “flavours” — they can involve a file transfer, ETL, or record-level processing, or some combination of these initiatives. Applications can obtain capabilities by reusing logical units, thereby improving efficiency.

Outputs and considerations

Scotiabank’s BI group is now supporting five major reporting outputs, corresponding to specific regulatory requirements. Because the output is used for reporting to regulators, Scotiabank has a “heavy focus on data quality,” and usies “auditable metrics” to ensure that the data on which the reporting is based can be trusted.

Because the five major initiatives have evolved in response to different requirements and in different timeframes, they are discrete, even if they benefit from Scotiabank’s integration focus. Scotiabank senior manager Tatjana Papic outlined a future vision in which Scotiabank will establish one master output from which project-specific outputs can be assembled flexibly to meet these (and future) demands.

Lessons learned/advice to BI users

Through the course of her presentation at the summit, Papic offered a great deal of insight helpful to firms trying to plot their own analytics strategies. In her presentation, she stressed several key themes:

  • Try to establish integration as a discipline. Scotiabank had two main integration objectives: to share infrastructure and licensing, and to reuse data, code components and best practices. Integration poses substantial challenges on both a technical and a business level. On a technical level, continuous change makes establishing integration frameworks problematic, while on a business level, “there is very little incentive to share.” It is easier to implement a stand-alone, project-based initiative than to build a framework spanning multiple objectives — but this will lead to increased maintenance overhead down the road, and may also result in inconsistent reporting, which would be very problematic for a regulated entity like Scotiabank.
  • Deploy incrementally. Scotiabank introduced new Information Builders platforms “in parallel mode,” moving some users to the new systems while other users were routed to the existing systems. This gave Scotiabank an opportunity to see how the system would function when it was exposed to actual users, without needing to support the entire user base from launch.
  • Start simple. It can be a “hard sell” to persuade users to accept a subset of target functionality in initial systems, but there is real merit in building from a solid base; as Papic said, “it actually has to work right…keep [initial implementations] simple and open, and then add” functionality incrementally.
  • Work on a “centre of excellence” basis. Scotiabank implemented a “hub architecture” even though adoption did not follow a hub adoption model. The Scotiabank team acts as a central resource and capitalizes on consistent data and reusable components, but works individually with users.
  • Drive adoption by linking closely with business executives and user “champions.” “If there’s anything I have learned from the [BI project] experience, it’s that none of this would have happened without senior management support.” She added that “you absolutely need the champions within the organization who are going to go around” and reinforce the need for and benefits resulting from the new platform.

Titan Insurance

Titan Insurance is a Cleveland-based company specializing in automobile insurance. It uses WebFOCUS primarily as a means of consolidating information from many sources into usable reports. Titan’s Tom Farynowski put it this way: “WebFOCUS provides connection. We have connections to DB2, several Oracle servers, Excel, and text files. Titan has data in Oracle (data warehouse) and Oracle (separate schema) and DB2 tables. I extract this information and use FOCUS logic (i.e. JOINS) to combine the information into one document (usually Excel but sometimes PDF). I have one application that has over 50 metrics where the information is gathered from 15 input files. An Excel File is generated dynamically based on user criteria (such as month/year or agency identifiers).

BI business objectives

Farynowski: “WebFOCUS can deliver reports in a real time method so that people don’t have to cut/paste information from one source to another. The business doesn’t really know it is sourced via WebFOCUS, but the reports it produces are invaluable…WebFOCUS allows users to read the data sources directly without the overhead of a data warehouse. Also, general users can utilize tools like Info Assist to generate reports in an efficient matter with only a little front end work by IT to setup the connections.”

University of North Carolina-Charlotte

UNCC is a major university based (as the name implies) in Charlotte, North Carolina, with more than 26,000 students and nearly 3000 staff members housed in 75 buildings located on the University’s 1000 acre campus. By way of background, UNCC’s Tony Vecchio stated that “UNC-Charlotte uses Ellucian’s Banner software to administer its student, financial aid, finance, human resources and advancement (fundraising) work activities. Banner uses Oracle as its data repository.  Banner delivers report-friendly views of its data in a parallel Oracle repository called the ODS (Operational Data Store). The vast majority of [UNCC’s] WebFOCUS reporting is based on the ODS.  About eight years ago, when the university converted to Banner to run its major business functions, there was a parallel effort to provide a web-based central reporting platform that initially consisted of the hundreds of existing reports manually created across campus but which are now automated through the use of WebFOCUS. Each of the business areas was provided a reporting platform (See Figure 1) that contained relevant reports which were static in format but allowed the ability to filter content (See Figure 2).”

UNCC graphic

BI activity at UNCC

Vecchio states that UNCC has “20 developers outside of central ITS (Information and Technology Services) that can write their own ad hoc reports against the ODS as needed. These people are responsible for providing WebFOCUS reporting solutions for hundreds of people in their departments… for automating hundreds of [previously] manual reports…so that hundreds [of business users] can spend time on other tasks.”

WebFOCUS deployment and business benefits

Vecchio added: “Simply stated, we use WebFOCUS for all of our university reporting needs. The difference or benefit of using WebFOCUS is automation and consistency of data in that there is one answer to a question…A data repository exists and everyone is required to use it versus a decentralized system in which departments across campus store their version of data — financial data for example. Secondly, and most important is that manual effort is replaced by automation saving many man hours and providing a much more consistent result. For example we converted the university’s annual financial statements to WebFOCUS such that with the click of a button statements are generated in minutes, replacing manual preparation requiring hundreds of man hours.”

Olive Media

Olive Media is the digital media division of Torstar. It sells advertising for Torstar’s online products, and also represents third party advertising space, from publisher sites like the New York Times and

BI business objectivesGary Neff, Olive Media

Gary Neff, BI specialist at Olive, states that “We had a lot of systems that were separate…four different ad servers, and also our CRM systems…we were building a lot of things in Excel, running reports from the different systems and then doing lookups. That is obviously not an effective tool…” The decision was made to aggregate all of the data into a warehouse, to improve business responsiveness by aggregating data so that the business can track how campaigns are doing, where Olive is relative to targets and budgets, and importantly, to “see things that aren’t happening” — for example, campaigns that are getting no impressions, and hence require some remedial action.

The BI system is used across the organization by management and operations, and also as a means of replacing previously-manual processes used by sales to report to customers. “The account management team was pulling reports from different systems, and then building smaller reports for each agency to say ‘here’s how your campaign is tracking’” — a “tedious” process that lacked “an overall consistency.”

Building on the platform

Once WebFOCUS was in place, Olive “started developing reports.” The first target was to automate “key company reports that people were developing in Excel.” Olive used Information Builders consultants to help build the initial reports, and used the knowledge gained from that experience and from Information Builders’ Focal Point online development community to expand its scope into other reporting requirements.

Business outcomes

Olive has used the WebFOCUS platform to build out several different business-critical reports, including:

  • An “executive flash report” that summarizes key business issues, with graphical views that illustrate other important factors like pipeline — “a lot of year-over-year tracking and year-to-budget tracking.”
  • Sales team reports that track agency performance, highlighting variances from the previous year, and capturing performance vs. target data.
  • Sales management reports that aggregate rep data, showing individual and territory performance. Neff described this as a “big win” for the BI team, because “the buy-in was instantaneous — there was nothing before like it.”
  • Operational team reports covering core issues like campaign tracking  that answer internal questions like “are we going to deliver on time?” or use of historical information to prepare delivery forecasts that supply information used to communicate with agency customers.

Yellow Pages Group

Serving more than 275,000 of Canada’s 1.2 million businesses, Yellow Pages Group (YPG) is an institution within the Canadian business environment. Although it is known primarily as a publisher of an annual telephone directory, it is evolving rapidly to meet customer demand for digital advertising — and in fact, in the last quarter of 2013, digital represented 45% of total company revenue.

This shift to digital brings new challenges to YPG. The print directory is a static, 1x/year resource, while online advertising needs to be dynamic. Mobility has an enormous impact on digital advertising: according to IT director Richard Langlois, mobile consumption is growing “nine times faster than web.” And YPG’s advertisers are mostly small businesses, who are not inclined (or have no time) to analyze digital data — making clear and actionable feedback from YPG essential to its digital success.

BI business objectives and activities

YPG’s BI goal is tightly intertwined with a key digital business target — to be an authoritative source for consumer and advertiser analytics (CAA). Its BI objectives include: improve the quality and consistency of the CAA it produces, integrate external data with its internal data, deliver user-friendly tools enabling ad hoc queries, and wrap all of the system functionality that masks the complexity of the underling 56 billion row database. One key target for the system was an ROI tool combining statistics from internal data (ad clicks) with external intelligence on conversion rates and values. The system is designed to let users change variables (such as the value of a sale) and to use the resulting insight to understand campaign goals and achievement.

Yellow Pages group connects its Big Data roadmap with its Enterprise Architecture roadmap
Yellow Pages group connects its Big Data roadmap with its Enterprise Architecture roadmap

The YPG BI project got off to an unusual start — the company’s marketing team developed and tested visual templates with prospective users, and then delivered the finished design to IT, with a mandate to echo the screens exactly as rendered in the proofs, and to fill in the functionality implied by the display. This approach is well-aligned with driving customer acceptance, but introduced several daunting technical tasks: it required YPG’s technical group to understand how to manipulate the interface at a very detailed level to ensure that the output expectations were met exactly (and in two languages!), while at the same time requiring a functional system that could rapidly integrate and process internal and external data and deliver meaningful feedback to the user — who in this case, would often be a small business manager/owner seeking quick and clear feedback that would them to understand whether current advertising expenditures with YPG were yielding positive returns, and how they should adjust their advertising approaches to drive more business.

As would be expected, the YPG BI project was (and continues to be) extremely complex. One point that Langlois stressed, though, is the importance of developing four parallel roadmaps: for data, technology, and applications, tied to YPG’s business arthcitecture. All of these aspects, he asserted, are important — and ensuring that there is a clear strategy for evolving in each area, in ways that don’t negatively affect performance in the other areas is critical. YPG uses an iterative Agile development approach to ensure that current initiatives in any of the three IT areas do not diverge from the overall plan.

Note: The full scope of YPG activities requires a “deeper dive” than we can provide in this post. will deliver a formal YPG analytics case study in 3Q14.



Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.