Reprinted from DM Review • April 2004 www.dmreview.com

Business intelligence (BI) isall about leveraging keyinformation to communicateideas and make sound, timelybusiness decisions. However,

many business intelligence solu-tions only deliver static reports andsummary graphical dashboards.These summaries often do not pro-vide the depth, flexibility and inter-activity required to make accuratebusiness decisions.

The situation is compounded bythe advent of large-scale data warehousesand the explosion of e-commerce.Businesses are now able to amassunprecedented amounts of dataabout their customers and operations.While this new environment is data-and information-rich, it is knowledge-and results-poor. Attempting tomake sense of huge data sets can be alabor-intensive, difficult and oftenfruitless exercise.

Today, management experts andsoftware vendors regularly promotereporting dashboards, scorecards andreports as the ideal way to monitorbusiness performance. They providevisibility into key performance indica-tors (KPIs) through simple visualgraphics such as gauges, charts andtables within a Web browser.Dashboards are appealing becausethey:• Present a wide number of different

metrics in a single consolidated view.• Roll up details into high-level

summaries.• Provide intuitive indicators, such

as gauges and stoplights, that are

instantly understandable (e.g., redbar means problem; green barmeans everything is on plan).

In many respects, a reportingdashboard can be likened to a dash-board in an automobile which pro-vides an “at-a-glance view” of the cur-rent operational state of the vehicle.Note that each gauge provides anindependent indicator of performance:the speedometer tells how fast thevehicle is moving, the tachometerindicates RPMs and the voltmeter isan indicator of charge.

Although the vehicle’s dash-board alerts the driver to conditionssurrounding the operation of thevehicle, it seldom tells the underly-ing cause of an abnormal reading. Alow-voltage reading, for example,could indicate a bad battery, a badalternator or some other problem.Some level of exploration is requiredto determine the actual cause of theabnormal reading in order to make areasonable decision on the correctivecourse of action.

In a similar vein, business usersneed to understand the reasons behindthe KPI results displayed in reportingdashboards for decision making.

Most reporting dashboards serveas a launch pad to access the underly-ing detail. The detail behind a KPI isusually accessed through a hot link toa report. The initial report often con-tains links to more reports for addi-tional detail that, in turn, may link toeven more reports. Cross-tabs/pivottables/online analytical processing(OLAP), spreadsheets and charts may

be leveraged within the reports or asseparately linked products/interfacesin the sequence.

The tension comes in trying tonavigate from summary to detail and

Are You Drowning in BI Reports?Using Analytical Dashboards to Cut Through the Clutter By Kurt Rivard and Doug Cogswell

in trying to correlate across the dif-ferent dimensions and indicators.

Analytical Dashboards Analytical dashboards represent a

new approach to dashboards. Ratherthan hide the details from the dash-board, an analytical dashboard is capa-ble of representing a vast amount ofinformation all at once. By using busi-ness visualization techniques, sum-maries and details can be selected andcorrelated on one screen in real time.A single visual interface can be used toempower users to engage in a discoveryprocess with their information, elimi-nating the need for the user to inte-grate reports, spreadsheets, charts andother tools.

Analytical dashboards provide theunderstanding of why KPIs are as theyare, what trends are driving them and

what really matters. They empowerbusiness users to navigate and inter-pret underlying KPI detail, replacingcomplex pivot table interfaces and theneed to string together “reportingthreads” that lack flexibility and pres-ent navigational challenges for busi-ness users.

Key Features of Analytical Dashboards

Analytical dashboards go signifi-cantly beyond current dashboards,charting or OLAP. Key features of ana-lytical dashboards include:• Big picture and all the details:

An analytical dashboard is not justan output report – it’s an interac-tive console. It shows the big pic-ture while also presenting thedetails. There is no need to transi-tion to another report, slice and

dice a cube or perform a databasequery to access the detail.

• Navigation with one interface:Drill down in traditional reportsand graphical front ends tends tobe an art that few can master –typically involving layers ofKPIs, down through OLAPtables, through to reports to thecritical pieces of data. Then, ifyou don’t hit the right piece ofdata, it’s back up, over and downagain. An analytical dashboardenables summary-to-detail navi-gation – all in one interactiveinterface with the flexibility tonavigate “sideways” and takeslices at any point in the analysiswithout having to backtrack.

• Correlation and discovery: Thecore to analysis is understandingcausal relationships across theintersection of multiple variables.Analytical dashboards are uniquein their ability to accomplish thisfrom an intuitive graphical inter-face. They allow the end user toselect subsets of multiple dimen-sions directly from the visualiza-tions in the dashboard and thensee the resultant intersectionimmediately in the same viewingscreen. The user is able to compareselected subsets with each otherand against the whole – a criticalstep in determining which path(s)to take to obtain their answer. Thisunique capability enables a dash-board viewer to perform powerfulwhat-if exploratory analysis with-out ever leaving the analyticaldashboard and discover keyinsights critical to business deci-sion making.

• More information: Analyticaldashboards can display tens ofthousands to millions of data ele-ments simultaneously. Details arenot lost in the summarization –they are right there in the display.

• Results export: Getting theanswer is step one. Exporting itinto standard reports or operationalsystems is step two. Analyticaldashboards are designed to facili-tate action on insight through theexport of results to other systems.

In essence, a properly structured

Reprinted from DM Review • April 2004 www.dmreview.com

analytical dashboard enables “selfservice” analysis by many peopleacross an organization. Such analysis isperformed intuitively by an end userwhenever he or she needs it – withoutdelay and without intervention by ITor other groups.

The interactive nature of the ana-lytical dashboard facilitates exploringand navigating through business datain an ad hoc manner – essentially adhoc queries against a preset structure.

The ability to perform guidedanalysis empowers the user to gainfact-based understanding of issues pre-viously hidden in the BI informationassets. This drives more accurate andtimely decision making and relievesmuch of the burden associated withanalyzing static reports.

Typically, three types of usersengage with analytical dashboards.First and foremost are dashboardauthors. These are the core analystswho understand the business objec-tives, the critical measures and per-formance indicators, and have theknowledge of how to apply reportsand dashboards to best representresults to a variety of stakeholders.They have an intuitive understandingof how to apply visualization, graphicsand reports to key data sets to createthe best representations for the organ-ization. They work with componen-tized tools that allow them to quicklyprototype, test and deploy theirreports and dashboards.

Business analysts and power userscome next. Each organization has alimited number of highly skilled andfocused analysts who are adept atsolving the “tough” questions andproblems. In some cases, these peo-ple will also be the dashboardauthors. In other cases, they will takethe structured dashboards andenhance, rework or recreate them byapplying additional components andanalytical techniques to answer theirparticular questions.

Dashboard viewers represent themajority of the users in most organiza-tions. They are business executives,line managers, store owners, front-lineworkers and a variety of others whoneed to see critical information in aneasy-to-digest format, but also needthe ability to interrogate it and getanswers to their questions as situationsand conditions change. Analyticaldashboards provide a unique andhighly valuable capability to provideand communicate the right fact-based

answers for this group.While each of these types of users

has very different needs and differentlevels of day-to-day interactions withthe analytical dashboards, they allinteract. Therefore, it is critical thatthey have a common tool with a con-sistent look and feel, and the ability toshare and communicate results.

An ExampleA financial planner might start

with an analytical dashboard similarto that shown in Figure 1. In thisexample, a bar chart shows the num-ber of funds in the portfolio by cate-gory, a scatter plot shows risk versusreturn for the individual funds, and adata sheet lists individual funds andtheir details.

There is definite value even if thiswere just a static report – the simpledisplay of several dimensions on onescreen with consistent coloring revealsthe following insights: higher returnsinvolve higher risk (there are no fundsin the bottom right-hand corner of thescatter plot – there are no funds thatoffer high rates of return for no risk),and international funds produced poorreturns relative to risk (the color goldrepresents international funds – theseare visible in the upper-left quadrantof the scatter plot; high risk and lowreturns).

However, the financial planner’ssituation is greatly enhanced if he orshe is given the ability to interact withand change the dashboard in real timeto answer client questions. Thisenables the planner to answer ques-tions during a live conversation (asopposed to contacting the client withanswers once he/she has retrieved andreviewed new reports).

If this were an analytical dash-

28 April 2004 | DM Review www.dmreview.com

Figure 2: Target Risk/Return Profile

Figure 1: Analytical Dashboard

board, the financial planner wouldsimply click and drag a rectanglearound points in the scatter plot thatrepresent funds with a targetrisk/return profile. This would imme-diately highlight the mix of fund cat-egories in the bar chart, and wouldshow that balanced funds offer a goodrate of return for moderate risk (seeFigure 2).

Then, by clicking to remove the“grayed” data, the financial plannerwould drill down to a chosen subset ofbest performers in the target profile.In this case, Fidelity, Vanguard,Merrill and Dodge offer large balancedfunds with reasonable rates of returnfor moderate risk (Figure 3). Using ananalytical dashboard, the financialplanner has shown his or her client thefacts behind the fund recommendationin a clear and compelling manner.

In this example, the financialplanner has easily navigated through anumber of multidimensional charac-teristics (risk, return, type, assets).The financial planner has shown thebig picture (with the bar chart show-ing a categorized roll-up of the funds),while at the same time presenting thedetails (the individual funds in thescatterplot with attributes in thetable). Big patterns and small anom-alies can be seen simultaneously.

The same analysis could have beenperformed using other business intel-ligence techniques, but executionwould be much more difficult. Itwould also be much more difficult forthe typical client to understand thefinancial planner’s analysis.

For example, OLAP and/or pivottables could be used. However, the datawould first need to be organized intodimensions and measures for OLAP.The user would need to read acrossvarious attributes and attempt to “see”the patterns, but they are not obvious(see Figure 4).

With pivot tables and OLAP, thefinancial planner can also drill downto see further details, but it is quitedifficult to see where the interestingdetails are – such as outliers in themortgage funds, or that the balancedfunds are tightly grouped together, orthat the largest funds (by net assets)tend to be growth funds. While it is

possible that using the correct descrip-tive statistics might reveal some ofthese patterns, the analytical dash-board makes these patterns immedi-ately visible and easily accessed with-out any statistical knowledge.

Traditional dashboards or reportscould also be used by the financialplanner. However, such tools wouldrequire the client to view various dis-connected charts that are not interac-tive and do not facilitate selection anddrill down. The client would not beable to isolate details, nor would theclient be able to see where thoseoccurred at the big-picture level. Thefinancial planner’s strategic discussionwith the client would be far lessinformative and far more difficult.

Benefits of Analytical DashboardsAnalytical dashboards make BI

solutions better. They complementother analysis and reporting tools byhelping people across an organization

cut through the clutter to find theinformation they really need in theglut of reports they deal with everyday. The benefits are better decisionsmade faster, with results communi-cated more easily and clearly.

Kurt Rivard is a senior sales consultant with ADVIZORSolutions where he works with business intelligenceand analytics partners to integrate business visualiza-tion into their product and service offerings. He alsoworks directly with Fortune 500 companies in thespecification, design and implementation of analyti-cal dashboards to address their business needs. Rivardhas more than seven years of experience applyingbusiness visualization to business intelligence, analyt-ical applications and analytical tools, and in thatcapacity has been instrumental in the pioneering ofanalytical dashboards. He also has more than 16 yearsof experience in software design and development. Hecan be reached at kurt.rivard@advizorsolutions.com.

Doug Cogswell is president and CEO of ADVIZORSolutions. Cogswell has led ADVIZOR Solutions sinceits spin-off from Lucent Bell Labs in 1999, and hasspent the last 10 years growing successful enterprisesoftware companies. He is well connected throughoutthe software industry and actively participates on anumber of boards, including the Chicago SoftwareAssociation (of which he is Chairman), theChicagoland Chamber of Commerce and OpportunityInternational. Cogswell can be reached atdoug.cogswell@advizorsolutions.com.

Reprinted from DM Review • April 2004 www.dmreview.com

Figure 3: Drill Down

Figure 4: Pivot Table