I received an email from Hyperion today announcing a webcast entitled: “Scorecards vs Dashboards — When to Use Which?” The relationship between scorecards and dashboards suggested in this title is misleading. Scorecards and dashboards are entirely different animals, not two things that you choose between. The balanced scorecard, as defined by its inventors Kaplan and Norton in the early 1990s, is a methodology for measuring and managing a business, which examines the business from the following perspectives:

• Financial
• Customers
• Business processes

A dashboard is a particular means of displaying information for the purpose of monitoring what’s going on in the business. Dashboards are a popular means to display balanced scorecard information, but they can display other information as well.

Business intelligence and performance management are confusing enough as it is without software vendors—supposed experts in these fields—adding to the confusion. How can we achieve business intelligence or manage a business if we can’t even keep our terms straight? Clear communication is essential to success in both of these important and intimately related efforts.

Take care,

Dashboards are designed to help us monitor what’s going on at a glance—or at least, that’s what they’re supposed to do. Unfortunately, any time a vendor combines more than one chart on a single screen, they call it a dashboard, no matter what its purpose, which creates a great deal of confusion. To give people a means to rapidly monitor what’s going on, dashboards must be designed in particular ways to take advantage of the strengths of visual perception and cognition and to work around or augment their weaknesses.

I keep my eyes open for anything new on the Web that goes by the name “dashboard,” and I mostly find examples that fail to communicate in a way that truly supports monitoring. I recently saw a press release from 8e6 Technologies, which announced their new product called 8e6 Threat Analysis Reporter.

8e6 Technologies, a security company dedicated to Internet filtering and reporting, today announced a new enterprise reporting suite designed to alert corporate executives and IT professionals about Internet threats coming from within their organization. With this product suite, 8e6 is introducing the new Threat Analysis Reporter, a real-time monitoring and remediation solution that displays an organization’s current security threat level, stratified across various gauges in an easy-to-read dashboard interface.

What does it provide?

A redesigned, Web-based user interface that is intuitive for users and provides better visibility to network health through a new graphical dashboard.

In other words, it uses a dashboard to help network administrators monitor the Internet activity of a company’s employees in an attempt to spot threats quickly and respond to them. Unfortunately, what they provide is data-lite—it not only tastes bad but it’s less filling. The product’s data sheet includes the following screen print:

If you were a network administrator responsible for monitoring real-time Internet activity, assuming that the red, yellow, and red zones on these gauges were defined meaningfully, you could examine this dashboard and detect potential problems (that is, the needle is in the red), but you would have to work much harder at it than the at-a-glance monitoring that dashboards ought to support. Study this dashboard for a few minutes to see if you can detect any of the flaws in its design.

…………………………………………………………….

Now that you’ve taken some time to critique this on your own, here are a few of the flaws that I noticed:

• Measures that are in the problem zone (that is, in the red) don’t pop out as clearly as they ought. With so much green, yellow, and red color on this screen, noticing that a needle is pointing to the red zone takes time. Not a lot of time, but much longer than it ought, especially for real-time monitoring. This dashboard attempts to assist in the search for trouble by changing the color of the gauge’s title to red when the needle ventures into the red zone, which might be an adequate attention-getter if red were not used so much elsewhere. If the color red only appeared on the dashboard next to items that needed attention, people would able to spot the problems much faster.
• All of the gauges work exactly the same, with the green zone on the left, yellow in the middle, and red on the right, but this doesn’t seem to match the nature of the data. If a high number of people in your company are accessing adult Internet content (the top left gauge), that’s definitely a bad thing (unless your company serves as a pornography watchdog). Low numbers in the green zone and high numbers in the red zone make perfect sense in this case. What I don’t understand is why a high level of productivity (the top middle gauge) is a bad thing (or security, network, engineering, and marketing).
• These circular gauges use of a great deal of space to say very little. They tell us a number, such as 351 for adult content, and if this is good, satisfactory, or bad, and that’s it. Apparently 351 is as bad as it gets for adult content, because the needle is pegged at the extreme end of the red zone. For a full screen, 10 numbers and ten qualitative judgments (good, satisfactory, or bad), isn’t very much information. I could put this in a simple table that would occupy a fraction of the screen space without any loss of information or ease of use. Better yet, why not provide a richer display of information about these important measures on the dashboard screen, without requiring people to drill or link to additional screens to learn more? For instance, it would be easy to add an actual scale to a gauge (preferably one that is linear rather than round, which is more space efficient), add a target or some other measure to which each metric could be compared, and add some historical information to show what’s been happening in the past leading up to the present moment.

I don’t mean to be picking on 8e6 Technologies in particular. Their dashboard is typical of what I’ve seen—poorly designed for monitoring and not very informative. Vendors that provide dashboards should take the time to learn the best ways to display information on this potentially powerful medium.

Take care,

If you develop dashboards in Excel, a new company name BonaVista Systems has just made your life easier. They have just released an add-in for Excel, named MicroCharts, which makes it simple to create bullet graphs (my alternative to gauges and meters) and sparklines (Edward Tufte’s small, bare-bones time-series graphs). BonaVista Systems first came to my attention when Andreas Flockermann, one of its founders, submitted the winning entry for the dashboard design scenario of the 2006 Data Visualization Competition, which I’ll feature in a November article. At the time, I had no idea how he created such nice sparklines and bullet graphs, but now I do. Here’s a snippet from the winning dashboard to illustrate what MicroCharts can enable Excel to do:

Click to enlarge

You might want to check this out.

Take care,

I returned last night from the midwest where I was working with some fine folks who recently started developing reports and their first dashboards using Oracle Discoverer and Oracle Portal. We spent most of a day investigating how well these tools will support their efforts to present information effectively. Until this week, I hadn’t seen Discoverer for several years, and I was hoping it had improved since its unimpressive beginning. What I discovered, however, didn’t bode well.

I recently ran across an article at eWeek.com, written in 2004, entitled “Users Applaud Idea of Oracle BI as Center of the Universe.” True-blue Oracle customers would love to pair Oracle’s great RDBMS with a solid end-to-end business intelligence (BI) solution. Unfortunately, Oracle BI tools don’t offer a real solution. The most important aspects of BI to business people — the ability to make sense of information and clearly communicate to others what they’ve found — are poorly supported by Discoverer and Portal. I heard a rumor that Oracle is shopping for BI tools to buy. I hope this is true. Discoverer has been around for years, but it still exhibits the developmental disabilities that often result from inbreeding. If Oracle wants to offer BI products that look like they belong to the same family as their fine RDBMS products, it’s time to bring in some fresh DNA.

Most BI products still do a poor job of supporting the analysis and presentation of information, but Discoverer is barely afloat at the shallow end of the pool. Even compared to Excel, Discoverer performs like a child, and not a very bright one. Its tables and graphs are primitive. Not only are they severely limited in functionality, they are also just plain ugly. For example, in this day of advanced computer graphics and high screen resolutions, there is no excuse for fuzzy, pixelated lines in a line graph. The quantitative axes of its graphs cannot even include tick marks, but only the numeric labels alone. And speaking of quantitative scales, it can’t format the numbers to display as percentages, so you’re stuck with numbers along the axis such as 0.1, 0.2, etc. rather than 10%, 20%, etc.. Data points can be included on lines in a line graph to mark the values, but neither their size nor their color can be changed.

To create a dashboard, you can take tables and graphs that you develop in Discoverer and publish them to a portal screen. Oracle Portal screens, like most dashboard and portal products, are divided into regions, and presentation objects such as tables and graphs can be placed into these regions. Unfortunately for the developer, however, there is no way to match the size of your tables or graphs to fit the space reserved for them in the portal. Once you go through the many steps necessary to insert the object into the portal, you must then examine how well it fits and go to another screen to set its size (width and height) in pixels. This takes some trial and error, because the width of the portal region isn’t measured in pixels, It’s measured as a percentage of the overall screen width. This means that you can eventually get it to look right, but only for a particular screen resolution. Change the resolution and your tables and graphs no longer fit.

You can imagine the problem that this creates when someone views the portal at a different resolution (that is, the table or graph will either be too large or too small to properly fit), but it’s even worse than this. No matter how good a job you do as a developer to size the object in the portal, when it is published to the portal and someone views it for the first time, the sizing that you worked so hard to get right isn’t preserved; it reverts back to the size it was when first created in Discoverer. Imagine what an impression you’ll make when the CEO looks at your new dashboard for the first time and it is completely unreadable — a screen full of tables and graphs that are only partially visible, too big to fit into the spaces reserved for them? The answer from Oracle technical support? The CEO can resize each object herself, going through the same laborious process that you did to get them to fit in the first place. Then, just for fun, she can change her screen resolution and go through the process again.

These problems are only a few that surfaced as I spent a few hours with my customer trying to get a dashboard to present a simple set of tables and graphs effectively. A thorough evaluation would no doubt result in a much longer list of problems.

I’m sure that BI developers who have worked with Discoverer and Portal for a while have figured out all sorts of tricks to overcome some of its limitations. Innovative individuals can often find a way to make things work — at least well enough to get by– but they shouldn’t have to. Why should the measure of IT talent be the ability to learn and tolerate overly difficult steps to do simple things and to develop tricks to work around software limitations? Come on software vendors, do your job. As for Oracle and its BI venture, perhaps its only hope is to purchase another vendor’s BI products that aren’t stuck back in elementary school.

Take care,

I’m going to share a little secret that many of us business intelligence professionals keep locked away from sight. Despite the fact that we spend much of our time supporting the information needs of finance professionals, we know less than we pretend about finance. Many of us have never taken a single course in finance and probably haven’t cracked a finance book more sophisticated than Accounting for Dummies. I’m embarrassed to admit that I’ve spent many meetings over the years discussing financial data, nodding as if I understand and peppering the conversation with finance terms here and there to reinforce the ruse, thankful that no one ever challenged me to define one of those terms.

I’m not completely in the dark. I understand the basics and have managed the finances of departments and of my own companies over the years, but my understanding is rudimentary and my use of the terms is shallow, which has at times elicited confused or suspicious looks from finance professionals when I must have used one of those terms incorrectly. Don’t you dare tell any of my MBA students at the University of California, Berkeley, that I’m a finance toddler. They might stop calling me professor.

If you’re in this boat with me, I want to recommend a new book entitled Financial Intelligence: A Manager’s Guide to Knowing What the Numbers Really Mean by Karen Berman and Joe Knight (Harvard Business School Press, 2006). The authors are the owners of the Business Literacy Institute and their many years of teaching people such as us has equipped them to explain the intimidatingly esoteric principles of finance in accessible terms.

If you ran across this book on the shelves at Barnes & Noble, you might have skipped right over it because it claims to be a “Manager’s Guide” and you’re not a manager. Not so fast. Most of the people that you support with financial reports and systems used for financial analysis are managers or are people who assist managers. Even if you’re not a manager, you need to understand the financial needs and interests of managers to support them and even to intelligently discuss their needs with them. If you read this book, you’ll suddenly be able to anticipate their needs, and think of how great you’ll look when you provide them with needed information before they even ask. You’ll also understand what people look for in financial data, how they use it, and why, which will enable you to design better reports, dashboards, etc.

If you’re a BI professional and don’t have a background in finance, you ought to get a copy of this book. If you’re like me, you might discover that reading it is actually enjoyable. It feels good to have the lights suddenly come on after years of living in a darkened room.

Take care,