View more articles in
Techniques >

Defining and Creating an Effective Tool for Management

By Melissa Stratman,
Coleman Associates

A Performance dashboard is a focused, clarifying snapshot of key data that communicates to all viewers—even a novice viewer—the status of efforts (performance) and what successes and failures the organization or project is having.

A performance dashboard is much different from a spreadsheet or report, and must be:

• easy to read and comprehend

• focused to convey a need for action in specific areas

• not overly dense with data

• distilled down to only a few key data points

• up-to-date

Like the one in your car, a dashboard must convey many key pieces of information at a glance and in only enough detail to cause an actionable response (putting your foot on the brake when you see the speedometer needle pass the speed limit mark). Click on this link for an example of a nifty performance dashboard: Happy Little Clinic Dashboard. Print it out and refer to it as you read through this article!

Easy to Read and Comprehend

A dashboard should require no explanation to anyone within your organization about the meanings of values or graphs. As the key function of a dashboard is communication, it must not be laden with small print or unnecessary details. The visual quality should be like a traffic light: it should cause a visceral response in the viewer. This response is often a result of clever choice of color, arrangement, font and graphical display of data. For example, the use of green, red and yellow to reflect good, bad or adequate performance is intuitive and easy to understand.  A very well designed and highly visual dashboard will cause the viewer to worry or feel satisfaction at even a quick glance.  

Graphs on a performance dashboard must be well designed to highlight trends and should be much more visibly understandable than rows of numbers in a spreadsheet. They should also use consistent axis labels to make it easy to make visual comparisons over multiple graphs.

Focus on Data to Influence Behavior

Dashboards focus the viewer on key data variations and discrepancies in order to influence behavior. Areas of concern or success are emphasized using arrows, inserts, circles or color codes around key data points. The old saying “a picture is worth a thousand words” is epitomized here: visual communication is what it's all about.

This visual nature communicates both quickly and accurately where energy and focus should be directed. Data is arranged by grouping related issues, demonstrating trends that cause the eye to be attracted naturally to key areas that require focus.

Most dashboards fail to communicate because they are too dense with data. Instead of being an efficient, one-page, visual, tell-me-what-I-need-to-know tool, many organizations label their collection of multi-page spreadsheets and databases a dashboard. These worksheets are not the same things as a performance dashboard—they are simply spreadsheets full of data.

Spreadsheets generally don’t get examined closely—except by the person or team creating them—because they are overwhelming, irrelevant and/or poorly organized. Without a visual emphasis on key points, all the data becomes equally unimportant. Remember, the key function of a dashboard is communication. While emailing spreadsheets could be technically considered communicating data, a dashboard uses elegant and informed design in order to communicate effectively.  Simplicity of data is the secret to effective, comprehensible communication.

Use a Few Key Data Points

Too often dashboards fail not because they lack key data points but because they don’t discern the few key points from all of the other data points available. A dashboard must be created by someone intimate enough with the data (usually a key team of process superstars) who can draw conclusions about performance from a few (no more than ten to fifteen) data points and focus the viewers on those points. Key data points must be determined by the aspirations of the organization or project for which the performance dashboard is being created. Take a look at the Happy Little Clinic Dashboard for an example of one clinic's important issues.

Many Electronic Health Records (EHRs) come with data dashboard functionality, yet they are difficult to tailor and they don’t allow for ongoing adjustment to suit the ever-changing aspirations and focus of the organization. In short, too much data becomes the enemy of a focus on performance and outcomes.

Use Current Information

Finally, a good data dashboard must be up-to-date to reflect data that just happened. It should never be more than one week old so as to be able to influence behavior. Organized but aged data is not a dashboard, it is a report—and a historic report at that.

Recall that a dashboard is a tool created to communicate and influence both behaviors and impact performance. Now, imagine that you are a driver of a car and your dashboard is altered so that you could no longer see your speedometer while you were driving. You could therefore not glance down nor react while the performance (your driving) is happening—thus you could not affect your outcome in any informed way. Instead of having a speedometer, imagine that you could get a printout—once you arrive at your destination—of  your speed and its variations. That report may well be ‘interesting’ but it is certainly not useful as it is too late to cause an action to correct outcome. An aged report is NOT a dashboard.

Why Does my Organization or Project Need a Performance Dashboard?

The primary value of a data dashboard is to focus everyone on performance and outcomes. Dashboards are powerful because unlike data reports, a dashboard taps in to the tremendous persuasiveness of visual perception and causes a physical reaction to what is displayed.  By using a dashboard effectively a team can translate that reaction into a response to make changes and improve performance.

Without a performance dashboard, collecting data usually becomes the responsibility of either no one or a small cadre of data wonks who cannot, by themselves, create a change in behavior. And those who collect, enter and graph the data cannot effectively communicate the outcomes shown in that data to those who can influence behavior and future outcomes.

Many health care organizations have data to drive their quality, financial or clinical operations. Where they fail to make the leap is in communicating the data—not simplifying, focusing, or making it available to those individuals who need to alter their actions to impact performance. Managers typically have the data, but front line staff have no idea how their actions affect performance.  Having a good data dashboard makes sharing and communicating outcomes easy and stimulating and it sparks both conversation and action.

Creating a Performance Dashboard—Getting Started

The first step to creating a performance dashboard—and often the most difficult—is to make a clear decision, from a project leadership or management team perspective, about what outcomes (data) are most important. The best way to avoid the universal conclusion that “it’s all important to us” is to instead focus on answering one powerful question: “What do we aspire to?” Everyone on the top management team should together answer that question in a very clear, concise way.  This step may highlight your organization's lack of focus on priorities—which perpetuates a lack of focus on the data.

Once you have your aspirations distilled to your top goals, then brainstorm all the data points that could get you to that goal. As you do your brainstorming, remember the adage that “when brainstorming, nothing is ever wrong”.  Be sure to encourage wild ideas about potentially revealing data and capture on a flip chart all possible data points that could help you monitor your progress and point you toward your goals. Once that brainstorming is done, there’s another tough step ahead—narrowing the possible data points down to a few key points that frame your dashboard.

Selecting Data Points for your Dashboard

From your list of brainstormed data points, now select the key data points that will help you reach your goal. Whittle them down by eliminating those measures that do not directly help you gauge your progress. Then review the remaining data points and select only those measures you really need.

Watch out for pitfalls that will take you off course and make your final dashboard much less useful. The most common pitfalls are often uttered by someone around the management table:

• "This is a data point we already have access to so it makes sense to include it."

• "This would be a very interesting “thing” to know."

• "I went to a conference last week and they said that we should be paying attention to this."

• "You can never have too much data about this area of operations."

• "It’s a confusing data point to understand but it’s important because…."

It is easy to latch on to a data point that you already have access to, but in order to make critical decisions, your data must answer critical questions. If you are working to improve the health of your diabetic patients, knowing their HgA1C is a good data point, as might be variations of their daily glucose levels. You could also record the color of shirt that they wear during each blood draw; after all, it would be easy data to collect—but just because it is available does not make it relevant. 

Bottom Line: If you have a data point, but it does not specifically get you to an answer that tells you how you are progressing toward your goals, cross that data point off the list and do not include it in your performance dashboard.

It’s very tempting to get caught up in looking at data that is interesting. Data has a reputation of being boring, so any point that promises to spice things up is certainly tempting, even to the most disciplined.  However, success isn’t achieved by glamorous distractions, but rather through diligence.  Any data point that promises to be interesting or exciting without drawing attention to your progress regarding goals falls directly into the category of 'bright, shiny objects". Bright, shiny objects must be crossed off the list of potential dashboard data points.

Another data point pitfall comes from outside the organization, usually because of someone’s recent attendance at a conference, class or meeting. It’s very easy to hear a compelling speaker and get wrapped up in a data point du jour. Before getting distracted, you must determine if that data truly helps you to directly gauge your progress in meeting your aspirations or goals. If it doesn't directly help you to understand your operations, then that data point is simply another bright and shiny object and should be ignored.

Yet another pitfall is the notion that you can never have too much data. “The more the merrier” works for many concepts, but not for a performance dashboard. Not only can you lose focus, but too much data also highlights a lack of understanding about exactly how each data point directs you to meet the goals of the organization. Since the maximum number of useful data points on a dashboard is ten to fifteen, each point must be chosen wisely in order to cover all key areas.

The final major pitfall of narrowing down key data points is that of using a data point that is so complex that it has little value to most who review it.  Stephen Wolfram in his book A New Kind of Science challenges the notion that complexity must be responded to with complexity rather than simplicity. Data for a dashboard must be easy to understand and any point on the dashboard must be able to stand alone and tell its own story. Esoteric data has no place on a performance dashboard.

(Cartoon from the December 6, 2010 New Yorker Magazine.)

Using your Performance Dashboard

Knowing what data goes on a performance dashboard and arranging it in a clear, concise and compelling way is the most challenging part of setting up a performance dashboard. Once the dashboard is created it should be tested on a small scale before being put into regular use. Understand however, that a truly useful data dashboard must be continually honed and adjusted in order to meet changing and evolving data needs and operational outcomes.

For more information on dashboards, check out Stephen Few's book, Information Dashboard Design: The effective Visual Communication of Data.