24 Feb Common mistakes in creating Management BI dashboards
KPI portals, known also as Dashboards, are the most popular means for delivering focused measurements to an organization’s management. However, many of the deployments of these dashboards do not pass the implementation test and are not being used. This article highlights some of the Do’s and Don’t-Do’s in the domain of a BI dashboard implementation.
What is a Dashboard?
The purpose of a dashboard is to graphically present a swift and clear overview of the organization’s status. This allows detection of areas where there is a need for action. Moreover, it allows for drill-down analysis in order to understand the required course of action.
The dashboard should facilitate answering the following questions for a manager:
1. What information requires my attention?
2. When should I take action?
3. What action should I take?
Dashboards include graphical objects since it is easier for a person to scan, process and understand visual information rather than textual information.
Why is it so hard to successfully implement a dashboard?
It is important to understand that even the use of the most advanced visual tools will not provide an adequate solution if the selection of measures does not fit the organization’s business goals and if the visual design is not done properly.
Implementation of a dashboard that provides an adequate solution consists of two parts:
What are the measurements to be presented?
What is the right way to present them?
Gathering a large amount of information on one page, without it becoming a clatter of data, is a challenge. The importance of choosing the right data to be presented to managers is critical. Wrong business measurement might lead to erroneous decision making.
Typical errors when implementing dashboards
- The IT as a sole leader of the dashboard implementation
- Choosing measurements that do not depict the organization’s objectives
- Choosing the wrong visual object
- Over-crowded visual objects on a single screen
- Overuse of “hazzle dazzle”, colors and decorations
- Unappealing UI design
Deployment of the dashboard without incorporating it in management routines
Error #1: The IT as a sole leader of the dashboard implementation
There is critical importance to the incorporation of a business sponsor in the process of management dashboard implementation. This is true even if the BI unit has deep knowledge of the business and of the requirements. The design of a dashboard needs to imitate the way of think process of the business user. This is typically a new design process and skill for the BI team.
It is important that the placement of information on the dashboard be done according to the thinking flow of a business manager. For example: if a manager notices a decline in sales, the immediate questions could be: “what products/services cause the decline?”, “Is the decline originating from a specific business unit?”, “What are the consequences on the company’s profit?”, “Could this be a result of a change in campaign effectiveness?”
Another aspect is the implementation/assimilation – a business sponsor that is involved from the initial design steps is more probable to make sure of the implementation across the rest of the business users in the organization. He is the one that will lead their involvement and cooperation through the whole process, from KPI definition, through dashboard design and to the implementation. The business sponsor can also help mitigate resistance to change in the implementation process.
It is important to note that even a perfectly developed dashboard can be abandoned and not used if the implementation process is done without adequate resources and care.
Error #2: Choosing measurements that do not depict the organization’s objectives
Organizations invest a lot in acquiring and storing large amounts of diverse data for the purpose of extracting business value from it. Sometimes, the mass of information that is brought forth is overwhelming and becomes a burden rather than a focused tool for managers to manage their organization.
The challenge therefore is to sort out the KPI and measurements with which the management has to focus on. Measurements which represent the state of the organization and the adherence to its targets and direct actions needed when necessary. Moreover, choosing a wrong KPI is likely to lead to misjudgment by management followed by damaged business and loss of efforts.
An example – an organization with a strategy of enhancing revenue from sales chooses the following KPIs: “quantity of products sold”, “sum of monthly sales” and “number of campaigns activated per period”. This selection of KPIs do damage. A rise in the quantity of products sold might not have the same latitude of impact on the revenue (since revenue depends on the selling price of each product). Moreover, counting the quantity of sold products without deducting the quantity of returns is an erroneous figure to present.
The sum of monthly sales, even if presented as a trend graph over a period of time, doesn’t present clearly the trend in sales. A Better presentation should be “% change in monthly sales” which should be measured against the previous period. The KPI “campaigns activated per period” will even do damage since customers will start receiving numerous and not personalized campaigns which in turn will reduce customer satisfaction and will not be good for business.
Error #3: choosing the wrong visual object
In the process of designing a dashboard their needs to be decisions as to what visual objects to use in order to best represent the wanted data message: table Vs graph, tree map Vs quarter diagram etc.
The objective is to represent the data in such a way that the viewer’s attention will be occupied by the most important part.
Here are some examples:
In which of the following graphs is it easier to notice the pattern of change over time?
Drawing 1: Line Graph Vs. Bar Graph
In which of the following graphs is it easier to understand who is the leading service team and the difference between team 1 and 6?
Drawing 2: Pie Chart Vs. Bar Chart
In order to understand the Pie Chart we need to read the figures written on it. On the Bar Chart you can figure it out with a quick glance. Whenever we are required to read text, the graph loses its value and presenting the effect will be the same as if they were presented in a table.
Error #4: Over-crowded visual objects on a single screen
The presentation of multiple data entities together on one screen has great meaning and value. The ability to scan in a single view multiple data figures allows for a cognitive correlation and understanding process. This would be different altogether if each data entity was presented separately.
It is common to find dashboards that couple together various data objects that have no relation between themselves. Furthermore, overcrowding the dashboard with many data objects causes a scroll-down effect that splits the view. In these cases, the load of information does not allow the user of the dashboard to focus on the data and to identify what requires treatment.
Drawing 4: Over-crowded dashboard containing unrelated data objects
Drawing 5: Overloaded and unfocused dashboard
Error #5: Overuse of “hazzle dazzle”, colors and decorations
Since dashboards are graphical by nature, a majority of the implementations tend to incorporate an overuse of colors and graphical objects. These can be impressive at first site but will not be effective for use since they distract the viewer’s attention from the data itself.
Colors play an important role in a dashboard and one should use them with special care:
Red colors have a negative meaning – a warning or below expectations performance.
Green colors have a positive meaning – performance meeting expectations or beyond.
homogeneous use of colors allows the viewer to correlate objects and eases the data processing.
The dashboard below is an example of an excessive use of colors. The use of a different color for each column has no value. Also, the variety and contrast of colors has no value. On the contrary, they put an unnecessary load on the viewer.
Can one understand from this dashboard what should be the current focus and what requires attention?
Drawing 6: excessive use of colors in a dashboard
Error #6: Unappealing UI design
A dashboard should be esthetic and attractive to the viewer’s eye. Especially if it is intended for frequent use. An unappealing dashboard passes a message of “unimportance”. Following the previous mentioned errors, the challenge is to find the right balance of using colors and decorations.
Drawing 7: An unappealing dashboard
Error #7: Deployment of the dashboard without incorporating it in management routines
The importance of a business sponsor for the implementation process of an IT system is obvious. Many perfectly designed and developed IT systems have failed due to not enough effort and attention during the implementation phase. But this is not enough for proper implementation of a KPI dashboard.
A management KPI dashboard which is not incorporated as part of the management reporting routines will not last.
The different stakeholders in the organization should know that they are being measured and monitored by their managers using the KPIs in the dashboard. The managers will raise inquiries based on the data in the dashboard and will expect to receive timely and accurate answers. The management routines can be at different levels and at different reoccurrence rate (Daily, Weekly, Monthly).
- The key for a correct dashboard design is understanding the thinking process of the managers, the important areas of performance for achieving business goals and the available courses of action for achieving these goals.
- Throughout the whole process it is important to allocate the right business sponsor that will guide and lead it. This will be followed by a commitment to use the dashboard.
- The dashboard should direct focus to the most important data in a clear and simple way.
- It is cruicial to remember that the technology’s place in this case is to support and enhance the business, not the other way around!
Daniella Elkouby-Fisch is VP of projects & consulting at Nogamy. Daniella has vast experience in leading implementations of KPI dashboards in many of the leading organizations.