Dashboard Display Best Practices

Dalam beberapa pekerjaan Dashboard System menggunakan Tableau, FusionCharts atau custom produk, kami berusaha menjawab kebutuhan display atau bagaimana data itu ditampillkan. Berikut adalah best practices nya

The Dashboard Panel
Unilyzer Dashboard Panel
Dashboard Display Best Practices:
1. Use Visualization
2. Show Key Metrics
3. Show Direction of Change
4. Show Degree of Change
5. Show Comparisons
6. Provide Data Source
Practice 1:  Use Visualization Objects – it is common knowledge that a majority of people process information better when it is displayed visually.  Using charts, graphs, and images is an effective way to present a lot of information in a meaningful way, and to do it with limited space.  Data dashboard displays should use the most effective visualization object, not just random charts and graphs.  For example, you would not use a Funnel or Pyramid object to show month-over-month revenue changes.  Here is a brief, short list, of display objects and how they can be used effectively:
  • Speedometers –  these are interested, cool, and fun objects to use on a dashboard display.  But they should be used in a meaningful way.  For example, a good use of speedometer objects is to show levels and where those levels are relative to boundaries.  For example, let’s say that you are an Human Resources Manager, and you want to maintain a compliance rate of 90% ,  a rate of 85% will get you in trouble, and a rate of 95% will get you a bonus. In this scenario, a speedometer gauge would be appropriate: set the range of the gauge from 80% (bottom) to 100% (top), then color 80-85% RED,  85% – 94% Yellow, and 95 – 100% Green.  Now, all you have to do is keep an eye on the needle, if it is in the yellow range you don’t need to turn attention to the matter, if the needle is in the Red then it’s time to investigate, and if it’s green you have cause for celebration, or at least, should pat yourself on the back.
Speedometer Visualisation Objects
Speedometer Visualisation Objects
  • Funnel – this visualization object is best used to represent a filtering process.  For example, if you are measuring sales conversions, and the middle steps are: 1) “Fill the Funnel” – drive traffic to your store, 2) present your product, 3) close sale.  Now, since this is a clear process, with a diminishing number as we move through the process, this is a good scenario for using a funnel visualization object.  For example, let’s say 1,400 people came to your store, 600 listened to your pitch, and 100 made the purchase – using a funnel object is a good way to represent this flow.  One could argue that a bar graph could show the same thing, if from left to right, you showed bars dropping from 1,440, to 600, and finally to 100 – however, the viewer in this case, would require an explanation because these would look like 3 independent data points with nothing indicating their relationship – on the contrary, using a funnel implies to the viewer that there is a filtering process happening – and this, allows the viewer to start thinking about the key drivers of each of those steps in the process. Bottom line: properly selected visualization objects communicate more effectively. For illustration purposes, figure 1. shows good use of a funnel visualization object to present data associated with a process – a sales process in this example.
    Figure 1. Professional Use of Funnel Object
    Figure 1. Professional Use of Funnel Object

Practice 2:  Key Metrics – present information that matters with sufficient detail, but not too much detail.  Usually the viewer will need high-level information about many different things and thus, should not be distracted with too much detail.  Once the viewer reviews all of the important metrics, he/she can decide on which metrics they would like to “drill down” on.  To drill down means to inspect the underlying detail by which the metric was derived or calculated.  Bottom line: show what the viewer things that make a difference, or things that are important for success.
Practice 3: Show Direction of Change - it is good practice to show the direction of change along with the metric.  For example, if you show profit margin of say 25% on your dashboard, you would also want to give an indication of whether this is an increase or decrease relative to a prior period, or to a target.
Practice 4: Show Degree of Change – show the viewer the magnitude of change. For example, in the image to the left,  the metric value for “Pageviews” was 13,217 for the period, the direction was down compared to the previous period, and the degree of change was 14%.  With out the direction and magnitude of change, the number of pageviews alone does not tell much of a story.  Another example of this is how theUnilyzer Social Media Dashboard uses it’s dashboard to show key metrics, the direction of change, and with a mouse over for the inquisitive viewer, reveals the magnitude of change.
Practice 5: Show Comparisons – a powerful way to provide insight with a dashboard is to provide comparative metrics on a side-by-side basis. The Unilyzer panel is a perfect example of how to build insight, or business intelligence, into your dashboard’s story.  This exerpt from the Unilyzer Dashboard shows the social media marketer, not only the key metrics for each social media channel driving traffic to their facebook fan page, but shows how traffic from each channel measure up to each other.  At a glance, the dashboard viewer gets more context and meaning to these key metric values and trends. This is a perfect example how to paint a clear picture of social media performance.  Thus, the Unilyzer Dashboard is one of my favorite Social Media Marketing Performance Management Dashboards.
The Dashboard Panel
Practice 6:  Provide Data Source – make sure to provide your view with information about the underlying data because if the values shown are not believable, the viewer will need to understand things like:
  • Data Ranges
  • Data Sources
  • Date Published
Knowing these things will help the viewer reconcile differences between their expectations and the values represented. It is not uncommon for management to have expectations of metric value results. These are based on ‘gut’ feel and not empirical data, so when/if dashboard values contradict their expectations,the data source information will be useful support.
Well that wraps up 6 Dashboard Display Best Practices. I hope this gives you ideas and perspectives that will help you develop your own dashboards.