5 Common Dashboard Design Mistakes

It’s always easy to criticize but you will definitely agree with the following 5 common dashboard design mistakes.

1. You try to explode the audience with data without context or objective

This is always the case since the BI project team constantly receives new requirements from different fields of users or groups. In the end, they need to come up with a design that fits all departments and users. The end result is disastrous since the only way to do this is the explode the dashboard with data and various targeted visualizations causing inappropriate display of data/media to end users The broad range of end users will not be engaged as the dashboard will contain metrics not specific to them and information they don’t understand. Moreover, there will be no actionable intelligence inside and resulting in low engagement rate or high bounce rate.

2. Key business stakeholders not being involved

Since BI is “Business Intelligence”, the focus is not on the tool itself. It’s “Business” and “Intelligence”. And where does all these come from? Certainly not from IT Tech guys. Requirements from business end users are important, where they provide you their pain points, challenges, ways to identify risks and opportunities. In the end, the goal is to help business end users generate revenue for the company. They will provide the correct metrics, the correct business view, the correct goals. IT is there to provide best practices and knowledge sharing.

3. Starting too big

One problem with involving key business stakeholders is that they have all the problems in the world. It’s hard for them to prioritize every single requirement as they are always inter-correlated and requires IT to include every metrics in the scope. Again, the end result is having no objectives. Don’t try to take on the world. It doesn’t work. BI is always never ending since requirements change. So start with areas where it is not prone to changes. Think big to cover the whole enterprise dashboard idea but start small. Think agile, think future, but be conservative to start with.

4. Corporate colors & theme

Having a corporate color and theme will always engage users. They are used to the colors used and the theme feels warm. Color is important, it gives a lot of information in terms of BI and you can’t have too much nor too little to it. Probably search around for “color psychology” and it may give you some extra insights to color. While deciding on which colors to use, check out the corporate colors. Ask simple questions:

Are the colors printable? Shades of black or shades of blue, can these be differentiated?

What about projectors? Can they be seen on the projector? If your charts and medias are done well, users will definitely use them in Power Points and emails.

Will users be comfortable with the colors? If I look at a graph with more than 6 different colors, will I still be able to analyze?

5. Collaboration & Interaction

BI is also about collaboration. Collaboration not just with other users but also interacting with the dashboard. Do you provide actionable intelligence? Users will always ask “What’s next”? While they drill down, it’s usually not the lowest level. Do you provide steps to further drill into details? Do you allow them to comment or email the relevant person to do some action?

Interacting with the dashboard also requires one huge point. This is called “Performance”. Imagine a user needing to wait 30 seconds before seeing results. If they were analyzing, then this waiting time will not engage them. 30 seconds is a lot in a dashboard. Users will want a fast dashboard due to the fact that they never know where they will end up. A business question will lead to another. The answer is never straightforward hence while they drill, they are thinking of other correlated metrics and the waiting time will interrupt their though process. When measuring performance in BI, we should be able to measure in milliseconds or speed of thought in Exalytics term.

There are a lot more points to point out on dashboard design mistakes but if you do achieve the above, you’re at 2/3 their. Others might include screen resolution, poor data arrangement, unattractive design/layout, excessive detail or precision etc. The list can go on.

If you have other thoughts to this, feel free to comment.


Author: Steve Yeung

Being in the EPM & BI field for more than 8 years, it's about time I contribute to newcomers! As a founder of MondayBI.com I wish to give you all the help I can. Feel free to give any suggestions or questions. Hope you will all enjoy this blog! William Wong Essbase Certified Specialist OBIEE Certified Specialist

3 thoughts on “5 Common Dashboard Design Mistakes”

  1. This blog post has some good advice about the best way to design dashboards. I am involved in designing OBIEE 11g dashboards, and the design is no easy task. I have learned that point number 3 is important for all BI designers to remember. Yes, your end users want good reports quickly, but you must start off slow, learn from what hasn’t worked, and expand vigorously on what has!

    Liked by 1 person

    1. Hi Kinsey, thanks for your comment. Starting a blog needs a lot of commitment and time, so your comments are very positive momentum for me. At least I know I’m helping others.

      I’ll give another advise for you. Your star schema, DW/DMs, ETL, data/meta-data governance, data enrichment, schema mappings are much more important attributes to the dashboard itself than any other. I would even say that MDM and clean data is more important than the dashboard, since the points I mentioned are the ones that takes a lot of time to do properly. Once all those are in place, modifying the dashboard is just magic. Even users can handle building reports, publishing reports & dashboards.

      So management involvement is critical. Ask your client how eager are they to do this properly. Do they want to repeat this work again or do it right the first time?

      If you start small, you will prove your methodology correct in a short time. If you follow their methodology, it might not be bad either since you can prove them wrong with little magnification to the whole picture.

      Liked by 1 person

  2. Hi, thanks for the feedback. As a functional user and manager of two subject areas in my organization’s implementation of OBIEE, I have had the privilege of working very closely with the senior programmer (easily one of the most intelligent person I know!) who actually built the star schemas for my subject areas. I don’t handle the repository or the business layer, but I know very well how to build effective dashboard designs. And you’re correct: Clean data is essential for any reporting project.

    With effective dashboard design, you are able to capture data errors quickly and effectively. Coupled with agents, finding and fixing database errors becomes–almost–effortless.

    Thank you for your work already on this blog. I look forward to checking in on new posts!


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: