Skip to main content
cancel
Showing results for 
Search instead for 
Did you mean: 

The ultimate Microsoft Fabric, Power BI, Azure AI & SQL learning event! Join us in Las Vegas from March 26-28, 2024. Use code MSCUST for a $100 discount. Register Now

Ritaf1983

When Big Numbers Become Big Problems

A few days ago, I had a conversation with a colleague about Data Visualization.

 

I explained that in data visualization, it is the relationships between numbers that matter, rather than the numbers themselves.

 

His opinion was different. From his point of view, the dashboard should have numbers because that's how users see the main KPIs.

 

In response to my inquiry about how exactly a single number could provide insight, he gave an example of the large BANS that appear at the top of the screen and, according to him, explained the situation in seconds. This is indeed the opinion that many dashboard developers hold, including myself in the past.

 

Why do Bans/cards on dashboards become so common?

Ritaf1983_1-1696513818174.png

1. " Experts" recommendations such as:

"Don’t bury the most important fact, your KPI, in a chart. Show it loud and proud as a Big A** Number (BAN)!"

2. Considering a dashboard as a stage for telling a story. The big and beautiful numbers go well with the idea of presenting the main point.

3. The desire to impress. Sometimes managers have a fantasy of seeing running numbers with indicators on the plasma, and that's how they think they know what the "pulse" of the business is. Because we have a tool that can make it happen easily, we run to meet their "desires".

The result - is "cards" with large numbers like in the picture below appearing in most of the dashboards.

Ritaf1983_2-1696513921994.png

As I began investigating the real goals of data visualization, I realized graphs are a functional tool, not a decoration. Dashboards are not a stage for a story, but a monitoring tool to improve a situation. All of this should be accomplished with a minimum of time and effort. Part of the change was that cards simply disappeared from the screen without being noticed.

 
Why don't BANS really help to monitor and understand the critical points of a business?
Ritaf1983_3-1696514111108.png

1. By using this format, the whole is separated from the parts. It is like putting a central piece of the puzzle on top, not in its place where it belongs. The result - on the one hand, the center part is separated, on the other, the picture has a "hole". It's not a good way to develop a dashboard.

2. These BANS take up a lot of space on the screen.

Below is a picture of a detailed table that is almost the same size as the Cards area.

Ritaf1983_4-1696514224139.png

3. There is an "advantage" to decoration, but it can also "hide" your insights. In the picture, it appears that the month was successful in total and that there was an increase from last year. But when we look at the full picture, one of the stores has seen a drop in performance of over 40%... Maybe not so good after all.

Ritaf1983_5-1696514256898.png

Final thoughts

 

The cards have simply disappeared from my last dashboards without me noticing.

 

It requires a lot of work and scientific precision to create a correct and detailed visualization. However, when I see the use of dashboards is leading to discussions and changes in processes, I know it was worth it.

 

*All graphs were created using POWER BI

 

Let me know what you think by giving us a thumbs up,

Greetings, Rita 

Ritaf1983_0-1696518148601.png

 

 
 
 
 

 

Comments

Great work, Rita! How did you introduce the change to the users? Did you simple publish a new version of the report? Did you announce it? And how did the users react? Was there feedback like: Where have my numbers gone? I want my old dashboard back?

@Martin_D 
Thank you for your feedback, and great question.

Of course, I don't sabotage anyone's reports without their knowledge

Work on the right visualization starts already in the definition stage if I start working on a new report/dashboard. At this stage, I delve into the business process for which the dashboard is being built, the required metrics, their granularity, what is a good range, and what is a range that requires improvement/immediate response.

The question I don't ask at any stage is how the user wants their reports to look. This is my expertise and "headache".

If the situation is the optimization of other reports, a conversation is still conducted with the user about the processes, and of course I update him that the visual appearance of the reports will change.

In the event of resistance to one visualization or another, and this happens because I don't build reports that everyone is used to.

I show him two options: his preferred option and the one that should be, and ask an informative question (not what he likes more), and then he realizes for himself what is the most efficient way and flows with me.

Here is a more natural-sounding translation:

Of course, I would never sabotage someone's reports without their knowledge.

The process of creating the right visualization starts in the definition stage, when I'm first starting work on a new report or dashboard. At this stage, I take a deep dive into the business process that the dashboard will be used to analyze. I identify the required metrics, their granularity, and what constitutes a good range of values. I also identify any areas that need improvement or immediate attention.

I never ask the user what they want their reports to look like. This is my expertise and responsibility.

If I'm optimizing existing reports, I still have a conversation with the user about the business processes involved. I also let them know that the visual appearance of the reports will be changing.

If a user resists a particular visualization, it's usually because it's not what they're used to. In these cases, I show them two options: the option they prefer, and the option that I believe is the most effective. I then ask them an informative question that helps them see the benefits of the more effective option.

In most cases, this approach is successful in getting the user to agree to the more effective visualization.

I hope this is helpful! Let me know if you have any other questions.

Thank you! I'll give it a try 🙂

@Martin_D 

Good luck! Please ensure that your version is meticulously designed by choosing the appropriate visuals and formatting them at the 'perfect pixel' level.

Thank you again for the interesting question; I believe it deserves its own post 😍