Improvise, adapt, overcome! How to increase adoption and user satisfaction of dashboards

[HamzaJ]

Improvise, adapt, overcome! How to increase adoption and user satisfaction of dashboards

Let's start at the beginning...

We have been using Sisense for over 6 years now. We started on a Windows deployment and have since migrated to the Linux version of Sisense. In those six years we have encountered all types of users;

• The Know-It-All Novice
  They claim to have extensive knowledge about data but often misinterpret basic concepts.
  
  
• The Excel Enthusiast
  Believes that Excel is the ultimate tool for any data-related task
  
  
• The Chart Junkie
  Loves creating charts and graphs, regardless of their relevance or accuracy
  
  
• The Privacy Paranoid
  They're extremely concerned about data privacy to the point of absurdity.
  
  
• The Data-Driven Dreamer
  Has grand visions of data transforming the world but lacks practical skills to implement anything.
  
  
• The Expert Fact-checker
  Compares all results with their analysis and submits support tickets on why the results are not the same

Although this is meant as satire I think we all recognize either ourselves or someone else in these personas. We have at least met them all in the last six years and every interaction with these users made one thing clear for us; dashboards (and in extension widgets) need to be

• easy to understand
• free of confusion 
• clean and concise
• well documented
• simple to use

With the ease of use of Sisense, in terms of creating a dashboard, we should be vigilant not to turn into a Dashboard Diva (Obsessed with creating dashboards for every possible metric, regardless of their importance). Spending time on a dashboard can make us blind to the actual understandability of a dashboard. Things that can be very logical and clear for us (... it is obvious that X Y and Z are like this), might be illogical and complex for regular users or even other designers. Our end-users are primarily doctors, nurses, and other types of practitioners. So not necessarily data power users. 

Due to this reason, we are always looking for ways to improve the 'readability' of a dashboard. It starts with simple questions such as;

• Does every dashboard have a title?
• Does every widget have a concise title and description?
• Is the color usage the same between charts?
• Is there a general description explaining the dashboard (we do it with a video)?
• Are all relevant filters available?
• Do we apply the dashboard pyramid (KPIs on tip, tables at the bottom)?
• No more than 9 widgets?

Although these are a good set to start with, it does not always results in a 100% score on readability. This led us to continuously look for improvements of the dashboards and its widgets.

Extending widgets and improving adoption

As mentioned before. we are continuously looking for ways to imrpove the readability of our dashboards and widgets. On the one hand we get valuable insights from this community and on the other hand we check-in regularly with our end-users to validate dashboards, widgets and assumptions. 

Filters

The filters are not intuitive for everyone. Often users will collapse the filter field, not knowing how to restore it and get lost or frustrated because they cannot filter on their parameters. 

We first started out with using Blox to create widgets that look like filters. A good example of this is given by Harik on Blox Filter . It is a easy way (once you know how) to create all kinds of widgets that users can use to filter. The only downside is the customization options and the limitations of Blox. We looked around and stumbled upon the Advanced Filters Plugin by Paldi.

It supports single and multi-select, date-range (with a calendar), filter dependency , easy excel-like search and the ability to create filter presets. Our users have been ecstatic about this feature, especially in combination with Filter Bookmarks . Which also makes it possible for designers to create preset filters that end-users can select and use

Widget filters

When we noticed the uptake in dashboard adoption due to easier filter usage, we wanted to further integrate this into the dashboard, especially apply this to widgets. Again thanks to Harik, we found scripts that can be applied to widgets to either create filter-buttons and dropdowns on a widget level. This works very nicely and gives the end-user the ability to change the insight of a chart. For us it made it possible to see the average treatment hours per discipline per diagnosis, either for all diagnoses or for a specific one. Without the need to filter all the data in the dashboard.

The downside is that applying code is tedious, especially if you want to apply it to multiple widgets. Plus not every designer feels comfortable doing this (and also it is prone to error when configuring it)

Fortunately Paldi's Advanced Filter widget does not only allow to create filters on a dashboard level, but also apply it to specific widgets. Now our designers are applying this to almost every dashboard. Simply selecting widgets and setting filters and they are done and ready to go.

Dynamic Widgets

Creating a consise dashboard can be a challenging task , especially when there are a lot of insights that are related to one and other. Before you know it, there are 5 similar widgets each altered a little. To combat this we have been diving into the ability to change dimensions and measurements in widgets. Similar to previous examples we did this with Blox or Javascript. There is a nice community post to create a dimension switcher in Blox (https://community.sisense.com/t5/knowledge/using-blox-to-dynamically-change-a-dimension/ta-p/8701). This one is on a dashboard-level and can even be used to change dimensions in multiple widgets ( we use this in one of our product-dashboards). 

If you are looking for a way to only change it on a widget-level, then this post can be of assistance. Eventually we migrated away from code and are using a plugin called Widget Toolbar which gives us the ability to , without any code, set per widget(s) the changeable dimensions, measurements, breakdowns and chart types. Due to the easyness of use our designers are implementing it in almost every dashboard. Here is a link to the plugin

Dynamic Values

There are use-cases where a value is dynamic and cannot really be captured in a database (unless your application has the ability to capture this information). For us it came down to changing ward-capacity and the ability to dynamically calculate products based on current registrations plus additional input (admission days and treatment hours). 

We started out with a Blox-widget that made it possible to let users dynamically input values that can be used in other widgets. A link to the how-to can be found here: https://community.sisense.com/t5/knowledge/input-parameters-using-blox-for-what-if-analysis/ta-p/183...

Again it was a code intensive solutions which was not suited for our designers plus values that were inputted were not saved when the page was refreshed. So we found our solution in a Paldi Plugin called Input Parameters this allowed us to easily add a widget with the amount of input parameters we needed within seconds. Afterwards we could easily use placeholders in our formulas and let it do the rest. A very nice addition is that the inputs are persistent (optional). So when a user came back a week later, the same inputs were still there 🙂

To be continued...

I can write and share even more great additions that we have implemented, perhaps this is interesting for a different post as this one is already becoming quite large and extensive 🙂

Let me close of by saying the following things:

• If you need another set of eyes to look at your challenge, feel free to reach out. Happy to help
• The power of Sisense is , amonghts other things, the connectability via Javascripts and APIs. We have created so many extensions, such as creating word-documents based on elasticube/dashboard-data via Blox. We are happy to share how we are doing this for the interested Sisense-ers.
• Because there are so many scripts available for which it can be difficult to find them, I have creat...
• Some of the above mentioned plugins are premium plugins offered by @Ravid_PaldiTeam , @Paldit . It might be not be an option for all, however they also offer great free plugins that I can really recommend:
  • Advanced Jump To Dashboard --> JtD without any scripts/coding
  • Advanced Text Widget --> Similar to smart label/blox but easier.
  • Advanced Tabber --> Widget tabber without any scripts/coding
  • Indicator Card  --> Combine indicator and charts into one
  • Advanced Tooltip --> Add additional information to a chart
  • Folder Organizer --> Organize all folders and dashboards easily
  • Default PDF Options --> Set default PDF options without the need to configure this in every dashboard
  • Share Dashboard View --> Share a dashboard URL with filters. As an OEM we support a lot of our customers and this makes it very easy to get their view right away
  • Avoid Over Filtering --> A handy plugin we use for our benchmark dashboards so the peergroup stays above a certain threshold