Value-Based-Design: What Is It & How Can it help your B.I. Solution?
- 16th July 2019
- Big Data & Advanced Analytics
- Luke Sharma
One of the hottest topics in data and business intelligence right now is storytelling. This links back to data literacy and the aim of storytelling is (you guessed it) telling a story with your data! Simply put, what does your data tell you, and what can you do about it?
Storytelling isn’t just putting together a pretty dashboard with lots of shiny visualisations on it. Let’s take this dashboard as an example;
This is a dashboard that I built in my “bootcamp” at Peak before we had training in dashboard development and let’s be honest, it doesn’t say much. Although it looks (relatively) nice – what is this data telling us? That there’s 12.84k sales in this car company, and that they are severely in the red. That’s about it really, the map looks great, but there’s no actionable intelligence. Is there anything that catches your eye and makes you think “oh we need to find out what’s going on there!”? Not really.
Telling a story with your data sounds like a simple and straightforward action to take – but how do you go about organising millions of rows of data or hundreds of separate tables into detailed, actionable intelligence?
At Peak, when we build our dashboards in Birst for clients, we have found that the most effective way to do this is by using Value Based Design. Value-Based Design is a theory that is focussed on displaying data in a way that the end user can get the most value (rather than just flashy visuals).
We do this in 3 steps;
1) Display – Displaying some data where we can spot an issue
2) Diagnose – Diagnose the business area, location or employee that is having an issue
3) Decide – Using the data to decide on a suitable action to take to remedy the issue
Using this basic landing page, we are displaying that this organisation’s billing amount and cases have been dropping year by year. We can see this in both the conditional formatting on the KPI as well as the spline chart.
So the user can drill into this “Billing amount” to diagnose the problem, which takes them to a more detailed dashboard:
On this dashboard we can clearly see where the company is losing business – conditionally formatted maps combined with the bar chart indicate that there have been big problems in the MidWestern States over the past 3 years. Let’s click into one of them to carry on our diagnosis:
So in Ohio – using these charts, we can diagnose that there are particular problems in the Aerospace, High Technology and Manufacturing industries. Most key is the Aerospace, as using the bubble chart and pie chart, we can see that this is the industry that generates the most money for this state. Let’s investigate further to try to make some decisions on how we can repair this area:
Using a detailed analysis, we can see that although the aerospace industry makes up a huge chunk of the money that goes into the company from this state, we only have 1 customer from this industry. Not to mention that from signing in 2015, this particular customer has been decreasing their billing! Perhaps we can decide to look at diversifying to provide services to more companies in the aerospace industry, or even just looking into the issues that this particular customer is having with this software (as a means to recouping losses).
Using value based design, in 4-clicks we have displayed an issue in the data (decreasing billing amount), diagnosed an issue in the company (Ohio sales have dropped, especially in the Aerospace area) and made some rectifying decisions (diversifying customers in the region, or improving customer retention in the area).