I have became very close friend with Tableau recently. While there are a lot of data visualisation libraries out there (matplotlib, plotly, seaborn, just to name a few), tools like Tableau has really made data visualisation way more effective.
However, I have seen the across different organisations that even though there are endless tools with the promise of “democratising data”, very few have truly lived up to it’s promise. Most people just go back to their manually updated excel, or just simple happy when someone just send over some number that they need via a Slack DM.
Can Dashboard “Democratise Data”?
I came across this via a blog post (Are Dashboards Dead?) and found many of the things it mentioned was very relevant. While all companies wanted to be “data driven”, we are probably not going to get there anytime soon. Unless you are a company like the recently IPO-ed Snowflake that has a product that is essentially a data business and require most of the staff to have a more than average joe knowledge on data, there is always a need to look beyond data but the actual business process of it. Not all the companies on this planet is “100% digital” yet and it might never will be.
Came right across my mind is how Instagram would be a good example of how to democratise data to the billions of users it has. The “Instagram Insights” feature for the business accounts is really about showing data to user at scale. Instead of “There is a dashboard for that”, they only show the absolute essentials in a relatively tiny smart phone screen.
While there are a few other screens with some other more detailed numbers, this one is essentially it just showing 8 numbers up there and that’s it. Of course there is an advantage that Instagram profile only allow that much elements to click on a post and hence there is less key data points to show, but the idea of finding the relevant number to show turns out to be the first thing for a dashboard. No, it is not whether a vertical or horizontal bar chart to show, this is probably not a priority most of the time.
The neglected skill: finding and communicating the relevant number to the relevant people
If the chain of thought starts from there, the core component of making a dashboard is more about finding the relevant number for the dashboard. The challenge doesn’t stop there, imagine even it’s in the same company or even the same team, the perspective of data could already be completely different. This “Activation” could be different from that “activation”. When you are building the dashboard, apart from thinking about dashboards, you have to find the relevant number as well.
An even more challenging thing is a dashboard are meant to share data and insight in a much user friendly method yet behind those beautiful bar charts and percentages, the assumptions behind all these different dashboards would quickly take over the conversation. Everyone would tend to held on to their own numbers, that’s what they have been dealing with day in and day out. Apart from finding the relevant number and plotting nice bar charts, it’s also about communicating and persuading the number across to the others, and sometimes, the whole organisation.
Assembling IKEA furniture vs finding the right piece of wood at Home Depot
So what does this all have to do with IKEA and Home Depot? It’s an analogy that I came up with after writing this.
If a dashboard is like a piece of furniture, then making a dashboard is like assembling the furniture that you came across in IKEA.
But most of the time, it actually more like finding the right piece of wood or the right screw to assemble the dashboard. Just like Home Depot.
And after all that, you have to convince people that this table that you build in this dimension is the most relevant piece of furniture in the office. Probably right next to that ping pong table.
So here we are, when we talk about Dashboards, we are really talking about IKEA or Home Depot. While both table have four legs and can held things, they are a different table. And yes, there should be a table there too.