Get MAD with your data
A good dashboard (application, report or however you want to define it) needs to follow some rules of best practice design to be understood by the end user as successfully as possible
Over the years I have built numerous dashboards and reports and when it comes to a BI application the most common and successful ones I have done follow the MAD framework.
The MAD framework is a 3 level approach defined as MONITOR – ANALYSIS – DETAIL that allows the end user to flow through data using a top down approach to view data at the 3 levels of detail as required.
This is really the ‘Dashboard’ part of the report. Often I hear a report or application being referred to as a dashboard, but in reality the dashboard should only be one or maybe two sheets of a larger applica. Think of the dashboard in your car, it’s the area where we can get the key pieces of information we require to monitor the consistent running of the car. A business dashboard is no different. It should convey, in the briefest of time, a quick snapshot of the required information to know at a high level the current status of the organisation.
If there is a change in the organisation performance such as the petrol is low, the engine is overheating or sales this year are down on last year, we use some form of indicator to point this out. It is this negative alert that encourages us to ask more questions and create an action to fix the performance. If it’s a positive alert, its reason to celebrate!
These sheets typically have more KPI’s and a few charts.
An example of a dashboard representing the Monitor part of the MAD framework
This next level gives us the ability to dig a little deeper to understand the issue, when we look at sales we can compare Sales Reps, Products, Branches, Customers etc to see how each of those are doing in comparison to last year and we can start identifying specific areas where there are issues.
Depending on the detail required each of these business areas could have their own sheet in a report to allow more analysis. At this stage it is good to have a tool that allows for selecting and filtering different dimensions (the describing data points like product name, branch name, salesperson name) so that you can drill in and ask more questions easily.
These sheets typically have more charts and maybe a couple of KPI’s, sometime a small table if it adds value
An Analysis sheet, in this case Customer Analysis
Once you have drilled in and pin pointed a possible issue that needs to dealt with -Your Actionable Insight, you may require some further detailed pieces of data. For example a Sales persons margin is much lower on average than others and you have identified that they seem to be offering more discount at POS on average than others. Once you have drilled down to this sales person, then identified the sales you might want to be able to view all the details of specific transactions.
These sheets typically are just a large table of transactional data that through the process of the Analysis level should have been filtered down to a quite specific set of data. You don’t want to get here and still have 1000’s of lines to trawl through.
A detail sheet showing the specific transaction details of a specific set of data base off applied filters
The MAD Framework Triangle
Normally the MAD framework is depicted in a triangle with the point at the top. From my perspective, when doing analysis I think it should be the other way around. Monitor gives us a high level view of all of the data. Analysis let us focus on a smaller data set that is more targeted. Detail is even more filtered and brings us to the pointy end of the process. Identifying the action required.