There is no doubt that Microsoft’s Power BI started a revolution by making a large scale, functional BI system available to all. Many are adopting from the cell oriented spreadsheet technology of the past to broad and fluid Tabular Intelligence and are investing time and effort in becoming masters of Power BI’s functional languages of M and DAX.
However, although mastery of Power BI’s functional languages is necessary, also required is knowledge of what the data story is. What’s a data story? Simplistically, it’s the ability to present data that will answer the questions the audience will ask.
As an example, below are two Power BI Sales Reports which draw from a common fictional data source. Both report the same Year to Date figures and variance by country and customer.
Nonetheless, Which do you think is presenting facts and which is presenting Insight? Moreover, what’s the difference? The answer is, although it's possible to present data that answers questions by the presentation of facts, the desired outcome is to offer data that stimulates a course of action, in a direction, at a sequence with a consequence and this is what insight is.
Figure 1 shows the sales result is good, it’s ahead of last year but, there are weaknesses in certain areas that can be addressed, such as;
France.
Some Categories in Countries eg. Condiments in the USA. And;
Some Customers are behind Last Years sales (not shown).
However, is there a sense of urgency or concern? I’d suggest not. Plenty of green (good) here.
Figure 2 indicates similar (to a certain degree) but, the narrative changes with the inclusion of measure that splits Sales by the value derived from existing, new and lost distribution. The insight is not from the value of Lost Distribution which is high, it is that 80% of the Sales Value is dependent upon New Distribution. Similarly, there is much that can be inferred from the customer makeup. Although returning customers make up 80% of the Sales Value, 20% comes from new customers. This organisation is dependent on new customers and replacement of lost distribution. Nonetheless, this is fictional data but, if it were not, it should be ringing alarm bells at the Executive level.
Figure 2.
For a BI Developer, determining what data is relevant is a matter of asking the right questions of those that will present and those that will consume the report. However, the issue with this is sometimes, the audience doesn’t know what it doesn’t know. In this case, the brief to a BI Developer is to create a sales report but, how would the Developer know which metrics to include when the audience isn’t aware such measures are possible.
The answer lies in the preparatory stages of developing a BI report. Not only is the time spent discussing report requirements necessary but, the discussion also requires to be directed to understand;
What does the current report lack (if it exists)? and, What questions remain unanswered?
What is the job function of the presenter of the report?
What job functions consume the report? And,
What strategic initiatives is the job function responsible for?
I may be overly cynical, but, I believe too much discussion focusses on which visual to use to create the data story. For me, this is secondary, a single figure in a card visual can convey more meaning than a multitude of graphs, maps and tables can, as long as it is the relevant figure.
Besides, there is an excellent reason why a BI Developer should spend much more time in the preparatory phase of BI Development. Although each iteration of a report further refines the requirements for the production version, trial and error is a costly and time-consuming method of development
Market Grunt specialises in BI Reporting for Retailers and Suppliers in the FMCG and Hardware industries in Australia. With over 70 years of Sales, Accounts, Marketing and Business management experience for small, medium and large companies, Market Grunt's unique combination of business experience will deliver the insights your business is looking for.