December 21, 2013 | By Rey Villar
This is part 3 of a blog series centered on the topic of creating BI power users.
Potential BI power users need the right data access method at the right time. That is, at the right time in their learning cycle. Since we remain agnostic to who uses our tools, we need to accommodate the learning (and working) styles of all: from starter data surfers to veteran data divers.
Simpler BI tools hide the underlying complexity of the data through visual point-and-click interfaces. Think dashboards and scorecards. A button push runs a data extract. Users do little more than navigate and read screens. This is the venue of the ‘executive’, ‘casual’, or ‘operational’ data user. The questions about what data should be presented have already been thought through by the application team. Not a lot of mental engagement required here, but it’s a stepping stone.
Drill-down is a natural extension of the dashboard. Selection of data values reveals a glimpse of the logical constructs hiding in the black box, but no database or code knowledge is required.
As the ‘visual’ tools increase in sophistication, they start to look more like logical file systems or data models. Users evaluate the business question to be answered and traverse a business-friendly model to gather the required data. User inputs allow data selection. Data extraction is no longer a purely point-and-click exercise. The mental transition – and visual translation to coding – has begun.
Many popular BI tools allow users to view the SQL generated by the tool. The ‘peek’ function is critical as it links the visual and logical paradigms. Reading the SQL code exposes the code to the end user. Users can toggle back and forth between the graphical and coded representations of their query. Over time, users gain insight into the business rules embedded into the code.
Your users go from crawling to walking when they can extract the tool-generated SQL code and make it their own. Code can be copied and pasted into a tool like TOAD or Teradata SQL Assistant. Users can view, edit, execute and tune their own SQL. Over time, the users learn how to answer business questions directly using SQL.*
More sophisticated tools allow users to go even further – to create multi-pass SQL, temporary tables, read/write data to/from any data source; virtually, to create their own applications. The ultimate BI power tool enables creation of complex procedures to extract data from many types of sources, manipulate and aggregate the data, and write it back in one of many formats. Powerful tools like SPSS and SAS can be used to explore complex business problems. This is the playground of the digital elite.
So what does the typical learning path of a fledgling BI power user look like under this program?
- Initiation to the BI world with a dashboard or scorecard.
- Data selection using point-and-click visual tools.
- Training on a tool where the data model (or similar) is the starting point to the creation of single-pass SQL.
- Beginner SQL training
- Access granted to use ‘show SQL’ window to capture code.
- SQL copied and executed in a database access tool like TOAD.
- Intermediate SQL training.
- Advanced query tool training.
- Advanced SQL training.
- Training on an advanced tool like SPSS or SAS.
Power users are created, not born. Graduating from simple clicks to complex SQL code can take several months to several years, but not all will successfully make the journey. As BI practitioners, we need to put in place the ecosystem to support the incremental growth of visual and code-based thinkers:
- Make available the tools that each type of thinker will gravitate toward naturally.
- Invite them to training.
- Add a support network to nudge them forward when they get stuck.
- Leave them to learn from each other and from the ‘data evangelists’ that will inevitably begin to emerge from the group.
BI power users, at last!
My next blog will explore the collaboration that sets the stage to create a community of power users.
* Access to this capability is a critical step on the path to becoming a power user. Unfortunately, many organizations stop their deployment of BI right here. Users with access to change SQL render a degree of unpredictability and added cost to the BI platform. Each organization needs to decide whether the rewards justify the risks and costs associated with unfettered access to the data. I would argue all day long that without the right to create, edit and test their own SQL, your end users will never make the transition to true BI power users.