One of the most exciting tales in modern analytics is the continued maturation of business-led analytics solutions, or Self-Service BI. Tools like Microsoft’s Power BI continue to make leaps and bounds forward, putting powerful and easy-to-use analytics tools into the hands of those who rely on insights from their organization’s data every day, all while helping to straddle traditional obstacles in the self-service space such as solution scalability, collaborative workspaces, and centralized oversight. But what about governance? When, where, and how does the traditional IT or IS department step in?

Self Service BI Data.jpg

While a tool like Power BI may help ease some aspects of governance, such as offering usage statistics and the ability to manage published content available to consumers from a single portal, other aspects can remain elusive: when should a data source become an IT-managed asset? When should a self-service data model become an IT-hosted enterprise model? Should they ever? To what extent should we favor “one version of the truth” over the “democratization of data”? Consider, if you will, the English Garden.

The English Garden

In the early 18th century, a new approach to garden-building was sweeping across the green spaces of Europe, replacing the formally rigid and symmetrical styles of the older French methodologies: the English Garden[1]. Seeking to embrace the natural contours and intrinsic “wildness” of the land, this new approach was a reimagination of the original landscape as an idealized pastoral form of nature, rather than a strict and symmetrical recreation of something apart from it[1]. Often including bodies of water, groves of indigenous trees, and traditional floral plantings, these new gardens were an offering of the best that both nature and traditional formal gardening had to offer.

To set up the analogy then, let’s think of an organization with no IT management of analytics artifacts (data sources, data models, or reports) as being like a natural landscape. While abundant, it is also untamed. Not just anything can grow there, because there is no higher governance than nature’s own rules – survival of the fittest.

Conversely, consider an organization in which IT policy governs and locks down all data sources, models, and reports. We can think of this as being like the old French-style symmetrical gardens. While there is an appeal (and a few advantages) to the rigid perfection, there is also an “unnatural” quality to it, as it has effectively replaced the land it grows on with an inflexible vision of what should grow there – survival by design.

A Traditional Give-and-Take

For a long time, IT’s control of analytics assets not only made perfect sense – allowing for focused and accountable application of data quality and cleanliness standards, security requirements, and captured business logic – it was also nearly essential. The tools required to clean, combine, transform, and centrally store data needed large amounts of computing power and very specific skill sets in order to be effective.

So too with building and distributing reports, where knowledge of esoteric data-centric languages like SQL and MDX (multi-dimensional expression language) were necessary prerequisites to report design, not to mention the inherent complexity of the report-building tools themselves.

These advantages and necessities came to present a metaphorical chasm within organizations everywhere, between those with the resources and skillsets to curate the data on one side, and those who need to make use of that data every day in order to do their jobs, on the other. As organizations began to create and retain more data than ever, this chasm grew wider and deeper. Greater volume, variety, and velocity of data, along with the attendant use-cases for their efficacy translated to significant demands for even the best staffed and equipped IT department. This then became our traditional garden: highly symmetrical, well-tended, and quite restrictive; often becoming quite independent of the surrounding landscape.

It was into this gap that Self-Service BI sought a niche, eroding some of the traditional necessities that once made centralized IT control of data and analytics a foregone conclusion. Power BI in particular represents one of the most thorough realizations of this vision available, bringing an intuitive yet robust platform for report design and data acquisition, and the vast capabilities of cloud-based computing, to make manageable what were once formidably large and complex data sets.

This “democratization of data” meant that those who most needed to make use of it had the means to potentially do so, without the risk of losing requirements in translating them to an IT department, or denial of service due to a lack of resources. Yet while some of the hard requirements to making use of data diminish with a self-service tool, what of the benefits that IT-governed data confer? Indeed, without some manner of overarching vision, the “democratization of data” can quickly begin to resemble a wild and overgrown landscape of competing and overlapping initiatives. Important standards around data quality and security can become nigh impossible to apply uniformly, let alone audit. In understanding the strengths and weaknesses of each platform, we can begin to tame the wilderness and build a better garden.

Visions of Harmony

Just as an English garden seeks to incorporate the natural contours and features of the surrounding landscape, so should a good analytics practice. The traditional give and take of enterprise, or IT-managed, BI and self-service BI need not be at odds, but rather can work together to form a more natural and complete practice.

From the perspective of IT, introducing self-service BI to the organization can serve as a path-finder to data sources, use cases, and business logic not yet considered or curated within the enterprise. Seeing captured data directly and the logic employed is often far superior to slogging through hours and hours of abstract interviews, all while risking that something could get lost in translation. Additionally, self-service BI can help drive adoption of existing enterprise data assets by shining a light on the benefits conferred by such. There’s nothing like experiencing first-hand the challenges and rewards to data acquisition and modeling for opening the door to an appreciation of those assets at enterprise scale.

From an organizational perspective, self-service BI helps introduce a common language by which analytics needs can be clarified and met. Without some degree of governance, the importance of having “one version of the truth” – or a common set of data and logic across all relevant areas of the organization – becomes immediately and obviously important. In this way, the introduction of governance from IT becomes a welcome means to taming the wilderness, rather than an imposition of abstract laws and requirements that can often seem completely out of sync with the surrounding environment. Additionally, IT skillsets that once represented an inescapable barrier to entry for report development can now present a route toward the useful enrichment of a business-led analytics initiative, rather than act as gatekeepers to it.

If you are looking for more information or need help finding the right solution for your BI and governance needs, contact us!

[1] English landscape garden. (2017, October 29). In Wikipedia, The Free Encyclopedia. Retrieved 15:47, December 14, 2017, from