Is self-service BI living up to the hype?
FYI, this story is more than a year old
Article by ZAP sales vice president Bill Tennant
Recently, Eric Buck made some thought-provoking points for users of Power BI, and those of self-service Business Intelligence tools in general.
In 'Self Service BI in an Enterprise World', Buck confirms that the explosion of data available to a business is transforming how everyone works and that the advent of self-service BI tools is “helping organisations leverage these large data stores and make their data work for them.”
But there’s an important caveat here - “One common theme though has been how restrictive and hard it has been for companies to provide both enterprise-ready reporting and accurate self-service BI options for their employees.”
“How do we control the sources of truth for our end users? What tools can we use that are simple for less tech-savvy individuals? How can we scale our BI solution to meet the needs of our industry?”
Self-service BI requires separate, self-service data management. Why?
Most mid-size companies may well be cognizant of the data value proposition, and they may have invested in BI tools to get the most from it. But what they don’t have is the budget to staff up a data science department to extract the maximum value - and the maximum ‘truth’ - from their structured data sources, for example, their ERP, CRM, HRM and financial systems - via BI.
To obtain this maximum value, all areas of data management need to be tackled on an ongoing, constant basis including
data governance, and
Traditionally, this has meant allocating budget and time to employ and coordinate a team of high-cost data engineers. Quite the opposite of self-service software.
The self-service BI world revolves around philosophies of ease-of-use and access to all. In a practical, day-to-day context this might mean multiple users with manual, drag-and-drop data modelling capabilities and access to Microsoft Dynamics and Power BI - a net result of dozens of conflicting data models, and dozens of users not knowing if they are looking at the right data from the right source.
The dangers of this are real and significant. As Buck warns, “BI is a very strong tool, but it can make you more confident in inaccuracies if there was an error in the data prep or the queries used to generate the report.”
Without even looking at the broader challenges of data source complexity and integration (from on-site ERPs to cloud-based CRM systems, for example), the bottom line is that most companies just can’t afford a data science department.
While Buck’s piece promotes very well the capabilities of Power BI it does not take into consideration that most mid-market companies don’t have a data scientist, let alone a team of them.