by Paddy Padmanabhan

Gartner BI & Analytics Magic Quadrant : What it’s Right About, What it’s Not

Mar 02, 20154 mins
AnalyticsBusiness Intelligence

The report gets it dead right on certain aspects of BI and analytics today. However, the reports are also constrained by a conformity to a certain structure and equating analytics with BI, a notion that’s past its shelf life.

The Gartner Business Intelligence (BI) and Analytics Magic Quadrant for 2015 is out. This is the annual anointment of vendors looking for validation from an authoritative “independent” entity, and my inbox is slammed with every one of the “leaders” hammering into my head that they’ve won top honors. Congratulations to all. And thank you for providing me with a free copy of the report.

The report gets it dead right on certain aspects of BI and analytics today. For one, it’s right about the importance of data discovery, and how a differentiated data discovery capability really puts one vendor ahead of another. With more data sources to deal with, internally and externally, the effectiveness of an analytics program depends crucially on the ability to integrate these disparate sources – structured and unstructured, internal and external – in a seamless and real-time fashion to gain analytical insights.

The report is right about the power shift (some may say power struggle) from IT to business when it comes to analytics. Business users have low tolerance these days for what they see as an asset-heavy, slow and inefficient IT organization, and they want to take control.

The report is also right about the increasing acceptance of cloud-based models. It rightly acknowledges that traditional BI vendors are cloud-enabling their platforms to address their needs for improved data discovery and visualization, and also to give business stakeholders a chance to be free of internal IT organizations. The authors also admit that traditional BI vendors are playing catch up against newer players like Tableau, the current gold standard for automated data discovery and visualization.

The report misses a few key aspects, though:

Many people I know would object to the inclusion of analytics and BI in the same sentence. BI is essentially about looking back, and analytics today is mostly about looking ahead. Examples: Amazon doesn’t try to tell you what you bought last week or the last year, they try to tell you what you might want to buy today or tomorrow. They’re trying to predict and prescribe what you’re going to do. Hospital systems today are less concerned with what their patients came into the hospital for last year. They are more interested in why they might be coming into the hospital in the next six months, and they want to intervene ahead of a potential admission or a re-admission.

Secondly, the report does not adequately address the increasing role of data sciences and the emerging Internet of Things (IoT). The report makes a passing mention of several startups that are exploring radically new ways of automated data discovery and pattern detection. To be fair, many of these startups are yet to go fully mainstream, and the IoT chatter may be more hype than reality at this time (as noted in my recent blog)  but one wonders if a Magic Quadrant Report on Analytics should perhaps be doing a little bit of prediction on the likely winners for the coming year.  

Finally, there isn’t enough discussion in the report on data security. This is an analytics vendor stack-up, and it’s all about data. In light of the recent Anthem data breach, IT departments and vendors alike have data security as their No. 1 concern, especially in the healthcare, consumer banking and retailing sectors which deal with sensitive personal information. The evaluation criteria for ‘ability to execute” tellingly misses this critical aspect.

In general, the Gartner Magic Quadrant reports are well researched. However, the reports are also constrained by a conformity to a certain structure that tends to continues to equate analytics with traditional BI, a notion that is well past its shelf life. To that extent, it has more historical than predictive value.