Jan 2 ’15
Enterprise Business Intelligence Is Coming Into Its Own
In the 2014 Magic Quadrant for Business Intelligence, the most recent, Gartner is drilling a familiar software term into the BI market space: “enterprise.” This new focus on enterprise derives from customer demand. Those customers, who in years past expressed preferences for data discovery, analytics and other business intelligence features, are now requesting more enterprise features in their business intelligence applications. This is a sign of widespread adoption and larger deployments, which is driving the need for better tools to manage the growing number of users and data sources—precisely what enterprise solutions are designed to do.
Enterprise features refer to those features that enable IT organizations to govern and administer software deployments, especially where large numbers of users (perhaps in the thousands) and data sources are required. For example, in business intelligence applications, shared repositories and centrally-managed security are considered enterprise features. According to Gartner, meeting enterprise IT demands, while at the same time providing tools powerful enough to satisfy business users, remains an elusive goal for most BI software vendors. Why has it taken so long for BI software vendors to build enterprise IT features into their products?
The first cause of tardiness in the expansion of enterprise features can be attributed to market forces. Based on market demand over the past decade, software companies have primarily focused on business intelligence features such as data transformation, data visualization and analytics. These features have been extensively advertised in glossy colors, high resolution and intelligent tag lines. Customers can’t help but notice the pretty pictures that are data visualizations created from these applications for advertising purposes, often with data sets of interest to the general population, such as sports and economic statistics. The result has been the creation of a new competitive environment in which vendors race to offer innovative features that outshine their competitors, and a rush to advertise those new features. Enterprise IT, which hasn’t been of high interest in the business intelligence software market until recently, has not received nearly as much attention over the same time period.
For many years, BI software customers have demanded more and better features. Consequently, we have seen major advances in business intelligence software, which has been transformed into applications that not only render data as compelling data visualizations, but also create an open-ended user interface, allowing business users to explore their data in the manner of their choosing, share their insights with other members of their organizations and create their own data mashups. Search capabilities have also been greatly enhanced, allowing users to quickly find disparate data points via generic search fields. In many BI software packages, the dashboard and report authoring experience have also been transformed into a relatively simple and intuitive task. Customization of how the fields will appear, such as a chart or a table, or which chart type, is a task that has become easy enough for business users with no background in data tools to master. These are just a few examples of how business intelligence applications have improved dramatically in recent years. But what does this have to do with enterprise IT?
As some vendors have improved their business intelligence applications, other vendors have followed suit. In a race to lead the market based on business intelligence features, enterprise functionality has taken a back seat. In summary, business intelligence features are sexy and make enterprise features seem boring in comparison. However, this imbalance should be temporary.
Once business intelligence features matured to their current state, a fairly recent development, the market turned a corner and a new emphasis on enterprise IT started to take center stage. In coming years, we can expect enterprise IT features to be weighed heavily by customers and to be a core aspect of software companies’ development strategies.
Enterprise IT and Business Intelligence: A Rocky Start but Quickly Improving
Another cause of late blooming of enterprise features within business intelligence applications is the nature of enterprise IT as compared to the nature of business intelligence. Given that business intelligence applications can be very complex, integrating enterprise functionality across all aspects of existing business intelligence applications has been a significant challenge.
In my work as a business intelligence consultant during the past 10 years, I have yet to see an ideal example of out-of-the-box “Enterprise Business Intelligence.” Progress has been made but features that are targeted at enterprise IT users often seem to be short of the mark. For example, as part of their “metadata management” strategy, many BI tools offer the functionality of a shared semantic layer and/or a repository of fields. The idea behind this is that a few users, perhaps with specialized knowledge of business processes, create a common set of fields to be leveraged by other users of the system. Critically, some of these fields contain calculations and alternate (user-friendly) names, creating an environment where data is easily located and calculations in different dashboards and reports are consistent. This is all excellent enterprise functionality, but the problem is that often these common fields only work in specific parts (or dialogues) of the system, and not others.
Certain types of reports or dashboards can be built completely from repository fields, but others cannot. This can be a major headache for a user who is attempting to build a report but finds that fields must be re-created and business logic re-defined for some of the fields in order to work with certain types of dashboards. This gap in functionality, and others like it, are often due to the complicated aspects of business intelligence applications. Besides the fact that business intelligence applications are complicated, they typically have extensive custom code and changing definitions. On the other hand, enterprise IT applications (or their interfaces) are relatively fixed over time, are not typically customized at the programming level and have relatively simple interfaces (since the whole point is an easily-administered system). The assumption I’m making is that it’s hard to integrate the simple, enterprise-focused application parts into a complex, feature-rich, business intelligence application that changes frequently and is comprised of many custom definitions and extensive custom code.
In conclusion, customers, who in the past were developing custom enterprise solutions for their business intelligence applications, should start to expect out-of-the-box enterprise functionality. This should be a good thing for everyone.