Business intelligence tomorrow… and what it means for today
Maximizing the cloud
Many comprehensive BI platforms function equally well, whether on-premise or in the cloud. Nonetheless, they heighten hybrid and multi-cloud deployments in several critical ways involving:
♦ Scalability: It’s difficult (and much less cost-effective) to match the cloud’s ability to scale with conventional on-premise options. The abundance of cloud data stores are perfect for data warehousing “because so much of the systems that support businesses today are becoming SaaS supported, [so] data is starting in the cloud and living its entire life there,” Potter commented.
♦ Transformation: Transformation is still a vital aspect of integrating data for applications such as BI. ETL measures will likely continue to persist, particularly when “automating ETL via automated scripts,” Potter said. When it’s done in the cloud, however, ELT—which minimizes the steps for loading data while leveraging the processing resources of the underlying data repository—makes that data “almost instantly available,” said Frank Bien, CEO of Looker.
♦ Reduced cycle times: Most cloud BI platforms have a host of connectors for the popular cloud resources impacting business users. Organizations can readily use them to connect to data sources, make data expediently available via ELT and schema on-demand options, and then swiftly get analytics to business users. “Older BI tools required 6-12 month planning processes before people could even start using them,” Schaefer said. “Cloud technologies take that lengthy cycle out of the equation.”
BI on unstructured data
Another exciting aspect of the transition from self-service BI tools to self-service BI platforms with uniform semantic, metadata, and classification capabilities is the incorporation of unstructured data for business use. Typical BI tools worked almost exclusively on structured data, and even today, the most sophisticated BI platforms offer only limited utility for semi-structured data. Nonetheless, the wide assortment of social, mobile, and alternative unstructured data sources that considerably enhance the analytics experience require “an environment where the meaning is accessible,” stated Robert Coyne, CMO of TopQuadrant. “This means you have to have models of the entities and the semantic types, properties, and interrelationships.” Self-service platforms accessible for BI (and almost any other use case) fortified by standardized approaches to data models, taxonomies, and relationships extend these benefits to alternative unstructured data, as well as to structured and semi-structured data. The singular alignment of all these types of data is necessary to deliver “the right information to the user looking to use self-service, to get exactly the data they need,” Polikoff said.
Embedded analytics at the edge
Other compelling developments in BI heralding its future relate to the expansion of access. According to Anand, increasing usage of mobile technologies is influencing comprehensive BI platforms predicated on “touch-optimized experiences that are very mobile-specific and also allow offline and transaction-based actions as well.” Somewhat parallel to the trend of mobile technologies with BI is its inclusion as part of edge computing, which eminently reflects the tendency toward embedding analytics. Truly comprehensive BI platforms are characterized by analytics engines that can perform the computations users need for analytics.
Progressive vendors in this space are able to embed this “analytics engine into any device for IoT use cases,” Potter said. The revolutionary nature of this aspect of Internet of Things analytics—which are widely predicated on filtering the results of analytics at the edge before sending them to a centralized location—is readily apparent. “Rather than collecting vast amounts of low-value data, we’re using our analytics engine to capture meaningful analytic events and push those back instead, so that you’re actually doing central analysis on much higher value information,” Potter commented.
Data strategy
The advancements BI is producing today and tomorrow are perhaps most profound in terms of their effect on data strategy. Now that BI solutions encom-pass holistic platforms for sustainable, governed, self-service access to data at fundamental levels (metadata, semantics, classifications, and data models), this discipline is transitioning from merely servicing business users to servicing the enterprise as a whole. Such platforms can facilitate the same benefits of reliability, expedience, and self-service to a range of other use cases. As such, it’s a pivotal implement for “enabling a customer’s data strategy to be integrated with their analytic strategy,” Potter remarked. This concept is relatively simple, yet unambiguously brilliant. Organizations can use the centralized mechanisms supporting these holistic BI platforms in any way they see fit. Developers, for instance, can leverage them to create the most meaningful applications, as well as to analyze their effects on business processes. Edge deployments, the incorporation of unstructured data alongside structured data, and the foundational capabilities to solve underlying business problems with data (the essence of data strategy) are all possible with such contemporary BI solutions.
The new mission
Due to its expansion from a random assortment of tools to centralized platforms delivering rapid data access to a breadth of sources in a governed, self-service manner, the very mission of BI is changing. These capabilities are enabling BI to widen its focus from business applications to enterprise applications, regardless of the use case. BI is “really about creating something that you can build on top of and go further to start to break the chains and solve the problems of [data] access and reliability,” Bien asserted. “As you solve those base needs, sort of like the base needs of hunger and shelter, then you can start to move up the value chain more and provide even more to people in your organizations.”
Moreover, the merit of these self-service platforms is unequivocally horizontal. Polikoff described a life sciences use case in which there are many vocabularies that companies are using—such as ICD9—and others. “They’re taking this information, their clinical trials or whatever, and they’re linking it to those common standard terms. This is how they know they can integrate it; this is how they know the meaning.”
In this respect, the mission of BI is only constrained by the goal of the organization deploying it, since holistic BI solutions fortify users to employ the tool they love on the platform they trust, Anand said. If users want to leverage other tools and interfaces, there are APIs for doing that, but they can still use the semantic layer to govern data for centralized or decentralized use, he added.