Multiple business intelligence (BI) platforms in an enterprise are here to stay. Respondents to an informal social media survey that I’ve been running for the past couple of years report that 25% of organizations use 10 or more BI platforms, 61% of organizations use four or more, and 86% of organizations use two or more. (Anecdotal evidence based on multiple interactions with Forrester clients support these findings as well.) No matter how hard enterprise IT organizations tried to rationalize and consolidate BI platforms, they were only going to be partially successful. While IT pros were tasked with platform consolidation efforts (laudable, but a Sisyphean endeavor) to get closer to the so-called “single version of the truth” and to achieve efficiencies and cost savings, business users fell in love with and didn’t want to easily give up their favorite BI toys (my apologies, a Freudian slip — I meant BI tools).

We called for the need of a “BI fabric” in our initial research in 2017 and later in our 2018 update of this report: Use BI Fabric To Optimize Your Multivendor Business Intelligence Environment, where we defined a BI fabric as:

Technologies and techniques that allow business insights pros to integrate, leverage, and reuse components from multiple business intelligence platforms.

Since then, this BI fabric baby has grown up — not too much, but slowly, taking baby steps. We expected a much more rapid growth in this market. Enterprise IT’s innate resistance to multiple redundant platforms and BI vendors’ lack of enthusiasm in investing in integration with direct competitors are the two main factors slowing down BI fabric adoption. But in the end, business users win, and we currently see BI fabric manifesting in one or more of the following architectures and technologies:

  1. A common BI portal — a single place to search, collaborate, create workflows and mash-ups, catalog, and secure objects (metrics, reports, dashboards, data visualizations, etc.) across different enterprise BI platforms. Vendors include Accton, BI Hub, Digital Hive, Knowi, Metric Insights, ZenOptics, and a few productized solutions from consulting organizations like Symphony by BRG Global Applied Technology.
  2. A common natural language query (NLQ) UI. A single NLQ UI connected to multiple BI platforms. Currently limited to AskData (recently acquired by SAP, and it remains to be seen whether SAP will keep integration with products that compete with its own SAP Analytics Cloud).
  3. A common data catalog — a single place to catalog all data sources used by multiple enterprise BI platforms. This is also a single place (a common BI portal can also be used for this function) for data source governance — tagging data sources with the levels of data quality and approved use cases, promoting data sets from development to production, etc. Most of the platforms from The Forrester Wave™: Machine Learning Data Catalogs, Q2 2022, plus productized offerings from global system integrators like LTI’s Mosaic Catalog can be used for this purpose.
  4. Data virtualization or a cubing engine as a common semantic layer. Data virtualization platforms like TIBCO Data Virtualization or Denodo, and cubing engines like AtScale, Kyvos Insights, or open source Cube Dev and a few others can be used as a common semantic layer for multiple BI platforms.
  5. BI platform as a semantic layer — a BI platform allowing other BI platforms to use its semantic layer as their own. The pickings here are slim, currently limited to players like Microsoft Power BI (supports platforms that use XMLA for Analysis for data connection), MicroStrategy (3rd party BI tools can plug into MicroStrategy virtual cubes), Google Looker (Tableau can plug into LookerML semantic layer), and GoodData (via query-by-API). Oracle Analytics Cloud and Oracle Analytics Server also support third-party connections to its semantic layer via Java Database Connectivity, but Oracle does not test nor specifically approves any competitor integration.
  6. Platform-to-platform connectors — capability to display a visualization/dashboard/report from a different BI platform. Current capabilities include Google Looker connectors to Qlik, Sisense, and Tableau and Yellowfin connectors to Microsoft SQL Server Reporting Services reports and TIBCO Jaspersoft.
  7. A common NLG platform. Most leading BI platforms come packaged with a native natural language generation (NLG) capability. But in majority of cases this capability is based on ML – automated, but leaving little room for customization, and mostly limited to a very short (we call this “short form” NLG) narrative/explanation of a data set. Hence, as we wrote in a recent report enterprises are increasingly looking for more comprehensive, longer (“long form” NLG), customizable NLG solutions. This is the realm of Arria NLG, Automated Insights, AX Semantics, Narrative Science, and Yseop. Organizations can continue to use multiple BI platforms but standardize long form NLG on a single tool.
  8. Metadata integration — import/export of BI platform metadata using an exchange standard. A couple of decades ago, seemed like it was evolving as a BI metadata exchange standard. But for now, only IBM, Microsoft, Oracle, SAP, and SAS support the standard. And sadly, these are mostly good for platform-to-platform conversions and initial deployments, as we are not aware of clients synchronizing multiple BI platform metadata in real time.
  9. Headless BI – nirvana state where every single BI platform component (data connectors, data pipelines, semantic layer, metrics, visualizations, reports, dashboards) are completely de-coupled, ivoked by APIs, and therefore agnostic (at least theoretically) to where, which BI platform each component comes from. GoodData and Google Looker come closest to supporting that architecture.

Like it or not, multiple enterprise BI platforms and the need for BI fabric are here to stay. We encourage enterprise buyers to urge their BI platform providers to embrace this new reality of a BI fabric. Want to research and talk more about this? Take a look at our BI fabric 2023 research update and schedule an inquiry with me.

Note: Forrester client access is required for research featured in this post.