This is one of a series of posts on business intelligence and related analytic technology subjects, keying off the 2011/2012 version of the Gartner Magic Quadrant for Business Intelligence Platforms. The four posts in the series cover:
- Overview comments about the 2011/2012 Gartner Magic Quadrant for Business Intelligence Platforms, as well as a link to the actual document.
- (This post) Business intelligence industry trends — some of Gartner’s thoughts but mainly my own.
- Company-by-company comments based on the 2011/2012 Gartner Magic Quadrant for Business Intelligence Platforms.
- Third-party analytics, pulling together and expanding on some points I made in the first three posts.
Besides company-specific comments, the 2011/2012 Gartner Magic Quadrant for Business Intelligence (BI) Platforms offered observations on overall BI trends in a “Market Overview” section. I have mixed feelings about Gartner’s list. In particular:
- Not inconsistently with my comments on departmental analytics, Gartner sees actual BI business users as favoring ease of getting the job done, while IT departments are more concerned about full feature sets, integration, corporate standards, and license costs.
- However, Gartner says as a separate point that all kinds of users want to relieve some of the complexity of BI, and really of analytics in general. I agree, but don’t think Gartner did a great job in outlining how this complexity reduction could really work.
- Gartner is bullish on mobile business intelligence, but doesn’t really contradict my more skeptical take. Even as it confesses that mobile BI use cases are somewhat thin (my word, not Gartner’s, and no pun intended), it sees mobile BI rapidly becoming mainstream technology.
- Gartner makes a distinction between “data discovery” tools and “enterprise BI” platforms. By “data discovery” I think Gartner means what I’d call the “pattern discovery” focus of investigative analytics. Anyhow, it seems that Gartner:
- Sees users as being confused about how the traditional pattern-monitoring kinds of BI fit with the newer emphasis on investigative analytics, and …
- … shares that confusion itself.
- Gartner observes that “Most BI platforms are deployed as systems of performance measurement, not for decision support.” It evidently sees this as a bad tendency, which is thankfully changing. Automated decisioning is part of the fix Gartner sees, along with collaboration. While I agree on both counts, Gartner oddly doesn’t also connect this to the general rise of investigative analytics.
- Gartner also had a catch-all trend of “new use cases”, listing some examples, but also sort of confessing it wasn’t doing a great job of articulating the point. I think that part of the difficulty is contortions as to what is or isn’t BI; Gartner seems to run into expositional difficulties whenever it touches on the core point that analytics isn’t all about performance-monitoring BI. Another problem is that Gartner doesn’t seem to have really thought through what does and doesn’t work in the area of analytic applications.
Here’s the forest that I suspect Gartner is missing for the trees:
- Even though all-in-one enterprise BI platforms are great at getting data to a multitude of endpoints …
- … and even though the number of endpoints for data are increasing (more users, more devices) …
- … all-in-one enterprise BI platforms fall short in helping the data be used once it arrives …
- … and all-in-one enterprise BI platform vendors will find it hard to catch up with other vendors’ data-use capabilities.
It may seem odd to say that all-in-one enterprise BI platforms will not catch up to alternatives, given that:
- Vendors are big, and have lots of money, for development and acquisitions alike.
- Until a few years ago, the best BI user interfaces came from the biggest vendors.
But for starters, please consider that most lists of enterprise BI all-in-one platform leaders would start with Gartner’s top five “completeness of vision” choices: IBM, SAP, Microsoft, MicroStrategy, and Oracle. Only one of those five is actually an independent business intelligence vendor; the others are trying to integrate BI with other parts of the technology stack. Even the vendors no longer believe in the classical enterprise BI separate-platform ideal.
On the other hand, integrating BI with other stuff to be best of breed across the board is hard. For example:
- SAP had a variety of BI efforts — sort of including NetWeaver — that didn’t work out very well. Those were deemphasized in favor of the acquired Business Objects.
- This year’s Gartner survey respondents slammed SAP Business Objects for functionality, performance, and “customer experience”.
- Oracle’s own BI efforts were largely dropped in favor of acquired products via Siebel and Hyperion (which itself acquired that BI suite).
- This year’s Gartner survey respondents slammed Oracle BI for difficulty of implementation, and also rated it low for both product functionality and product quality.
- Gartner’s list of Microsoft “cautions” is extensive.
In fairness — IBM seems to have left Cognos’ products pretty much alone, except for some questionable “appliance” marketing bundles.
Now let’s look at three categories of “analytic application” technology opportunity I cited in a recent post:
- Operational applications enhanced with some analytics so as to improve routine business processes. This could become a strong argument for buying Oracle BI to go with Oracle applications and Business Objects BI to go with SAP apps. But note the problems I cited above even with lesser degrees of integration.
- Information services enhanced with some analytic technology that retrieves (and perhaps also helps analyze) the information. It’s hard to see why those would be well-integrated with an enterprise’s BI platform standard.
- Analytic-application-specific “platform” technology. This is pretty much the antithesis of relying on a general cross-industry platform product.
Let me be even more specific.
- I think large application vendors will eventually do a good job of integrating their pet BI products with their applications. However, I do not expect this to be a short or easy process, or one that blends well with efforts to make those same BI products be across-the-board competitive.
- I think there’s a huge opportunity for third-party BI vendors to play-nicely-with-others and integrate with all the other applications out there, packaged or SaaS (Software as a Service) as the case may be. Almost everybody in that arena could stand to improve their game.
- I think today’s BI vendors should be able to lower their latency to handle near-real-time use cases, provided that the data naturally goes into a relational or OLAP database (operational or analytic as the case may be).
- I see little evidence that today’s leading BI vendors have a clue how to handle non-tabular data. I’m not optimistic that their likely eventual move to buy that expertise will do them much good.
- Big BI vendors will have trouble competing in the cool-visualization market. In particular:
- Business Objects’ acquisition of Xcelsius hasn’t amounted to much, and the visualization side of its Inxight even less.
- IBM isn’t usually very good at user interfaces.
- MicroStrategy’s architecture could cause difficulties.
- Microsoft could be an exception, given its UI skills; Oracle could go either way.
- More generally, I think it is tough for a BI vendor to be great both at performance-monitoring BI and investigative analytics. I further think that both are extremely important.
Finally, as is always the case when I argue for best-of-breed over dreary-corporate-standard, I must concede that for sufficiently small enterprises, an all-in-one offering is probably best. The same may go for any single department of a large enterprise.