A remarkable number of vendors are involved in what might be called “specialized business intelligence”. Some don’t want to call it that, because they think that “BI” is old and passé’, and what they do is new and better. Still, if we define BI technology as, more or less:
- Querying data and doing simple calculations on it, and …
- … displaying it in a nice interface …
- … which also provides good capabilities for navigation,
then BI is indeed a big part of what they’re doing.
Why would vendors want to specialize their BI technology? The main reason would be to suit it for situations in which even the best general-purpose BI options aren’t good enough. The obvious scenarios are those in which the mismatch is one or both of:
- Kinds of data.
- Kinds of questions asked about the data.
For example, in no particular order:
- Splunk is focused on machine-generated data, which can arrive in a steady stream (e.g. into Splunk’s own data store) or perhaps in batches (e.g. into Hadoop). Splunk’s BI tools are focused on questions like “What’s actually going on in our machines?”
- ClearStory is focused on third-party data, and has tried to innovate strongly in platform technology and user interface alike.
- Platfora first attacked the problem of providing in-memory-style access to Hadoop volumes of data. Recently it also tackled the UI and data structure challenges of entity-centric event series analytics.
- Metamarkets is focused on online advertising data (even more so than when I previously wrote about the company), and the velocity/volume tradeoffs that involved.
- PivotLink, originally a general-purpose SaaS BI/data store provider, decided a while back to focus solely on the retail industry, and get as far into analytic application functionality for that market as it sensibly could.
- StreamBase tried to offer a general-purpose real-time BI product. But LiveView’s initial UI was specific to the financial trading industry, and the product never prospered outside the one vertical market in which StreamBase had traction.
And that’s just a list of companies in my current or recent client base. There surely are other examples with which I’m less closely acquainted. (E.g. Glassbeam.)
And by the way, Michael Bloomberg became one of the richer people in the world offering a combination of specialized BI, manual reformatting, private networking and similar embellishments to a collection of essentially public sets of data.
In another possibility, since BI is traditionally used on a fairly stand-alone basis, special needs can arise when BI is to be tightly integrated into concrete applications. I’ve written more about Workday’s efforts in that regard than about larger vendors’.
Tying these observations into some of my previous posts:
- Building a data technology platform into full commercial maturity and generality is a long and expensive task. But building one copy you operate yourself can let you come to market sooner. Not coincidentally, some of the specialized BI startups are SaaS-only (Software as a Service), notably those who are most aggressive about their platform-technology innovation.
- I’ve recently argued that the platform side of BI technology is crucial to innovation. Most of the vendors I’m thinking of seem, based on their development priorities, to feel the same way.
- I’ve further argued that navigation is a more important part of the UI than visualization. The specialty BI companies don’t necessarily agree with that. On the other hand, they do have pretty good navigation capabilities.
And finally, this all seems consistent with my claim that even if you want to sell on the basis of analytic application benefits, you’d best have a strong proprietary technology story.