Parallel Evolution: ECM and BI

With all the M&A activity of late in the ECM space, I thought it would be a good time do some comparison between the ECM and the BI markets.

Recall that prior to joining Mark Logic (the content world), I spent nearly a decade in business intelligence, or BI (the data world). Since few people have spent a lot of time on both sides of the data/content divide, I think I can bring an interesting perspective.

Both the BI and ECM spaces grew up as applications/tools on top of relational databases (RDBMSs). ECM has its roots in document management. BI has its roots in ad hoc query and reporting. Both started as cottage industries with many specialized vendors serving many subcategories of what has converged today to become a broader, more general market.

For example, in BI, Cognos specialized in online analytic processing (OLAP), Business Objects specialized in ad hoc query and reporting (Q&R), and Crystal Reports specialized in enterprise reporting (ER). In ECM, Documentum specialized in document management, Interwoven specialized in web content management, and FileNet specialized in imaging.

As time passed and as shake-outs occurred in the subcategories, the subcategory leaders naturally decided to move into adjacent subcategories. Often this starts out as product development , some of which succeeds and much of which fails. Sometimes it starts out as third-party licensing. Usually it ends with an M&A frenzy because there are actually two things going on.

First, each vendor is watching the progress of the others in getting the “complete suite” product line. In BI, the complete suite is ETL, Q&R, OLAP, ER, planning, budgeting, data profiling/quality, and analytic applications. In ECM, the complete suite is document management, web content management, imaging, records management, workflow, and digital asset management. (See this Gilbane Report dated 2/04 for more.)

Second, each vendor is watching the relative size of the other vendors in this process. No one wants to be left behind in the race for scale, or “critical mass.”

The problem with the enterprise suite-itization strategy is that it usually results in worst-of-breed product lines. In reality, the leaders in each of the N subcategories buy the losers in the other N-1. Think: Business Objects buying Acta or SRC, Cognos buying DecisionStream, or Hyperion buying Brio long past its sell-by date.

Sure, there are sometimes exceptions where leaders buy leaders (e.g., EMC/Documentum buying Captiva or Business Objects buying Crystal). But the typical outcome of this phase of market evolution is:

• Highly uneven product lines with respect to product strength/quality
• Little or no product integration across the suite (“we’re fully integrated, at a price-list level”)

This leads most industry analysts and pundits to dismiss the suite-itization. This happened a little bit in BI; it happened a lot in ECM. In fact, at the very first ECM conference I attended in 2004, I heard that “enterprise content management isn’t” or “ECM is a myth” about 10 times before the opening day’s lunch.

Before moving to the present day, let me rant briefly about suite-itization strategies. While the idea of building suites generally strikes me as good, I dislike the way most vendors implement suite-itization. Why?

• Because the strategy is seemingly irresistible. Once the sirens sing, no one can resist. Once one vendor starts building a suite, everybody follows. It’s virtually impossible to think of a company who resisted suite-itization, once started in their category. It would be nice if a few contrarians tried to hammer out victory on best-of-breed basis, if only to see what happens.

• Because the strategy ignores subcategory-level innovation. I think most vendors immediately switch all their R&D effort to product-line integration, completely overlooking the opportunity to innovate in the subcategories. So, innovation basically stops in the market.

Until recently, BI and ECM evolution were striking similar. The same things happened to different vendors, in different markets, serving different customers. But the evolution of the markets and the drivers of that evolution were viritually identical.

Until now. Two things have happened in ECM that haven’t happened yet in BI.

• Cannabalism. Contenders in the suite-itization battle buying runners-up (think: OpenText buying Hummingbird). Early M&A activity is about product-line completion and size. When a contender buys a runner-up they are presumably getting about 80-100% product overlap, so it’s almost exclusively about scale.

• Major vendor incursion (think: IBM buying Filenet). The big boys (e.g., Oracle, Microsoft, IBM) in effect become suite-of-suites vendors. This is already true across many categories today. Thus far, the big guys have either dabbled in or avoided both ECM and BI. Take Oracle, for example: their BI tools (e.g,. Discoverer) have been perenially weak. It appears only through accidentally picking up nQuire (as part of Siebel) will Oracle finally have good BI technology. Similarly, Oracle’s “tsunami” launch appears to have had anything but a big-wave effect on the ECM market.

So ECM seems to be heading in a new direction. If things evolve similarly in BI, you can expect to see cannabalism (such as Business Objects buying Hyperion) and incursion (such as IBM buying Cognos) in the future. People have speculated about such things for years. If BI continues to evolve in parallel fashion to ECM, then perhaps soon the speculators will be proven right.

For more on ECM market consolidation, check out this post from Tony Byrne.

Note To Readers On My Hiatus
Sorry! When you’re a CEO-blogger, you’re precisely that; not a blogger-CEO. So from time to time, there may lags in posting. My goal remains to post 1-2x/week. Hopefully, I’ll do better at meeting it in coming weeks.