Pure-play BI vendors strike back

Operational BI is just one of many ways to compete against the acquisition-hungry big guys

It is only fair then that I let the other side have its say. As you might guess, pure-play BI vendors don’t subscribe to my bleak view of their future. My first comeuppance came from Sanju Bansal, COO of MicroStrategy, a 17-year-old BI pure-play.

Keying off IBM’s acquisition of Cognos, I proclaimed in my Nov. 12 post that BI will be subsumed into the database, citing the fact that Oracle can now offer Hyperion technology, Microsoft has announced reporting and analytical tools for SQL Server, and IBM can now add Cognos analytics to its DB2 database.

Bansal parried: “It would be hard to justify spending $3 billion, $5 billion, and $8 billion for Hyperion, Business Objects, and Cognos, as did Oracle, SAP, and IBM, respectively, and give them away for free,” Bansal told me in no uncertain terms.

Rather, Bansal says the acquirers are buying access to customers in order to cross-sell, upsell, and tap distribution opportunities, meaning that SAP’s acquisition, for example, allows it to leverage Business Objects’ strong presence in the midmarket.

Bansal adds, as many have before him, that customers have a plateful of heterogeneous systems that require a BI company willing to optimize for a wide array of databases and platforms.

A point well-taken and true. You have to believe that when SAP buys Business Objects, it is going to focus on integration with NetWeaver first and foremost. Finally, to make amends with Bansal, I’ll allow him this free plug for MicroStrategy.

As price points and architectural decisions change, pure-plays allow for flexibility, Bansal says, noting that if you want to switch from, say, SQL Server to Oracle to Teradata over time, you have complete freedom to do so if you go with a company like MicroStrategy.

The counterargument to that is familiar: It is better to deal with one vendor and a single platform. Even if its BI solution is not perfect, it will get there.

However, consider this. What if IBM, Oracle, and SAP spent a total of $15 billion on three aging dinosaurs?

That is what Charles Nicholls, CEO of SeeWhy Software, and Måns Hultman, former CEO and now chairman of QlikTech, believe to be the case.

Nicholls talks about the new focus on operational BI, as opposed to strategic or managerial BI. Despite promises from the acquirers that they too will offer “operational” BI, Nicholls believes the companies they acquired are in fact too steeped in the old way of doing things: collecting data in a data warehouse and writing reports after the fact.

From a business perspective, Nicholls claims that somewhere between 40 and 50 percent of a BI vendor’s budget is allocated to connectivity to middleware, applications, and the database or to integrate with all the various platforms and systems — both off the shelf and customized, as companies often tweak what they buy.

At the same time, the target audience of these companies is small. Traditional strategic BI is targeted at the top of the pyramid, about 5 percent of the total number of PCs at a given company. So how can they remain profitable selling into such a small market as their overhead increases?

On the other hand, Nicholls claims a company such as SeeWay, which targets 45 percent of all the PCs at a company, has a much healthier business model.

Operational BI is suited to a service-oriented world with loosely coupled services strapped together with an ESB (enterprise service bus). Today, valuable information is going through that bus, not through the applications. As a result, operational BI makes a great deal of sense, as it can be used by many more employees.

“We analyze data continuously, as opposed to storing data and running a query,” Nicholls says.

The beauty to me is that now BI imposes itself, in a good way, into business processes by changing the processes as the analysis warrants.

Take, for example, QlikTech. This company offers an architecture it calls “associative in memory” technology that is far different than a typical OLAP report generator. It searches in a nonhierarchical way, which means it is not constrained by searching in a predefined path where an IT engineer has to decide for you which way you want to analyze or attack the data. Hultman says the only way to realize an associative approach is to place the analysis in memory, so a user can change sorting order, or the way data is organized on the fly.

The real point is that perhaps I was overly hasty. Perhaps point-solution BI vendors will not only survive but flourish and that the recent acquisitions by the giants only serve to reinvigorate rather than replace the pure-plays.

Recent Reality Checks on business intelligence :

• IBM Cognos deal highlights a resurgence in upgrading the database

• Reading between the lines of the Oracle bid for BEA

• SAP’s Business Objects acquistion: The death knell for point solutions?