The Dream of Downstream
October 9, 2008
“When Will BI Head Downstream?” asks Shadan Malik. The question is one that challenges business intelligence vendors to expand their market. The story appeared in eCommerceTimes.com on June 27, 2008. You can read the full story here. (This is a story that has a pop up ad, so you could encounter a dead link or be unable to access this story.)
The write up reviews the consolidation of the business intelligence industry. You know the mantra by heart: IBM bought Cognos, SAP bought Business Objects, etc.
For me the key point of the essay is this statement:
The rapid expansion of these vendors has led them to focus on integrating platforms rather than creating new tools, which has resulted in complex and expensive solutions and stagnant innovation.
Business intelligence makes such good sense. The arguments advanced by vendors whet the appetite of executives for:
- Dashboards. The idea is that a busy manager can see the key data at a glance. The metaphor is the speedometer and oil light in your automobile. Who would not want to have a single heads up display just like a fighter pilot?
- Consolidated information. Most organizations have data scattered around just like a fraternity has junk piled in every room’s nooks and crannies. Finding anything is the equivalent of rummaging through a pile of sweaty gym clothes and mismatched socks. Who would not want an organized view of information?
- Real time alerts. Most organizations know when something bad happens by happenstance. A sales person takes a client to lunch and the client says, “You guys are not going to get the contract for our new data management system” or “You read about a competitor’s winning a contract in the Wall Street Journal”. Who would not want to know important events in real time?
The article takes a final jab at business intelligence vendors with this observation:
Organizations face the challenge of increasing loads of data, and traditional spreadsheet reports are not providing the quick and easy information access that is required for them to complete their job effectively. The majority of organizations have a variety of software applications that constitute their data infrastructure, and this data is often locked within the silos of these independent applications, making it difficult to report on combined data and have a holistic view of the business.
Annex Greener Pastures
Towns grow by annexing smaller communities. The idea is part of the American approach to getting big. In Kentucky, Louisville annexed Jefferson Country. The result was the Louisville was going to be as large a metropolitan area than Boston or some other big city. The arguments were that consolidation makes the city cheaper to operate because duplication of certain services would be eliminated. Well, Louisville is not Boston, and the annexation of adjacent communities has not paid off. Many people just moved farther away from the city, preferring the inefficiencies of life outside the reach of bureaucracy.
The notion of annexing markets is similar. Business intelligence vendors and their owners want to move downmarket. I anticipate that smaller business intelligence firms will be gobbled up. Prices for business intelligence services will be cut as vendors offer more business intelligence services from the cloud. Microsoft and other vendors will bundle business intelligence into larger packages.
However, the push is initiated, the idea is the same: move downmarket in order to generate more revenue.
Sprawl Brings New Problems
The problem for business intelligence is that business intelligence, like enterprise search, difficult to implement. The intuitive goodness of doing “business intelligence” glosses over the expensive, time consuming work that is required to make the systems work as the customers perceive them as working. The reality of business intelligence is very different from the jazzy, easy-to-use services that marketers plant in the minds of customers.
A short list of these issues includes:
- The data are a mess and it is expensive to clean up simple things like names in many different forms and formats
- Information is just plain wrong. Commercial databases cannot easily remove flawed data, and most businesses lack the resources specialists have. Bad data are endemic.
- Real time is a pipe dream. If you want to analyze streams of real time data, get ready to write big checks for specialized hardware and systems. There’s a good reason why the Exegy real time data engine costs six figures a box.
- Business intelligence requires that users understand what mathematical procedures do in order to generate results. A value such as 0.04567 does not mean much without some deeper understanding of what was analyzed and how and for what purpose.
Downmarket is Not Downstream
The metaphor of moving downmarket reminds me of sitting in a rubber tube and floating down the river near Moose Jaw, Maine. As long as I kept still, the current did the rest. I was clueless about the rapids around the bend, but progress at the outset seemed automatic. How nice, I thought, just before I entered the Class 5 rapids.
Business intelligence, just like enterprise search, may behave the same way. Business intelligence and search are becoming utilities, but they are when packaged for the downmarket cut down to an inner tube. The idea of seeing how much of a product sold in real time is okay when the product units are well defined, the reporting systems working well, and the data flows leisurely. When the real world of messy data and fast changes become evident, the vendors are in for a wild ride.
Observations
At this time, simplified business intelligence, just like the basic key word search systems, give the appearance of delivering “good enough” results. The challenge arises in these conditions:
- The flows of data exceed the system’s capabilities, allowing the user of the business intelligence system to experience the thrill of losing control and being carried along unaware of what the heck is coming next
- A crash because the customer cannot handle the business intelligence system no matter how simple its controls
- A greater loss of control than previously because the automatic nature of the system allows managers to relax, trusting the new system.
My hunch is that more business intelligence vendors will offer simplified versions of certain business intelligence functions. The adoption of these tools will over time accelerate. The market does not yet know what it does not know. Like search, only when the system does not perform to expectations will the problems surface. It has taken decades for the dissatisfaction with search to become widely known. Business intelligence may, in my view, follow the same trajectory.
Stephen Arnold, October 9, 2008
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
