CAN DATA MINING CROSS THE CHASM?
by John K. Thompson
Can Data Mining cross the chasm? A provocative, yet timely question which deserves a thoughtful answer, or at least an informed opinion. Let me attempt to provide one or the other, or maybe both depending upon your opinion. In Geoffrey Moore's books, Crossing the Chasm and Inside the Tornado, Mr. Moore has laid out the Technology Adoption Lifecycle (TALC). These books and the TALC discussed within them have become the high-tech bible for planning and executing not only marketing strategy, but also the overall business strategy for emerging markets and companies striving to attain leadership in those nascent markets. To attempt to have a conversation about high technology startup firms and their products without referencing the TALC or related jargon is nearly impossible.
I have read in various publications that the data mining market is "embryonic in nature", or "ready to cross the chasm", or "in danger of losing chasm crossing momentum". For those not well versed in chasm speak, embryonic in nature is indicative of a market that is far from growing fast enough to cross the chasm. Ready to cross the chasm indicates that the firms classified as innovators, and early adopters, have installed the products in sufficient numbers to sustain a significant number of vendors and the early majority firms are beginning to consider using the technology. The last comment is self-explanatory, but assumes that the market was or is moving at a sustained rate that would indicate high levels of interest and purchase activity. The first two comments are at opposite ends of the pre-chasm spectrum. How can that be?
For such a young market the data mining space is incredibly complex. Consider the number and types of firms that are competing for leadership in this market. First, there are firms that have been through Chapter 11 or similar types of bankruptcy protection, which allowed them to be recast into software firms from the ruins of their proprietary hardware businesses. Then the majority of the major players that have been successful in the Business Intelligence marketplace jumped headlong into the market 18 to 12 months ago.
Along the way the major statistics based software vendors decided to throw their hats into the ring. And let's not forget all of the government funded R&D that comprises the technology bases of a variety of niche vendors as well as some of the major players. Then last but not least are the true garage shops that are cranking out algorithms by the boatload, who for the most part are looking to license their wares to anyone who has money to do so.
So, in a simple back of the envelope kind of segmentation exercise, we end up with at least 5 categories of vendors. When you take into consideration that each of these five segments approaches the market from a slightly different angle, and the product and service offerings are broadly divergent from vendor to vendor, you end up with a cacophony of messages which are blasted at the market creating one big noisy confusing space.
We have a young market, which really has some quite old roots, but we'll call it young for the sake of discussion. It has a large number of players, who are offering a wide variety of products and services that range from near business solutions to pure technology in raw form. And, then we blend in the analyst community that weighs in with opinions that range from "Data mining is DOA" to "The data mining market will be an $8.0 billion operation in the next..." Hyperbole and confusion reign, but only if you look at the market from that 10,000 foot level. When you begin to look more closely at the market, you begin to see which currents are moving swiftly away from what seems to be nothing more than a chaotic whirlpool.
Many people have said that there are approximately 50 to 60 data-mining vendors actively building and marketing software and services. I think that there might be closer to 100. Each week I see new firms popping up, but most of them I rarely see again. How do I see this shaking out? Well, here is my informed opinion.
First of all, the larger and most serious players are moving into vertical markets around industries like telecommunications, insurance, and financial services. Most of these companies will continue to enhance their core technologies and extend their applications through internal development and acquisitions. When there are discussions around crossing the chasm, these are the firms that are the closest to the jumping off point.
Secondly, pure data mining tool vendors who either are trying to move to a vertical orientation, or are actually executing on a well defined plan to achieve vertical applications or solutions need to be introducing those products in the next quarter or two. Time is running out on these firms. These firms, unlike the firms in the previous category, do not have the luxury of existing revenue streams from products or services to sustain them while they move to a solutions orientation. I am skeptical as to whether many of these firms will exist at the end of 1999. One or two will survive and may actually have a chance to capture a vertical and then the opportunity to contend for a leadership position in the broader marketplace.
The Business Intelligence firms are mainly focused on query, reporting, and On-Line Analytical Processing (OLAP) offerings. These firms jumped into the data mining market so as to not be left behind in the rush to grab some of the money to made from data mining. Many of the personnel that came into these firms either through acquisitions or outside recruiting have left those firms to head back to companies that understand "serious analysis" as one of my contemporaries put it. These companies will maintain the lightweight products that they have purchased and superficially integrated with their core offerings. Nobody in this space is vying for the leadership position in any serious manner.
The traditional statistics vendors may gain the most for the least amount of investment. These firms, for the most part, are not concerned with crossing the chasm. They see the data mining bandwagon as a way to expand the existing markets that they play in, which are primarily governmental and academic. But more importantly, the increased visibility and market coverage focused on data mining has provided an easier entry into the commercial world that they have had a difficult time selling into. A couple of the firms are thinking about making a run for the abyss, but most are happy to take additional market share where they can.
And finally, there are the government sponsored firms that have not weaned themselves off the government teat, and the garage players. The play for them is to license their core technology to the firms in the other categories. These people either do not understand the commercial market at any level or they are just too small and too focused on the merits of the raw technology to be considered anything other than a supporting act.
So, there you have it. The race boils down to the big players taking the vertical markets one by one, or there is a possibility of a few small vendors, who are intelligent, focused, and who execute with a ruthlessness of a winner, who will win a vertical or two and live another day to fight for the brass ring.
And, Oh by the way, if you didn't think it conspicuous by its absence:
there is no horizontal, or general market for data mining tools.
---
John Thompson, Vice President -- Marketing, Magnify, Inc.
I'd like to hear
your thoughts. You can reach me at jkt@magnify.com