Eric Apps, president of Angoss Software, here develops on his belief that we are only begining to scratch the surface of what web mining can mean to customer interaction and online sales.
Both pure-play ‘dot-commers’ and traditional businesses establishing a web presence are keenly aware of the need to build online brands and strengthen online communities. In five short years they have moved from being sized up based on their “hits” to “eyeballs” to “page views” to “clickthroughs” to “conversions”.
What does data mining have to do with this?
In the same way Wal-Mart’s much publicized “beer and diapers” analysis affected merchandising strategies and layout planning in retailers, Amazon.com’s “next book” selling suggestions have come to typify the business value data mining brings to an on-line experience. And from a data mining perspective, both examples merely scratch the surface of the possibilities.
While the metrics used to measure e-business success continue to evolve, it is already clear that data mining technology can help e-businesses better reach, understand and grow their online communities, regardless of how they define and measure their own success. And success is a very site-specific concept (as long as you don’t have venture capitalists breathing down your neck to get the business moving!).
To see this growing interest, you need only look at whatis happening. Major technology vendors like IBM, Microsoft and Oracle are moving into this area, offering data mining infrastructure support in their e-business platforms. E-businesses, as well as the application vendors and alternative service providers serving them, can leverage these capabilities.
Applying data mining in the web space is a logical extension of the core technology now well rooted in many large traditional bricks-and-mortar businesses in North America and Europe. A wide range of public domain and home-brewed data mining algorithms have been used for over 20 years to improve decision-making.
At their core, these technologies have been based on data exploration, pattern recognition, knowledge discovery and the generation of predictive models.
For non-practitioners, data mining as it relates to human beings is really nothing more complex then understanding a few key concepts. The vast majority of people are creatures of habit – that can be discerned. We are also increasingly “digitized” and known to our banks, telecom carriers and retailers only by our “electronic footprints”.
At its core, data mining revolves around these precepts and – more broadly – the notion that what has happened in the past is generally the best indicator of what is likely to happen in the future.
In the same way that traditional data mining tools complement traditional data analysis tools (such as query and reporting and OLAP solutions), web-mining tools like KnowledgewebMiner from Angoss Software complement basic analytics solutions for e-businesses.
Web log parsing and reporting solutions such as Net*Analysis, webTrends, NetTracker and others, for example, help businesses identify which audiences they are reaching and how they are reaching them.
Outbound e-mail and campaign management solutions, such as those offered by Kana, eGain, Delano and others, facilitate the management of this communication. At a broader level CRM and cCRM solutions, offered by hundreds of technology vendors in a broad range of geographic, industry and functional areas, all seek to capture and manage these electronic footprints.
Web mining solutions add incremental functionality, helping e-businesses derive even deeper levels of understanding of activity at their websites to guide their strategic and tactical planning as well as their day-to-day operational activity.
* Angoss Software Corporation, based in Toronto, Canada, develops advanced data mining and data visualization software solutions for use “at the intersection of the warehouse and the web”.