CONVERGENCE? OK, BUT WHAT ARE WE CONVERGING ON?
by John K. Thompson
This week I am thinking about the much-hyped word, convergence. It is one of those terms that means something different to each individual, especially people in the high-tech field. Convergence has been widely discussed for the past few years in relation to the merging of traditional consumer goods and networked appliances, and regarding the melding of television, voice, and data networks. Now there are discussions in the Business Performance Management (BPM) marketplace about convergence. My first question back to anyone that I encounter talking about this is, "What are you talking about?" Not to be rude or confrontational, but unless the speaker frames the discussion fully, there is absolutely no chance that speaker and listener will be on the same page.
In thinking about convergence in the BPM marketplace, I have been considering three distinct product areas: On-Line Analytical Processing (OLAP), Data Mining, and Knowledge Management. The convergence of these products may happen in a physical sense sometime in the future, but for now, the convergence that I am talking about is conceptual. Convergence of these three technologies, and others not considered here, will need to be considered and planned for by vendors and then carefully evaluated by firms intending to implement such an all-encompassing BPM architecture.
Of the three product segments mentioned, only OLAP is broadly adopted across a wide range of industries and application areas. Data Mining is experiencing growing pains as it slowly penetrates specific industries and applications area, and at the same time grows the market segment with the introduction of text and video mining technologies. The nascent Knowledge Management market is experiencing similar growing pains as it is being molded by the emerging startups with newly minted products and the older players that are migrating from the Document and Image Management market. The market, being the ultimate arbiter, will decide the fate of the later two technologies. For the sake of this article let's assume that both market segments survive and grow and join OLAP in the BPM market.
With the BPM market containing the three technologies (with probably three firms in each segment), firms have a framework containing sufficient choices for them to evaluate and build a decision support backbone that supports their need to investigate, evaluate, determine, and monitor courses of action. Traditionally decision support systems built from technologies like OLAP and data mining have relied almost solely on quantitative or numerical data. Standard reports and analysis of tables, or if the systems are advanced, the numbers are hidden behind a static graphic, or a visualization that allows viewing from all sides, but in the end it's still columns and rows.
With the introduction of Knowledge Management, firms can now start to build infrastructures that collect, classify, catalog, and manage soft or qualitative data and information. One example of an application that will grow with this technology is Marketing Automation. Group Brand Managers or Division General Managers may want to know what the ad spending has been on all promotions that are targeted at a specific age group by gender and geography. Certainly the budgeted numbers and expenditures that have been loaded into the data warehouse can be pulled via a reporting or OLAP tool, but what about the unplanned expenses? Consider that the artwork for the direct mail, as well as the 30 second spot for radio, both had additional costs that were not budgeted, were approved on the fly, and not entered into the legacy systems and therefore never made it into the data warehouse. What about those expenses?
With a system designed to collect, and manage the paper flows or qualitative sources these types of activities and deviations from plan can be captured and analyzed. This scenario sounds plausible, but will all companies be able to implement such a system to capture even a portion of the paper flows? Of course not. Systems to capture these types of data and information, at first, will only be implemented by the largest of corporations who have the staff, the processes, and the need to drive down costs through thorough process optimization and automation. As these systems are purchased and implemented, vendors will grow the functionality and begin to move more packaged software into the market were smaller firms can utilize the more mature technology and applications.
This being the second go-round for the newly named and positioned Knowledge Management, not all of the players in this space will label themselves as such and with good reason. The vendors that have a better chance of success will be those that position themselves as offering an application or solution to an existing and known problem. Generic or horizontally focused technology waiting for a client to build a solution will not work. The horizontal or generic approach is falling by the wayside for all but the most simplistic of technologies. In the enterprise software marketplace, of which BPM is a segment, vendors must offer at least a partially crafted approach to a problem. Clients do not have the time to learn the new technology, implement the software, and build a custom solution. Clients want complete applications or custom applications that can be finalized quickly and implemented to solve the problem.
How do these three product segments converge? Data mining, in all its forms, analyzing numbers, video, and text, will be used to discover new patterns and relationships as well as validate some of the more novel hypothesis. As I have written about in previous columns, data mining will need to mature into accessing traditional storage systems such as RDBMS and other repositories, and when it does there will be an increase in the adoption rate, but I believe that data mining will take its rightful place in serving as the discovery and validation mechanism in the beginning of the BPM loop. The traditional analysis of quantitative data will continue to be performed by reporting tools, and OLAP systems aided by graphics and visualization technologies.
The change in mind set that we will begin to see is the use of throw away data marts where sensible and practical and the beginning of the use of saving decisions and processes that have yielded positive and in some cases spectacular returns for further analysis and aggregation with other decisions. Ultimately this practice will yield libraries of decisions that can be examined for behavioral trends and common elements. Knowledge Management will work in tandem with OLAP and reporting systems supporting the existing analyses by collecting qualitative data that has traditionally slipped through the cracks. As previously noted, the adoption of Knowledge Management systems will be most rapid where vendors offer clients a clear and defined benefit where the clients recognize an existing need.
Much of this is speculation, but there is a need to build a BPM framework. Organizations want to make better decisions. Managers want to be able to clearly understand why one program worked and why another did not. The troika of technologies discussed may or may not fit the bill. That is up to the ingenuity of the vendor community.
We know clients want it, but how will we best deliver it?
---
John Thompson, Vice President - Marketing, Magnify, Inc. I'd like to hear your thoughts. You can reach me at jkt@magnify.com