SAS Bashes Those Who Overuse the Word Analytics
November 10, 0003
A happy quack to the reader who sent me a link to “Understanding the Eight Levels of Analytics. SAS Executive Says There’s More to Analytics than Meets the Eye.” SAS—a company steeped in statistical and numerical recipes—has a grip on the business analytics market. I even have a colleague who thinks about solving problems by coding the solution the SAS way, learned at Cornell University where she obtained her doctorate. I’m easy. We use SAS methods. The SAS executive is commenting about non-math outfits appropriation of the word “analytics” without defining it nor backing up the use of “analytics” with substantive systems and methods. The article included this passage which I found suggestive:
Everybody has analytics,” he [SAS senior vice president] said. But what they are actually offering in terms of analytic capacity in support of solving business problems remains in question. “So you bring all the data together and you put it in some form in which the end user can gain access to it, but what are you doing with it?” he asked.
The guts of the article is a summary of the eight types of analytics. What resonated with me is the penchant of revenue chasers to appropriate terms that add sizzle, cachet, and magnetism to less exciting types of software. I won’t name any search vendors who use this method, but if you visit a trade show, you will see that quite a few vendors of information retrieval systems are busy explaining that their systems perform analytics. The demos often show a search box retrieving a complex report. Google Squared and WolframAlpha offer similar services with an increasing number of training wheels tacked on to each, but neither company is using those systems to add excitement to their underlying technology. The Google and Wolfram services provide a glimpse of an alternate approach to obtaining information from a word or two key word query.
The SAS approach is to chop up analytics into eight segments. Most of these are quite simple; for example, reports and a click to reveal underlying data in a spreadsheet. Others are more complex and are not going to be easy to wrap into a key word search system. Google is working on this problem if I understand some of the lingo in that firm’s technical papers and patent applications.
The point is that companies may buy something that sounds good but won’t deliver data that is accurate or ultimately appropriate to carry the decision load place upon the outputs. With math skills in most organizations below average, more attention to numerical recipes is needed. Talking about math is not quite the same as understanding what’s going on and what the results can or cannot support.
Stephen Arnold, November 10, 2009
Comments
- Subscribe to Beyond Search
Feature archive
News archive
-
