Monday, November 21, 2005

A bit rich

You may have seen SAP's latest advertising campaign, which breathlessly claims that "A recent study of companies listed on NASDAQ and NYSE found that companies that run SAP are 32% more profitable than those that don't*. Based on a 2005 Stratascope Inc. Analysis". There are at least three interesting features in this advert. The first is that little asterisk at the end. If you read the (small font) footnote you will see that this excludes all financial services companies. A little odd until you realize that financial services companies these days have two particular characteristics: they make a lot of money, and they rarely use SAP. Could their inclusion have, perhaps changed the results somewhat? Of course one could ask Stratascope, the market research firm headed by a chairman and President Juergen Kuebler, a nine-year SAP veteran, and I'm sure they will give an unbiased and objective opinion. I am going to take a wild guess and say that including financial services companies would not make the figure look better.

However by far the most interesting aspect of this advert is its sheer chutzpah, with its implication that if you use SAP then you will be more profitable: 32% more in fact. Lest the subtleties of statistics have escaped the denizens of Walldorf, I would like to remind them that because two datasets have a positive correlation, it does not mean that this correlation is caused by anything. For example, I observe that my increasing age is well correlated with the steady rise in global temperatures. As far as I know, there is no direct link. Similarly, one could observe: "the stork population has gone up, as has the human population. Hence storks must create human babies". For example, I can tell you that four of the UK's five most admired companies last year were Kalido customers (true) yet to say that one implies the other is absurd.

It is particularly implausible to make such bold claims in the case of IT systems of any kind, which may well have distinct and real benefits in individual cases and projects but whose influence on overall productivity has generally eluded economists entirely. The studies there have been are controversial e.g. the 2002 McKinsey study that showed that, other than a few sectors (retail, securities, telco, wholesale, semiconductors IT) there had been no productivity growth whatever between 1995 and 2000 in the US despite heavy IT investment. This study was looking at all IT investment, of which ERP is only a small fraction.

So overall, SAP's claim excludes a key industry sector to selectively improve its results and in any case makes a claim that is logically spurious and has no supporting evidence. Other than that, excellent. All par for the course in software industry marketing.


Post a Comment

<< Home