Enterprise Technology Blog

Of the many, many unsolicited emails I received last week only one had a subject line compelling enough for me to take a chance on opening it: “SAP and Oracle: Who’s Ready for Small and Medium-Sized Businesses?”

Being that SoftBrands is a Gold Channel Partner of SAP which focuses on the SME space - as well as SME sites of SAP’s large enterprise customers - i was curious to read the research and see what conclusions were drawn.

The study was conducted by Nucleus Research and, much to my surprise, the findings were not very friendly to SAP. The final line in the conclusion states: [T]he data today shows Oracle has a clear lead over SAP in the SMB marketplace. 

Ouch!

Now, I’m not an expert on methodology, but one of the most important factors for achieving accurate results is sample size. Simply put, if your sample size is too small in proportion to the entire population (say 5 of 5000), the data cannot be considered reliable. 

Another factor is clearly defining the profile of those surveyed. Failing to put the research into the proper context will lead to questions about the conclusions drawn.

On both counts the Nucleus Research report falls woefully short.

Let’s start with sample size. Nucleus Research did in-depth interviews with 29 SAP customers and 27 Oracle customers. I don’t know how many customers Oracle has in the SME/SMB space, but SAP has thousands.  With sample sized this small the data is so volatile I can’t fathom how any reliable conclusions can be drawn on just 27 and 29 interviews. 

The second point goes to defining the population surveyed. In this case, the report says only that Oracle and SAP SMB customers were contacted. But it fails to define SMB.

What criteria were used to determine if a company is part of the SMB market-space; revenue, employee count, single site enterprise, a combination of these and others?  There is certainly no consensus on what characteristics constitute a SMB enterprise so a definition is certainly in order.

Another glaring omission is what the surveyed companies were running.  For the SAP customers, were they on All In One, Business One, R3.  What about Oracle with JD Edwards, PeopleSoft, Siebel. Two companies each with multiple products.

A small and poorly defined sample raises several questions about the conclusions drawn by Nucleus Research and dimishes the credibility of the report. 

In this case, two strikes and you’re out.