SAP (NYSE:SAP) is both gaining and losing market share depending on what markets you are talking about and how you measure them. But customer count is a better indicator anyways. SAP released its 20-F this week using an exchange rate of US$1.32 to the Euro to express 2006 results in dollars.
SAP’s methodology of using the US$:Euro exchange rate at the end of the calendar year is consistent with past year SAP SEC filings. But, as I noted in my January 16 post, be careful in comparing these dollar-denominated results with those previous filings. Because the Euro rose 10-11% vs. the dollar during 2006, the dollar-denominated results in the documents indicate a 23% SAP total revenue increase for 2006 over 2005 (and other possibly misleading information in terms of income and EPS). The dollar-denominated results overstate SAP’s market successes in 2006, although SAP was successful in my opinion.
This is not an accounting or reporting issue but an IT market investment research issue. Playing it straight in its 2006 20-F, SAP only shows an explicit year-over-year revenue, earnings and EPS comparison in Euros. You have to go to two different documents to see the 23% increase that I mention above. You will also likely see growth numbers like mine in leading IT research firm share tables later this spring. Fair’s fair, of course; a few years ago when the Euro went the other way vs. the dollar, these same currency-conversion methodologies in share tables artificially depressed SAP’s growth and share.
In Euros, SAP total revenue is presented as growing at 10 (11% at constant currency). But this is not the best way to view the financial results either. SAP warns against relying solely on the lower Euro-denominated numbers because they are based on U.S. generally accepted accounting principles. I also believe that all currency-change methodologies skew market-research results for any company that gets a significant amount of revenue from different geographic areas.
Why should we care? Because we make investment decisions based on market share. Who’s hot and who’s not—and why or why not—are constant questions? But the answers are so dependent which markets are being measured that I question the premise. For example, SAP plays broadly speaking in two software markets, applications and middleware:
• In what SAP calls core enterprise applications, SAP has its own convoluted measurement for market share analysis (see Note 3 of this SAP press release). I don’t like it because—among other reasons— SAP is not transparent about who the other application suppliers are in its 30-company index and where the 4% forecast comes from against which SAP compares itself. There are over a thousand companies in the packaged applications market and as a group they account for much more than the $16 billion in “software revenue” that SAP cites. In the 20-F, SAP cites IDC data that says the applications market grew over 7%, not 4%. (“Software revenue” has a specific meaning in the SAP lexicon.)
• In addition because SAP is now moving rapidly up the middleware share table charts (55% growth denominated in Euros; perhaps over a billion dollars in 2006 if you use the US$1.32:Euro rate), its applications revenue growth was less than 10% denominated in Euros (just doing the arithmetic with the total number, the NetWeaver number, and the remainder from the 20-F)
Although I argue below that I don’t think it matters, here’s what I think of the 2006 applications market in terms of share for those that do care:
• SAP’s 2005-2006 applications revenue growth is around 10% after triangulating against NetWeaver’s reported growth and currency conversion adjustments
• Oracle (NASDAQ:ORCL) grew its applications business around 9% in the 12 months ending November 30, 2006 (but the ERP part of that was as high as 15%)
• Infor, a now privately held amalgamation of approximately 20 different formerly independent ERP and standalone application suppliers, was probably flat.
• Microsoft (NASDAQ:MSFT) grew its business applications business about 15%-20% in CY 2006, most likely breaking through the billion-dollar barrier
• Sage grew 22% during the year ended September 30, 2006 but it appears as though at least half of that growth was based on acquisitions.
• Intuit (NASDAQ:INTU) grew its QuickBooks and Professional Tax lines single-digit
• Lawson (NASDAQ:LWSN) had no growth in 2006 on a proforma basis that factor in its merger with Intentia.
So, as described in my January 17 comment to my Jan 16 post, it’s likely SAP held or grew application share in CY 2006. This back of the envelope analysis does not include industry-specific “ERP” players including Cerner, FiServ, GE Healthcare, Meditech, McKesson, SMS/Siemens, and others that I include in my full review. Nor does it include the non-engineering standalone applications providers that should be included to do a truly apples to apples comparison. Furthermore, this is not the same as “ERP” market share, in which SAP likely did lose a point or two in 2006. If anyone cares about the difference, add a comment to this post and I will reply.
How about a quick look at the middleware market (using Oracle’s definition):
• IBM (NYSE:IBM) still leads dramatically with somewhere around $10 billion in annual software license, maintenance and like revenue, but with much of it is tied in to IBM systems and services deal of course
• Well behind with approximately $1.5 billion in “middleware” revenue would be Microsoft, not estimating a value for the middleware built into Windows operating software and for which Microsoft does not charge separately
• Although BEA (BEAS) has still not filed detailed results because of stock-option accounting problems, if its maintenance revenue for the 12 months ending January 31, 2007 was the same percentage of overall services revenue as it was for the 12 months ending January 31, 2006, than BEA’s software license, maintenance and like revenue was a little above $1.2 billion.
• Oracle came close to that number in the 12 months ending November 30, 2006 as described in my December 21 post
• SAP NetWeaver is close on Oracle’s tail and growing at least as quickly
So SAP had a good year from a market share perspective: it held its own in a market it dominates (applications) and is catching up rapidly in a market it considers its growth area (middleware).
But is there a better way to measure SAP’s (and others’) progress than the slicing and dicing of share tables that can be manipulated by market and by currency. In January 2007, SAP put out new statistics that in the long run will be much more revealing than converting euros to dollars, or counting widgets, or debating the definition of middleware. They are customer count and customer-count target by type. SAP highlighted two strategic metrics:
• Its progress against a previously stated goal of 100,000 customers by 2010. This is a metric that gauges SAP ability to enter the small and midsize market arena (which for SAP means midsize market as opposed to truly small businesses)
• Its progress in getting R/3 users to transition to SOA.
Against these targets, SAP was proud to announce 6000 new customers in 2006, increasing the total to 38,000
• Almost 13,000 on Business One (up from 9000)
• Almost 9,500 on All in One (up from 7700)
• 15,500 on classic SAP applications.
(The fact that there was no growth in classic SAP’s installed base is not surprising given SAP’s already impressive penetration of large enterprises.)
SAP also pointed out that they expect 66% of customers to convert to Enterprise SOA by 2010. This is a metric that gauges large-enterprise loyalty and ability to add to the SAP footprint within an enterprise (more than just ERP for example).
From an IT Investment Research perspective, it will be tough to meet the 100,000 goal, unless 45-50% of customers are midmarket by 2010, because of the competition from Intuit, Microsoft, and Sage. But it’s a worthy goal that could be reached by peeling off Infor, Lawson, and former J.D. Edwards/Oracle customers. In the long run it will be a good way of measuring all applications suppliers. Because of the consolidation in the applications market and the move down market by the larger suppliers, there will be far fewer brand overlaps. These have been the argument against customer census number measurements in application market research in the past.