In a continued set of posts, I am examining the three dimensions of SAP over the last few years, and into the near term future. This one focuses on the introduction of the platform at SAP and the change it introduced to customers and software suppliers alike.
Undoubtedly the most dramatic set of decisions SAP has taken over the last 5 years revolved around the shift from an applications only company based on a closed proprietary platform to basing our solutions on an open standards-based platform (counter to ORCL's advertising). The decisions manifested in the creation of NetWeaver – supporting duality of ABAP and JAVA in a unified stack – and the decision to open SAP’s collection of processes to consumption through Web Services technologies as we have called them Enterprise Services.
While ABAP is a great domain specific language, designed and optimized for data centric transactional applications, it has never caught up as a general purpose programming language outside the domain of few aficionados. At the same time, SAP and the SAP customer base has a huge vested interest in making ABAP successful, to the tune of probably $100B in software that is implemented and running 7x24 at tens of thousands of customers. The need to support a new programming language did not emerge from the desire to improve ABAP – quite the opposite – it came from the desire to attract new talents trained in Java into the SAP camp and letting them leverage the greatest collection of enterprise processes by writing code in a language they are used to from school. The end result of the effort is a reviving of ABAP and basis as the most popular provisioning process platform, and NetWeaver's Java as the most integrated composition engine on top of those processes.
We had to decide at SAP two major decisions at the time: the first was whether we should build a platform or sign up with one of the vendors like IBM WebSphere or Microsoft .Net. After a serious evaluation it became very clear that if we choose one we have to choose all middleware platforms. As result of the fact that none of the vendors agreed on a real set of standards and they all diverged by more than 90%, we have come to the conclusion that we have to build our own platform. The harder derivative question around the technology platform became "should we keep this new platform for ourselves or sell it as an open middleware platform". Had we kept the platform to ourselves we would have only caused every customer to have NetWeaver as one more platform with high integration costs between technology stacks.
At first it seemed as though we will have no chance to penetrate the middleware market, but as numbers show we had done an amazing job. NetWeaver had launched in earnest in January 2004 – only three years ago. A year later, in Q1 2005, Merrill Lynch reported that we have reached the brave position of 1% (yes, ONE percent) of what the call "intent to buy" market share. Since then, NetWeaver created a significant adoption wave and ASUG reports more than 50% of SAP customers have picked NetWeaver as their strategic platform. Similarly Merrill Lynch reported in early January that NetWeaver reached 16% intent to buy market share in middleware platforms. That number is amazing especially when put in context - .Net had 17% market share, but every other Java platform (WebSphere and Fusion included) had less than 16% intent to buy share. That positions NetWeaver as the Java middleware market leader – and that is after 3 years.
The main driver in my view was the decision to combine processes with technology in the creation of a business process platform – manifested through the opening of processes through web services – the Enterprise SOA strategy we have embarked upon a few years back. As it happens, we started building composite applications under the xApps brand a few years back, and as we started building more and more of these composites we collected a list of processes residing in the suite that we exposed through an API. This is like extracting a dictionary out of a textual body of work (in our case the shakespeare and the bible of processes). Once you wrap those processes with web service interfaces you create an SAP dictionary and the big question was – do we keep THAT secret sauce to our selves or expose it to ISVs.
The question is the same as a country that enjoyed tariff protection around its borders forever and now decides to accept free trade over night. It is a scary proposition, and a tough decision, since you can open up and get a wave of innovation as an up side but also cannibalization of the account base as potential downside. The decision came down to not only open up but proactively evangelize the platform to ISVs through SDN and even help them get funded through the NetWeaver fund. The amazing wave of innovation that swelled around SAP was rewarded by more customers getting xApps, as well as SAP being able to acquire ready made solutions like GRC and xMII from Virsa and LightHammer.
The combination of technology openness and process openness created a new SAP proposition – one that is a platform for many companies to build on, inside a company or inside an industry. The adoption of an open ERP together with the current adoption of NetWeaver (500 installations going live every month!) makes that vision into a reality.