Having attended SAP TechEd Berlin keynote via the virtual link I was expecting greater things from Vishal Sikka, SAP CTO at the Las Vegas reprise this week. SAP managed to disappoint by adding nothing to what we saw in the earlier keynote presentations. Boring? No. Enlightening…not really. Steady? Yes. But all of those descriptions cannot mask the seismic changes SAP is undergoing or the path it is trying to persuade its two million plus developer community to follow.
The keynote themes were familiar: cloud, mobility and in-memory. But it is the detail that we learn what this means for customers.
Vishal Sikka is rapidly growing in stature as someone with whom analysts can do serious business. Apart from his tech credentials and the blessing of SAP co-founder Hasso Platter, Sikka is both charming and refreshingly honest in ways that put predecessors and many PR managed vendor CXOs to shame. I was fortunate to spend three sessions with him; at times trading barbs but at others digging into the serious business of how SAP will sidestep the increasingly long shadow that Oracle is casting over the enterprise landscape.
In this post I will concentrate on high speed compute, what I term situational applications, cloud reality and customer protection.
HANA - high speed compute and decision management
Last week I referenced the veiled barbs thrown in Oracle’s direction. This week Dr Sikka was more direct. In a presentation deck covering HANA, its in-memory compute theme, SAP talks about 10x the performance of Exadata at 10% of TCO. In hard dollars, SAP talks about having built a trial system for $532,000. It will therefore pitch this as a $5.0 to $5.5 million cost saver. However, the delivered solution cost is more likely to run $14-15 million. If that seems breath taking then it is important to understand what SAP is really pitching and the near term opportunity.
Today, SAP is concentrating on HANA as the cornerstone upon which to build an initial $100 million pipeline for value add, high speed, decision support analytics among its top customers. One example they showed was around responding to a complex set of events to determine optimum retail stocking. An example might be ice cream sales predictions in specific geographies and climates.
I argued this is not a new idea. Enron invented a whole series of predictive analytics solutions in the late 1990s. Wal-Mart has been using the ‘beer and nappies’ idea for many years. SAP responds that the fundamental difference from what went before is in speed, relative simplicity, cost reduction and richness of the data which can be used to make better planning decisions or expose trends. Dr Sikka wraps this up as a value based proposition.
Teradata, Oracle and others will not stand by but SAP believes it has a good shot at this because it can come at the potential for revenue generation as a solutions provider and not simply as a box shifter that has to manage integration expectations. I question how that will work in the real world because as we know, data extraction is not a trivial task. We can only wait for SAP to share its fortunes on future financial analyst calls.
The broader question comes in where HANA goes from here. Dr Sikka was cautious not to oversell the notion of HANA as a future OLTP appliance though that is clearly the direction the company wishes to go. If successful then ‘New DB’ which is bundled into HANA becomes the competitive threat against which Oracle will push back.
Situational applications
TechEd kicked off with Innovation Weekend. This is an event where SAP invites any developer to sign up for a 30 hour competitive project brainstorming session. At the Las Vegas event, teams came up with all sorts of weird and wonderful ideas but the one that garnered most attention was a not for profit campaign management solution that included social media goodness.
SAP developers are used to poring over millions of lines of code but what Innovation Weekend is designed to achieve is an appreciation of some of the newer analytics and presentation technologies that can be easily stitched together while leveraging existing SAP technologies. In one case, the team came up with a solution that didn’t require a single line of code. If that sounds shocking then it should also speak to the fact SAP is surfacing solution toolkits that enable rapid development of apps that serve more than a Hello World purpose.
What was striking though was the way developers who had never met one another were able to rapidly conceptualise and build working solutions that can stand business scrutiny. This is a clear departure from the way in which developers have been separated from the business and viewed as glorified plumbers used to spending months and years coming up with working solutions.
Viewed from my line of business stance, this is a significant step in a direction that SAP has to both inject into its own development environment and foster among the partner ecosystem. If successful, then it paves the way for SAP to shed its slow, clunky, not invented here mantle. Does it have all the moving parts in place? Far from it.
I spoke with managers tasked with getting internal buy in. They, along with Dr Sikka readily acknowledge that moving what I term ‘the entitlement engineers’ - those who have been with the company for decades and who represent a deep repository of SAP specific expertise - into this brave new world will be a tough ask.
If SAP is smart, it will find ways of using its ecosystem top talent to spread the word. For example, I see a good case for assembling SWAT teams of SAP Mentors and other top developers who understand the value of these new approaches from a development perspective, lift them out of their existing positions for short periods and take them on the road to show their peers what can be achieved. Showing what can be done at TechEd is fine for reaching maybe 10,000 engineers, but this method allows SAP to hit the mass of developers it needs to persuade and especially those based in Germany.
Alongside, it will need to build a volume based portfolio of applications that can be bought not sold into customers.
Hello cloud
SAP’s critics ridicule its late entry into the cloud arena. If what I am seeing represents the company’s response then its critics need to re-assess their thinking. The Business ByDesign group is now starting to build momentum. SAP shared some of its marketing strategy with me and while I have reservations, it is at last moving in the right direction. Critically, BYD marketing is being corralled into a broader on-demand/cloud marketing strategy that is divorced from Business Suite activities. This clears up many of the confusing dotted line responsibilities that mired the company in painfully slow decision making. Most important, the key budget holders are people that have fresh ideas and can execute upon them. A simple example comes in the way SAP is making contracts far easier to understand.
In an SAP Mentor session designed to throw open questions about how SAP addresses and develops the channel, I was impressed by the way SAP is committing to making it as easy as possible to become a BYD partner. In doing so, SAP committed to having clear guidance in place by the end of the year. The all important cost question came up and here, SAP was open about how it is attempting to come up with a model that both satisfies its internal revenue recognition policies while not killing the channel with outrageous sign on fees. There is much more to come on this but the important difference is that SAP has finally got some clarity into its thinking and direction. Based on past meeting conversations, I sense those who have been interested but hesitant now feel more comfortable going forward.
However the real surprise came in a chance meeting with Cap Gemini. It has been actively deploying Amazon cloud based development systems and in a video I shot with Chris Kernaghan, one of its lead developer/architect, (see top of post) he shared how Amazon based SAP systems can be spun up in minutes at initial cost savings of around 60%. This is not a complete answer and current deployments are limited to single database instances. However, in a post shoot discussion, Chris said that it is only a matter of time before Cap Gemini is able to deploy production ready systems. If successful, then SAP’s customers will have fresh options that dramatically reduce deployed cost.
Customer protection
In the past, SAP has concentrated on encouraging customers to upgrade. The forced march of ECC6 has not been the rip roaring success the company had wished for and neither should it have been. SAP should not be rewarded for inflicting exhausting upgrades even though that might keep developer teams happily employed. The company has finally come to the realisation that a good chunk of its installed base wants to stay on older, stable releases. That does not mean those same customers do not wish to benefit from say mobile applications but they don’t want to walk the one or two year path for something that may be useful but is tangential to the main ERP. Up to present, the only way customers could achieve those benefits was through the upgrade process. Now, SAP is offering an alternative in its Gateway product.
Gateway offers a simple REST API through which customers can take advantage of new capabilities. Is it all sweetness and light? No. Gateway is rough around the edges and has a number of limitations. However, engineers I spoke with seemed broadly pleased with this direction and expect Gateway to offer realistic opportunities that are currently closed to those on R/4.6 and earlier. I encourage readers to bookmark Enterprise Geeks for an upcoming podcast where Ed Herrmann of Colgate Palmolive dives into the topic (among other things) with Dr Sikka. (UPDATE: the podcast is now live.)
Concluding thoughts
During one of our sessions with Dr Sikka I joked that he had failed in his keynote. “So the Innovation Weekend folk get 30 hours to build a new solution and even with four days between keynotes you couldn’t come up with something fresh? Fail !!” The point was not lost on a thoughtful CTO who genuinely believes his technical vision can be transformational but without the disruption which has worn out many of its customers.
The paradox however is not in that direction. It is in SAP’s ability to eat its own organisational dog food. The ability of any company to succeed in executing against a given strategy is often dependent upon the extent to which it is able to carry its employees with that vision.
Dr Sikka readily acknowledges the enormity of persuading powerful interest groups within the company that this new direction is the right way to go and will enhance careers. There are many ways to approach the problem all of which are fraught with risk. However, SAP is unique in that Dr Sikka has the backing of the one man to which SAP employees pay attention: co-founder Hasso Plattner. While Dr Plattner is not involved in day to day operations, he shares a passion for the company that is not diminished with the passage of time. It is a gift he can persuasively communicate when he chooses to do so.
Equally important, I get the distinct impression that Dr Sikka has the trust of co-CEO’s Bill McDermott and Jim Snabe. They can clear road blocks in ways that Dr Sikka cannot on his own. While US analysts will angst over this triumvirate operational team’s ability to maintain a single focus and vision, they forget there is way more power in a three handed team than relying on the top down voice of a single CEO. It may not be quite as agile as impatient analysts would like but it will get SAP to places others cannot follow.
This leadership’s team’s needs to demonstrate to the thousands of SAP engineers that the new software development landscape is both exciting and rewarding. Dr Sikka has put a stake in the ground both in the keynotes and in the discussions he is having with engineers, Mentors and analysts. Dr Sikka equates what he sees combined with the new SAP toolkits as equivalent to the golden days of late 1970s and early 1980s experimentation with Visicalc and the Commodore Pet. It may seem odd to those brought up on PHP, scripting languages, open source and Ruby on Rails. It is exactly the right approach for SAP.
