SAP HANA gut check

In the run up to SAP TechEd Las Vegas there are three major themes we can expect to hear a LOT about:

• HANA - SAP's pitch for high speed, in-memory analytics and applications
• Mobile - some of us have been privileged to see some pretty spiffy demos of new applications with plenty of iPad love thrown in.
• Analytics - activists in the BI space have been moaning about delays to GA for BI4. SAP hopes to shut them up.

For this post, it is HANA that intrigues me. According to Bill McDermott, SAP's co-CEO, this represents one of the fastest growing pipelines in the company's history. He has consistently said at each of the last few financial analyst calls - and in between - that HANA's pipeline is running at €10 million ($14.5 million) a week. Last week I heard that mobile represents a $0.5 billion pipeline. Yet by years end the company says that HANA and mobile will have delivered $100 million in additional revenue. That doesn't sound so great to me. What's the story?

Software SNAFUs

Most of those I speak with believe SAP is betting much of its future on HANA - period. Yet its early adopters are not always having a great time. From what I can gather, a slew of bugs, combined with development gaffs have left SIs frustrated at not being able to do the simplest of things and in some cases, having to completely reload data sets each time they stop/start a workload operation. At least that was the case until very recently.

Internally, I understand there is a gap between the invention teams who are doing the baseline innovation and the developers who have to carry it forward. I cannot figure out where the dividing lines fall. There are named individuals who are taking communications and product responsibility but in recent times, it has been necessary to escalate all the way to the executive board in some cases. That should not be necessary. And no-one I have spoke with can provide me with a clear answer as to where the buck stops. It is partly symptomatic of a company run by consensus but someone must have the last word. The good news is that SAP is engaging and acting upon customer related issues at a speed that is unprecedented

Another of the big complaints comes from wondering where the applications are that sit on top of HANA. There are many potential use cases in the works (many centered around analytics) and last week, Sanjay Poonen, corporate officer told me that there will be HANA apps on show at TechEd. Given the recent past experience, they'd better be prepared to blow a few people's socks off.

To its credit, SAP is responding as quickly as possible and at the highest levels. Release 13, which came out towards the end of last month resolved many hairy issues but there's still plenty of elbow grease needed to heft this baby. But even so, it does not bode well for the future. When you have a recent history of frustration, people are less disposed to being kind. I fully expect some robust conversations on this topic at TechEd.

Despite the understandable grumbling, much of which comes from the gap between hype and reality, it is a little soon to be overly harsh. SAP is using HANA development as a way of revving up its agile approach to software development. This is ambitious for such a solution and they need to work out all the kinks.

But the issues don't end there.

Hardware bingo

HANA is a combination of three things: software, services and hardware. It is this last part that is driving some colleagues crazy.

SAP is trying hard not to be in the hardware business. It does not want to be corralled alongside Oracle as a full stack provider. But it has specified and certified certain machines upon which HANA can run. This is because its licensing model is tied to the amount of memory a HANA machine uses. The current supplier list includes Cisco, HP, IBM, Dell and Fujitsu.

The problem comes when the SIs try get a competitive bid together. Prices appear to be all over the map, despite the relative similarity in components needed to build a HANA box. There are some important differences and buyers will need a clear understanding of the different processing capabilities of the CPUs on offer.

What's more, with SSD prices falling - yes - you need storage - it is hard to know what is a fair price to pay for these machines. This puts SIs in the unenviable position of having to act as hardware negotiation brokers for their customers, a task with which they are not overly familiar but which they can learn given time. Except they don't have time as customers are asking for proofs of concept and those are proving difficult to fullfil. And to add a little spice, it is a challenge to crank out a specification from the available price lists on the vendor websites without coming up with apple and orange comparisons. I've tried.

And despite what I say about SSD pricing, when you're getting into terabyte territory, prices are at  nosebleed levels. Will customers wait for prices to fall, especially if ROI becomes problematic in the C-suite? SAP is convinced the use cases among charter customers are dramatically compelling. Fine - but what happens when you start selling to the next tier, those who will be forking over for the full enchilada?

There is precedent for the negotiator role. Salesforce.com tier 2 SIs are helping large customers negotiate license contracts in the absence of a well defined price list for enterprise and in the context of integration or customisation projects. It's not the same but it is a similar principle.

Will SAP SIs get embroiled in this issue or will there be fruitful times ahead for Vinnie? I'm testing the waters and we will see what emerges. My own view is that SAP should take ownership of the problem because after all, it is their solution when viewed as a package.

IBM under the gun?

Which brings me to a problem that is harder to resolve. IBM is both a hardware and services supplier for HANA. How is it going to put together a competitive bid when other SIs can come in with commodity kit from Dell or possibly HP? Should it offer competitive kit if that fits the bill? Best procurement practice would suggest that is the right way to proceed.

But that leaves IBMs hardware people potentially out in the cold. They want compensating the same as the services people but right now it looks like IBM overall will have to take some sort of hit if it wants to be in the game. That's going to be about as welcome as a wet fart in a spacesuit, even with some of its most trusting customers. IBM can assemble bids that compensate for what will be an imposed problem but it will still leave an unpleasant taste. As a side note, will HP be able to compete  in combined bids? In their current state I'd argue - unlikely. And from what I know of pricing, very unlikely.

SAP will argue choice and its large customers will have deals that get them kit at favorable prices from their vendor of choice. But the fact remains these machines are an expensive component in a POC that form part of an indivisible package.

Final thoughts

When you stand back and look objectively at the totality of HANA what we see is a super hyped piece of technology that has got customers excited in advance of where the technology is ready to deliver. Early success reports sound spectacular but if you listen carefully, then you discern that much of what we are seeing is anticipated rather than in use, in anger. We need to hear customers talking deep value and that's not really happening. If anything, some appear to be taking a punt.

In new development, SAP is trying hard to unshackle itself from its sloth like past. It is developing quickly but not at a speed that matches the marketing hype. A bit of throttling back on the general marketing and more discussion on deep success is the required medicine.

You don't flip an organisation of this size overnight or even in a year. These are growing pains. Coordinating feedback and having a way to fast track problem resolution as an everyday part of the development lies at the heart of keeping the SIs in the game.

The hardware problem is not going away and without a concerted effort on SAP's part, customers are going to be both confused and out of pocket. Can SI's fill the gap? They may have no choice but in the meantime, what to do about IBM?

I'm sure SAP is thinking about all these issues and many more but evidence is needed to help smooth out the bumps. TechEd represents the perfect forum for that to happen and I hope that SAP sees what is written here in that light.

While TechEd remains a largely technical conference, the commercial issues will need attention. How those get hashed out will be just as interesting to watch as the TechEd 'screwdriver' sessions.

Back to the numbers. Those same early adopters are likely not paying for the software because of their 'all-you-can-eat' license deals that entitle them to everything SAP has to offer in exchange for a very large check. When those deals wash through we will then need to see how the pipeline turns into revenue. Given the issues I've outlined, it will be 'interesting.'