Wednesday, December 21, 2011

Can SAP be the #2 database vendor by 2015? [Guest Post]

Note: This is a guest post by John Appleby (@applebyj) of BlueFin Solutions. Opinions are his (but I value his opinions!).

Last week 150 people travelled to Boston to listen to SAP talk about their future focus. There was a lot of talk about cloud, muted by the fact that SAP have just acquired HCM cloud vendor SuccessFactors and therefore are unable to talk about the acquisition until it is completed. But their in-memory technology, HANA, was also in centre frame.

And during the conference, SAP senior executive Steve Lucas announced that he intends SAP to be the #2 database vendor by 2015. It wasn't a throwaway comment and hyperbole is not Steve's way. He was clear on who this meant overtaking and clear on this difficulty of this journey. But is this realistic, or just pie in the sky?

For those who don't completely understand: SAP HANA is a next-generation database. At a 30,000ft level it can do anything that Oracle, IBM or Microsoft can do, but hundreds or thousands of times faster, because it runs in the main memory of a computer system rather than on slow spinning disks. I've used it and it does exactly what it says on the tin.

Why doesn't SAP HANA have deeper market penetration?

Put simply it is because SAP wanted it this way. Whilst HANA truly is a general-purpose database, SAP first announced it as an analystics appliance for the 1.0 release. They also priced it really high and didn’t' offer a discount – list pricing can be as high as €180,000 for a 64GB HANA "unit", depending on which version you require.

And what's more, SAP sells solutions and HANA is a platform, so the global salesforce doesn't quite know how to sell it in volume - yet. They didn't want to sell it in volume in any case because they wanted to introduce it slowly to market – building stability, references along the way and avoiding expensive and embarrassing global escalations.

So by the end of 2011 we should expect $100-150m of HANA sales, which is 3-5% of SAP's total revenue. Not particularly significant, right? Well in September they released HANA as being supported for SAP's Business Warehouse software, which allows large-scale data warehouses. And this is where it gets interesting: there are 17,000 existing BW customers, and HANA would provide business benefit to all of them.

What is the wider market opportunity for SAP HANA?

It starts with the SAP BW product, which has 17,000 customers. HANA can replace the database that BW runs on – which is typically from Oracle, IBM or Microsoft. The benefits of this are huge – much faster reporting, data loads, far better agility and a better business experience overall. HANA literally transforms SAP BW.

But that is just the start because the fact that BW runs on HANA means that SAP can allow any of its software to run with HANA – replacing the existing database. And how quickly it chooses to allow this is a factor of how quickly it decides that HANA is mature enough to do this. Larry Ellison from Oracle claimed that there is no in-memory database anywhere near to being able to run as the database for a transactional system at the beginning of 2011. HANA has not yet proven that it is ready yet, but this is exactly what SAP intends.

And even that is just the tip of the iceberg. Because let's be clear: HANA can be used for anything. It supports industry standard connections like ODBC and JDBC and anything that runs a database can be run on HANA – just much faster than ever before. Other vendors have in-memory technologies but the way HANA is designed means it is much more general-purpose than what Oracle and IBM have to offer.

So what is possible by 2015?

Well this is the billion dollar question. From a personal perspective I am deeply impressed with what SAP have done with HANA in 2011. It has gone from being vapourware (January) to being a real product on the price list (July) - albeit immature. And by September a second – much more mature – release was released to market that supports BW. That's an enormous amount of progress in 12 short months.

And I already have projects underway that use HANA as a general purpose database for things like the Sybase Unwired Platform – enabling real-time enterprises with iPads, providing decision making on the move based on events that happened seconds before. There's no doubt that the technology has huge potential. The question is – what happens next, and how fast?

It also depends what you mean by #2 database vendor. For example Oracle say they are the #1 SAP database vendor. Yes – they have the most large systems. Microsoft claim the same, because they have the most customers (a lot of smaller customers run Microsoft). And guess what, IBM claim the same – because IBM have the biggest SAP databases. SAP are going to have to be clear when they explain what they mean by #2.

But based on what I've seen – expect early ERPs to be supported by SAP in 2012 – including the BusinessOne ERP suite – ERP lite, if you like, designed for organisations with 1-100 employees. Expect SAP's SME (10-1000 employee) ByDesign cloud suite to run on HANA and also expect HANA to support standards: because today whilst HANA supports ODBC and JDBC, it does not support 3rd party systems to connect directly into this.

In 2012 we should also expect to see proper support for larger (>1TB databases). HANA compresses around 10:1 to 5:1 compared to other databases so 1TB is 5-10TB of standard database, but it hasn't really been proven to scale properly yet. Expect to see this happen in 2012, as well as scenarios that require disaster recovery and other technical stuff like integration with enterprise monitoring suites like Tivoli.

Why is SAP taking its sweet time?

It seems to be that the answer is pretty simple. The last major database to go to market was Microsoft in the mid-90s with SQL Server. They made an acquisition and it was still awful until the release in 2000. SAP doesn't want this stigma and is therefore phasing the rollout, making the software expensive and thereby limiting the number of customers.

For example there was no shortage of customers out of the 17,000 to join the NetWeaver BW early adoption program that SAP calls Ramp-Up. But there were 50 slots and those were easily filled with customers who had realistic projects that would go live. Little by little they build references, quality software and trust within the customer base – but not growing so fast that there are major project failures. There have been a few instances where HANA was oversold in the early days and those projects were managed carefully – directly by the SAP leadership team.

Projects that use HANA as an ERP database have been deliberately avoided, as SAP Chief Technology Officer Vishal Sikka told me. He described how if Porsche's BW data warehouse were to stop working, they would raise a priority support call and SAP would sort it out. However if their ERP system stopped working and production of cars ceased, he would get a personal phone call from a board member.

So in 2013 expect support for the ERP suite to start to come. And by then, HANA will be sold for every conceivable scenario through 2014 and 2015. And the interesting thing is I don't believe that a killer scenario exists for HANA yet, because all the scenarios right now are really about doing what you do today, but faster. Let's start thinking about what you can't do today – and might give you a huge competitive advantage.

And can SAP be the #2 database vendor by 2015? Really?

Honestly I'm not sure, but it is definitely the right goal. HANA isn't about high-performance analytics – it's about changing the way that customers do business, with a technology enabler. For this reason, SAP have to be looking at going after the database market – and helping customers get a huge competitive advantage along the way.

I honestly suspect that the biggest challenge that SAP have along this way is enabling the salesforce to educate customers on how beneficial it would be – because the SAP salesforce is aligned around industry verticals and Lines of Business. A global reorganisation is underway to change this, and the way that HANA is explained and sold is at the core of this, but HANA is a platform and will need to be sold as such. That is a serious piece of organisational change for SAP and shouldn't be underestimated.

Regardless of whether Steve's goal is met or not, SAP HANA is perhaps one of the most interesting technologies I have seen in my 14 years of working in Enterprise IT, and well worth serious consideration, whatever business you run. If you are a SAP customer I would go a step further and say that you should look at building your HANA roadmap, based on your business imperatives compared to the product maturity and availability roadmap.

Disclosure: SAP paid for John's travel and expenses to the Influencer Summit in Boston. Mine, too, btw, and SAP is a client of mine (Dennis).


  1. It will be a HUGE surprise for me if SAP becomes the #2 database vendor by 2015, for a few reasons.

    1. No general purpose DB matured in 4 years - not ORCL, not DB2, not MS . I am so happy Vishal chose to not go aggressively into ERP on HANA. Good call - it is NOWHERE near doing that now.
    2. Current top 3 have dedicated salesforce who will come hard at SAP now - and SAP does not have enough dedicated DB sales force to counter in short term.
    3. Current top 3 incumbents are not stagnant - they are still growing and now control 75% of market by revenue. Lucas explicitly confirmed in Boston that he wants to be #2 by revenue. That threshold is pretty high to get into #2.
    4. As HW prices fall - with SSD, RAM, Fusion I/O card etc becoming cheaper - will SW continue to demand a premium with performance as the big benefit?

    Now - there is one possibility. All DB vendors define markets arbitrarily - and I think you alluded to this above- and SAP can probably define something along those lines to claim #2. They wouldn't be the first to do that.

    So all things considered - I don't see this happening. But being an SAP consultant - I would love to see SAP claiming that spot and proving me wrong.

  2. John - Thanks for a great guest post!

    Vijay - You said a couple of things I agree with, but a couple as well I don't agree with. SAP does have a dedicated DB sales force - Sybase. It doesn't seem the Sybase folks are involved enough in HANA yet, but hopefully that will change. Still, I think SAP has a tough row to hoe to get to #2 by any measure. I guess they'll probably pick some metric (e.g., "new database license sales to enterprise customers") and declare victory ...

    -- Dennis

  3. It was an extremely interesting statement made by Steve Lucas and indeed he said it is going to be by revenue. But I never read this as a comment that it is SAP HANA database (HDB) alone that is going to be #2 database by sales, as I would read from the John's thoughts here. It is obvious for me that the driving arm would need to be Sybase, which was not mentioned in the blog at all.
    We are all in the violent agreement that it depends on what and how is measured. For the simplification I assume we are looking at the commercial RDBMS market, leaving aside open-source RDBMSes and non-relational ones, with known split:
    #1 - Orale
    #2 - IBM
    #3 - Microsoft
    #4/5 - Sybase / Teradata
    So, it is not that SAP starts from zero to get to #2 spot. Current SAP's db portfolio is HDB, Syabse (ASE, IQ and some smaller) plus SAP's very own MaxDB (~800 users today). The issue is that Sybase share is around 3-5 percent, while current #2, i.e. IBM, has around 22-25%. The percentage difference is huge, so you cannot close this gap without some trick. And the trick in my opinion is in SAP+BOBj customer base. I think it is a no brainer that db growth in coming years will come from approach to replace existing databases by Sybase ASE for BusinessSuite customers (I will not bet for HDB in this segment now), by HDB for BW customers, and by Sybase IQ for BObj BI ones. It is a huge potential number of installations, and every replacement with any of SAP's databases reduces competitors' installation by one as well. The calculation of revenue will still be tricky though.
    Thanks for bringing this topic to discussion.

  4. Vitaliy -

    Good points as always. Thanks!

    - Dennis

  5. I don't think that is a possibility in near future. Not just because HANA is new it will take time to mature, but also because the leading vendors will be very aggressive in retaining the spot.

    1. I was a little surprised when several SAP execs made this claim, since such a shift has not happened ever before in the history of enterprise software. However, HANA is off to an amazing start ...