SAP Rumored to Launch Hana-Based Cloud Planning Solution Next Week


Well, it’s pretty clear that SAP will be announcing something in Chicago on Monday, judging by the above re-tweet from SAP head of EPM product management David Williams, and by the content in the SAP Conference for EPM program, itself:

SAP Keynote and Panel Discussion: Next-Generation EPM in the Cloud

In fact, SAP’s vision is to be “The Cloud Company, powered by SAP HANA.” What does this mean for SAP? What does it mean for SAP’s EPM portfolio? And what does it mean for you as a customer? In this interactive keynote, SAP executives, partners, and customers will dive into how the cloud and software as a service (SaaS) are impacting finance transformation and how the next generation of EPM in the cloud has arrived!

If looks like a product launch, walks like a product launch, and quacks like a product launch, then it’s probably a product launch.  Judging by the description, SAP will be launching a new, cloud-based, Hana-based EPM solution.  We’ve heard elsewhere that it will be focused specifically on planning.

A Brief Rant on Hana
Before diving more into EPM, let me comment a bit on SAP’s Hana strategy, which I find quite confusing.  In my estimation, SAP has two problems with Hana.

  • An unhealthy obsession with the database market born from their historical dependence on Oracle.  Instead of letting time make databases irrelevant (as cloud services do), SAP chose to enter the database fray both through the acquisition of Sybase and the development of Hana.  In my mind, SAP would have been better off to simply let databases commoditize (and put their apps on Postgres or a NoSQL system).  Instead, they did the exact opposite — and market it heavily with an ingredient-branding strategy around Hana.  The message isn’t “our apps are better.”  It’s “our apps are better because they’re on Hana” which is both only partially true today and requires the logical leap of faith that being Hana-based would invariably make one app better than another.  (My cake is better because I use cane sugar.  Maybe.)
  • An illogical desire to conflate Hana and cloud.  The two concepts are orthogonal.  Hana is a column-oriented, in-memory relational database system.  Cloud computing is a delivery (and business) model for software.  You can build cloud services on whatever kind of database you want — the world’s most successful cloud company,, builds atop Oracle.   A key idea of cloud computing is that infrastructure becomes irrelevant.  Just as you don’t know where Salesforce’s data centers are, what brand of servers they run, and what operating system runs on those servers, you don’t need to know what database system they run.  That’s the point.  So to conflate cloud and Hana is illogical and confusing.  SAP badly wants Hana to mean “cloud” and if they keep pushing it eventually will, but then it won’t mean column-oriented, in-memory database.  Because the concepts are so different, Hana can mean one or the other, but it can’t simultaneously mean both.  (Heinz used to mean pickles.  Now it means ketchup.)

While I am on the record saying that SAP’s Hana strategy will “work,” I believe that is not because it is a good strategy, but rather because it is a strategy to which they are highly committed.  And with enough financial might,  you can drive almost any marketing message home (e.g., “with a name like Smuckers, it has to be good.”)

Back to EPM
Nevertheless, the new EPM solution will invariably be Hana-based and thus a good deal of the launch presentation will therefore describe the supposed benefits it inherits from so being, but is being Hana-based a good thing for an EPM system?

Hana is an in-memory relational database.  Because multi-dimensional analysis is so critical in EPM, virtually every commercial EPM system in the market runs atop a (typically in-memory) multi-dimensional database.  (In fact, in-memory multidimensional databases predate in-memory relational databases by a decade or two, dating all the way back to TM1.)

So, SAP’s new EPM solution will not only be the world’s only EPM solution atop Hana, it will the world’s only EPM atop a column-oriented relational database.  Other than BPC, of course, where Hana is positioned as an accelerator.  But if being Hana-based made an EPM system both great and cloud, then why do we need the new offering?  Because Hana has nothing to do with cloud.  (See other rant.)

Net/net:  While SAP will invariably position multi-dimensional databases as evil (“rogue cubes,” as Sanjay Poonen called them when announcing BPC on Hana), I’m not convinced that running an EPM system on Hana is a great idea, as opposed to running on a multi-dimensional database. (But SAP doesn’t have one of those).  Performance and scalability will be the test here and time will tell.

Keep It Simple
The other big message coming out of SAP of late is simple.  (So the company that stands for applications and complexity talks most about databases and simplicity.)   Nevertheless, the simple message may provide a clue for what’s coming so let’s take a look at SAP’s 21-page statement of direction on simplification of their business intelligence solutions.

As business intelligence (BI) solutions from
SAP have evolved over the years to address new
customer needs, introduce new innovations,
and take advantage of emerging technologies,
the portfolio has grown to encompass a large
number of client tools. While these tools provide
best-of-breed experiences in specific use cases
and together provide a comprehensive BI suite,
it can be difficult and confusing today to choose
which tool to use.

From this you might conclude product convergence on the EPM side, but the introduction of a new cloud-based solution actually increases product line complexity as SAP will need to differentiate, among other things, when to use BPC, when to use BPC on Hana, and when to use the new offering.  This is common in large vendors with complex product lines and is sometimes called “having to sell against yourself.”

The only EPM-specific information I could find in the 21-page statement of direction on simplification was this:

Rounding out these two central tools, we plan to offer a single interface for Microsoft Office integration, based around the edition for Microsoft Office of SAP BusinessObjects Analysis software. Our intention is to address use cases covered today with SAP BusinessObjects Analysis, SAP solutions for enterprise performance management (EPM), add-in for Microsoft Excel, and SAP BusinessObjects Live Office software through a single add-on that is planned to provide access to any data, analysis, and planning.  We also anticipate embedding live visualizations and dashboards created by SAP Lumira and SAP BusinessObjects Design Studio within Microsoft Office documents.

This means a single new Office interface is coming that includes EPM and goes beyond it.  Complex in the sense that it will another interface and a new one, but simple in that it should replace several different interfaces over time.  I’d guess this is all about BPC and not the new cloud offering, but I can’t be sure at this time.

Deja Vu All Over Again
To be a little snarky, I feel compelled to remind readers that SAP has already announced a cloud performance management solution, SAP EPM on Demand a little more than two years ago

sap epm 1

And which they subsequently shut down

sap epm 2.

So What Does This All Mean?
We’ll obviously know more after the announcement, but I drew several conclusions from recent history and anticipated moves:

  • Cloud computing is continuing to transform enterprise IT and that the megavendors continue to increasingly realize that.  They are squarely out of the “denial” phase of the market transformation and working actively through both M&A and in-house development to offer cloud services.
  • The Innovator’s Dilemma is a very, very hard problem for businesses to manage when dealing with disruptive change.  Not only do megavendor incumbents have legacy products that add complexity and risk cannibalization, they have legacy business practices as well.  I’ll be very interested to see SAP’s pricing and hope that they’re not copying Oracle’s strategy of “any color you want as long as it’s same number as on-premises” (to paraphrase Henry Ford).
  • SAP believes that finance is increasingly ready to go cloud.  Because finance has generally been slow to “go cloud,” I view this as yet another sign of increasing cloud-readiness in finance.
  • While SAP remains highly committed to Hana, they are “out on their own” in running an EPM system on a column-oriented, as opposed to multi-dimensional, database system.
  • SAP’s new cloud EPM offering will definitionally be a v1 product and likely take a few years to reach maturity and functional completeness.

2 responses to “SAP Rumored to Launch Hana-Based Cloud Planning Solution Next Week

  1. I am an SAP Employee (and so couldn’t possibility comment on Monday’s announcement :-), but here are some of my personal thoughts on the HANA rant:

    (1) “SAP would have been better off to simply let databases commoditize.” SAP was/is under huge pressure to offer faster, more flexible applications and it’s clear that company sincerely believes that in-memory “HTAP” systems offer unique and dramatic advantages for enterprise applications (as do veteran database analysts such as Donald Feinberg: If a credible alternative in-memory, ACID-compliant system existed, maybe the story would have been different.

    (2) Conflation of HANA and the cloud. Just as SAP believes strongly in HANA, it also believes it’s the right platform for cloud solutions:

    Because hybrid combinations of on-premise and cloud are going to be the rule for the majority of companies for the foreseeable future, using HANA for both has advantages.

    But HANA also helps the transition in a different way. SAP customers expect any cloud-based solutions to be compatible with their existing on-premise systems. One way to do this would be to provide the same apps in the cloud. But SAP has a lot of code (an estimated 400 million lines) and rewriting it all for cloud is an enormous task.

    But just as most of the data in an ERP system consists of caches and aggregate tables to ensure performance, it turns out that a large chunk of application logic involves updating and synchronizing all those caches. When HANA eliminates aggregates, it’s not only the amount of data that falls dramatically, it’s also the number of lines of code – making the cloud transition easier, at least in theory:

    This is, of course, separate from the marketing aspect of using Cloud and HANA together, but as you point out, SAP is strongly committed to an ingredient-based (or “permission to believe”) strategy, believing it has a unique head start in a technology that will underpin all applications in the future (moves by other actors in the industry would so far seem to back this up).

  2. Timo,

    Thanks for reading and thanks for your comments. Let me respond to your two points.

    1. To say that SAP, an applications powerhouse, should not be messing around in the database business is not to say anything bad about column-oriented databases or in-memory technology. I am “all for” column-oriented databases and in-memory computing. My point is I don’t believe SAP should have put themselves in that business. Might they succeed anyway? Yes, quite possibly. Might it have been the right strategy 10-20 years ago? Yes, quite possibly. But today you have highly-scalable RDBMS alternatives being, open-sourced by major vendors, in use across thousands of companies and supported by major vendors (e.g., Hadoop, MongoDB). Would SAP not be better off leveraging these next-generation stores (that now do exist) as opposed to building its own? I have followed column databases since C-Store and am one of the few who’s actually read Hasso’s paper from SIGMOD 2009. I have a great handle on databases. That’s not the question. The question is should SAP be making one? That is far less obvious.

    2. I am aware that SAP believes that Hana is the right platform for cloud applications, but I could point out that way more real cloud applications run on Hadoop, BigTable, Cassandra, MongoDB (and for that matter, Oracle) than run on Hana. But again, I am not oppose to Hana and I am “on record” saying I think it will work due to SAP’s huge commitment to it (and that commitment does matter). But I am simply pointing out (again) that building and delivering cloud applications can be done on top of virtually any infrastructure. Moreover, since you opened the topic, I found it somewhat ironic that Hasso himself did ” a post starting with “it is amazing to me how little the benefits of the Suite on HANA are understood or even known in general and by the members of the Americas SAP Users Group”. Moreover, I would argue that it is always difficult to argue why people should be doing something that they are, in fact, not doing. The fact is, very few people are building cloud-based apps and services on Hana. NoSQL and regular RDBMSs probably 100x outfactor Hana. As I learned from you when we discussed the second serve in tennis — let us assume players are correct to hit a softer second serve and then try to figure out why. Let’s not assume they are wrong to do so and try to prove it. Hasso’s post feels a lot like arguing the latter.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.