Can SAP HANA bring real savings? Yes and the numbers are impressive

SAP HANAI don’t think anyone will disagree that SAP’s HANA marketing has gone stale. Speeds and feeds didn’t do it. Talk of a platform hasn’t done it – yet. Even having more than 1,100 startups in its HANA funnel hasn’t done it. If I mention HANA in analyst company, the collective eyeballs usually roll as colleagues grumble about  HANA being a technology play with few applications behind it. Even news that SAP is putting the Business Suite on HANA draws little more than shrugs and yawns.

Some analysts regularly complain that SAP’s ongoing costs are among the highest in enterprise land, that SI’s continue to take huge chunks of change out of customers and after 40 years, SAP still draws a bad rap for project failure.

What has been missing is a discussion around total cost of ownership or TCO in the context of SAP HANA.

SAP itself has talked about HANA offering the opportunity to simplify the SAP landscape. In recent times, the company has obliquely referenced a reduction in footprint of 10, 20 or 30x with newer iterations leading to 50x footprint reduction. But what does this mean? Less storage – a thorny issue in its own right – less hardware resources. Another pain point…??

Part of the problem rests in the fact there is little incentive among SIs to do the math necessary to make a solid case for HANA. Why simplify when complexity is your billable hours’ timesheets best friend?

One SI has bitten the bullet and undertaken a study on behalf of a large government department that has an SAP estate. They’ve looked at the WHOLE estate, not just one piece that can be easily picked off, and undertaken a thorough review of what is achievable.

I cannot get into the detail for reasons of confidentiality. What I can say however is that in this particular case, the planned savings are substantial. With HANA including licenses, hardware, project and support, it will cost them $17m over 5 years compared to the current ‘as-is’ cost of $30 million. That’s a 43 percent reduction. How did they get there?

There is a combination of factors in play, not least a renegotiation around the SAP estate and the replacement of proprietary UNIX based systems with x86 boxes running HANA instead of the existing database. In other words, the SI has taken a whole world view of the situation in order to understand, cost out and assess savings well beyond a database replacement.

When viewed objectively, this is a classic case of disrupting the existing model in order to bring in a level of renewal that is being demanded by governments around the world.

Although not quantified at this stage, there should be other tangible business benefits beyond the obvious TCO arising out of this simplification project. According to my sources, the department is viewing this as a stepping stone towards re-imagining reporting requirements in what will be a real time environment. This will open the door for fresh opportunities for SAP to upsell new analytic applications at a later date.

Impressed? As far as I know, this is the first time numbers of this kind have been put into the public domain and while very much futures based, is confidently expressed in the ‘stuff’ I have seen. We need more cases of this kind to show how it’s done and the benefits that accrue.

Endnote: EMC has recently talked about performance improvements of 400 percent in its data retrieval activities. The interesting part about the EMC case is that it makes a good marketing case for its own solutions and those of VMWare alongside HANA.

Disclosure: at time of writing, SAP is a premier partner. It had no hand in this story.

Featured image via Fasttrack

Den Howlett
Following 20+ years in finance and IT related roles, Den Howlett became a freelance writer/analyst/commenter specialising in enterprise IT. He has been calling the vendors to account ever since.
Den Howlett

@dahowlett

Disruptor, enterprise applications drama critic, BS cleaner, buyer champion and foodie trying to suck less every day.
Den Howlett
Den Howlett

Latest posts by Den Howlett (see all)

  • dahowlett says:

    paulvero21 – Paul: let’s not confuse facts for opinion. By gating point I mean the choice of input system…or rather method.

  • paulvero21 says:

    dahowlett
    Hi Den, “laughable” is a bit harsh perhaps, but my impression is that there is still plenty of HANA marketing and hype, and its not just me who thinks that – witness the recent ZD Net story “SAP executives stopped short of saying the company’s HANA platform would cure cancer, but not by much”; SCN blogs are titled “HANA Speeds Cancer Research and Results”, etc.

    Regarding your first question, what exactly do you mean by “gating point”?

    PV

  • cochesdiez says:

    jonerp awedome. I will see that I add more pov to this topic
    #thanks

  • jonerp says:

    .cochesdiez re: our convos on CIOs in transition, hopefully you saw this http://diginomica.com/2014/01/29/riffing-cio-cloud-dilemmas-sina-moatamed/

  • cochesdiez says:

    jonerp thanks a lot buddy

  • jonerp says:

    cochesdiez happy bday dude :)

  • dahowlett says:

    paulvero21  I think we need some specifics on this. To the point on OLTP – does that matter when the gating point is often the input systems?

    As to my claim that marketing has gone stale – how is this ‘laughable?’ My claim is based upon the emphasis on ‘speeds and feeds’ whereas there is a much bigger story IF SAP can get its act together on platform benefits.

  • paulvero21 says:

    EMC are right to be sceptical about ROI. 
    Consider this real-life example. A 3TB ERP database, performing perfectly well.  Uncompressed it’s about 9TB.
    SAP guidance for Suite on HANA suggests RAM requirements are 5TB for current data, so say 6TB allowing for minimal future growth.  6TB single node is not yet a supported configuration but no doubt will be very soon.  A 6TB server will be an expensive bit of kit, and then there’s the one for DR…

    We know that HANA architecture will not outperform row-store for true OLTP, though code rewrites by SAP to benefit only HANA will bring it in-line with other database performance.  For batch processing SAP will make HANA outperform using stored procedures, an approach that they could use for other databases but will not, to gain competitive advantage.  However we know that much cheaper solutions such as SSD can speed batch up for other technologies (admittedly not as much as a stored procedure approach).

    We also know that we can get good BW performance by other, much cheaper means.  The true HANA benefit lies solely in its ability to give real-time answers, not delayed by e.g. hourly loads to BW using side-car approach.

    Businesses should really ask themselves, how many operational reports are so critical and performing so badly to justify the huge investment?  Sure, for some businesses, this will be the case – but I bet they are a tiny percentage.
    The biggest driver for HANA is hype, and I find the claim that the marketing has gone stale to be laughable.  If customers aren’t buying in fast enough that’s because they aren’t all fooled.
    Outside of SAP-land, SAP cannot force people to so the necessary code rewrites to make OLTP performance adequate, so it will remain a datamart-only product with stiff competition from elsewhere.

  • paulvero21 says:

    praba01

  • johnstuddert says:

    Exciting times for sure. The only cold water I’d throw on this is that with any study there’s a danger that the SI is tempted to make the numbers look as appealing as possible where there’s any scope for interpretation (unless perhaps they’d be ruled out of any subsequent implementation, but that seems unlikely), so the real value is usually in how the numbers pan out after the implementation/transformation.

    This would go doubly so for such a new and relatively untested area, given the very fact that as you say, so little discussion and hard numbers are available on TCO for this type of complete transformation.

    You obviously have confidence in the experience and reliability of the SI involved though, and even if the immediate numbers don’t end up quite so appealing there are other future advantages in the transformation as you say. Appreciate the insight, keep it up!

  • cochesdiez says:

    timoelliott I love the “Except when you buy it and run it” !!!! ++++1

  • TimFisherUK says:

    dahowlett certainly interested in a chat. DM me and we can sort something.

  • dahowlett says:

    TimFisherUK Want to share numbers?

  • TimFisherUK says:

    dahowlett diginomica agree with this. We’ve been saying the same. Need to take a whole landscape view or the case can struggle to stack up

  • applebyj says:

    JakeEchanove dahowlett diginomica usually it is a bunch of things. HANA is usually an enabler rather than a driver.

  • holgermu says:

    applebyj dahowlett Good point. Then 8 months is not soo old. ;-)

  • dahowlett says:

    JakeEchanove more than that…

  • JakeEchanove says:

    .dahowlett diginomica wld be the same with other DBs and migrating to x86? Is it actually HANA that is driving the savings? applebyj

  • applebyj says:

    holgermu dahowlett Makes sense on commodity. Why charge for older content?

  • holgermu says:

    dahowlett What is commodity now – wasn’t 8 months ago. Time flies.

  • dahowlett says:

    praba01  Hi Bala and thanks for asking. I am satisfied my sources are giving me good information. 
    On the EMC thing – they are working through this but are pretty clear about value – just not expressed in ways I’d prefer to see. We have covered EMC before when it was piloting re: BW

  • holgermu says:

    applebyj dahowlett It’s the – then – pretty new technology of #HANA measured against them. So nothing new.

  • praba01 says:

    Dennis,
    “$17m over 5 years compared to the current ‘as-is’ cost of $30 million” – Significant, no doubt. As you stated,there’s a combination of factors in play. Two major factors come to my mind:

       a) As far as I know, public sector solution is not available on HANA yet. I hope SI knew something we don’t know on time & efforts needed to make PS available on HANA.
       b) What are current DB vendor’s plans for next 5 years? 

    From EMC Article: 
    “Many SAP customers are struggling to forecast ROI on potential HANA investments, as the licensing and infrastructure costs can be daunting.  While there will be short term limitations (we fully expect HANA to be a mainstream offering within the next 5 years), virtualizing SAP HANA allows a reduced infrastructure demand and thus cost management. ”

    I’m a bit confused:  EMC article seems to suggest TCO may be more with HANA migration but ROI will be even more; at this point, customers are struggling to forecast ROI. However SI’s report suggests HANA migration would lead to lower TCO. These reports contradict each other, no?

    Best regards,
    Bala Prabahar

  • applebyj says:

    dahowlett holgermu Hmm not sure you should need a research report for this. Hardware, SW license, people, maintenance. 4 dimensions.

  • dahowlett says:

    holgermu why would I go behind a paywall when I can get the numbers for free ;)

  • holgermu says:

    dahowlett Good point – interesting cost savings for CRM to CRM / HANA – my first research report… https://www.constellationr.com/research/behind-roi-deploying-sap-hana #inMemory

  • dahowlett says:

    greg_not_so WYSIWIS…

  • greg_not_so says:

    dahowlett diginomica Can SAP HANA bring real savings? Yes, and the numbers are impressive. http://ow.ly/tfzPa” < p->x hw/sw switch?