Return on HANA investment - two SAP partners weigh in
- Summary:
- Return on data is far from easy. Here's video and analysis on how two SAP partners are getting it done for HANA customers.
When it comes to return on data, enterprise customers have a sticky wicket indeed: solutions for (sic) better results are everywhere, but the proliferation of data is also expanding at an intimidating rate. Net-net: the cost of getting a return on data remains a considerable problem. In SAP's case, HANA use cases are maturing to the point where we can talk numbers, not just speeds. Den dug into this with a Molson Coors video use case and analysis.
But despite HANA's potential, many customers remain wary of the cost-versus-return. Cloud licensing options are a welcome addition to the HANA mix (though with a 1TB limit currently) - but in-memory at scale is rarely cheap. Making a bad purchase because you were seduced by sexy technology is also expensive for IT leaders - from a career standpoint. That means you better have a tangible business outcome charted, and confidence your team has the right approach. At least that's the sentiment I heard from customers I spoke with at the SAP HANA 2014 show in Orlando this week.
In the past, SAP's partners have not been as aggressive as I would like on the issue of cost innovation (actually that's a gross understatement, but anyhow...) On the HANA 2014 agenda, two partners, Dolphin and Teklink, tackled the HANA TCO/ROI issue head-on in their presentations - albeit in different ways. Den and I got them both on video; here's what we learned.
Dolphin: reduce HANA TCO via better info management
Dolphin addresses the BW on HANA use case, one of HANA's key scenarios for the existing SAP install base. Dolphin does not implement BW on HANA. Instead, they help customers tackle information management issues before the installation begins. End result? Reduced HANA TCO via smart use of near-line storage and a leaner BW data model:
During our conversation, Dr. Werner Hopf framed the problem:
The cost of storage is getting less expensive, but the data volume being processed eats up all the technology advancements. At the end of the day, the cost of ownership is still up pretty high. You always have the two dimensions, performance and processing capabilities versus costs. You're always operating within limits there. Now HANA definitely pushes the boundaries in the direction of performance, but a lot of companies are very concerned about how much does it cost to operate that system.
Hopf believes proper information management, including savvy use of near-line storage, can bring HANA TCO in line with the BW on HANA benefits. The catch? Companies have to spend time getting this architectural plan right before they dive in (also: Hopf points out that the intensity of data used by a company/industry is a key variable to consider). To back up his views, Hopf cited a successful Dolphin project, an energy company in California:
One of our customers is an energy provider in California, and was one of the first HANA pilot customers for SAP. We helped them implement a near-line archiving strategy that greatly reduced the growth rate of their PW and HANA environment, and ultimately enabled the company to achieve a positive ROI for their HANA project.
Case study notes on Dolphin's web site provide more detail on this project's numbers, including a 50 percent reduction in database growth (and stabilization of database growth going forward), and data provided for analysis 'at a lower cost, without performance loss.'
TekLink: change HANA from a technical to a business conversation
TekLink's approach moves speed for its own sake aside, and puts the HANA business case in the center. This means engaging process owners in a HANA conversation they may have never had before, and building out the project financials around those requirements:
During the shoot, TekLink's Tony Guetersloh explains how open-ended conversations with process owners lead to metrics that guide the project. I asked him if process owners are at all surprised to be included in these conversations:
With most of them, the reaction is, 'Wow! No one has ever asked us that before. Usually IT has always dictated what we could do and what we can’t do and no one has ever asked us what would be the possibilities if we could whatever we want.' Usually conversations go very well. I think the trickiest part then is actually putting a dollar value on that. If you could do this - how much would that really save your organization, or how much additional revenue or capabilities value would that really enable?
Aside: I liked the openness Guetersloh expressed about not pushing every project to go forward. That's the advisory mindset today's buyers expect, and let's face it, not every conversation about business possibilities can be translated into HANA reality. But in some cases, the project does move from business case to proof of concept to go-live. Guetersloh gave one example of a large consumer packaged goods company (think candy and confections):
We did our value wheel approach with them. They had a scenario where they don’t have necessarily real-time views to their inventory, and so they have to keep (items) a bit longer in inventory in their distribution centers, because latency around their loads and ETL and all that. They told us that if they really could get a real-time view, they might be able to reduce inventory and reduce their working capital by $7 million a year, which is real value savings right there.
But as Guetersloh notes, you can't get dreamy with real-time - you still have to take risk factors into account. So they plugged the numbers into TekLink's risk-adjusted business formula, which begins with a calculation of Net Present Value (NPV):
We baked that into our value wheel, and then we also took it with a grain of salt and said, 'Okay. This is best case scenario, but now, what’s the real probability that even if this went well, what is the probability you would achieve that much savings?' Then I said, 'Well, maybe between 60 to 80%.' We applied a discount rate to that - just like in a finance model - so that it factors in the risk, so we’re not just going to an executive saying, 'Yeah, you’re going to get millions of dollars, just sign a check,' but actually saying, 'No, we’ve actually qualified this, applied a discount rate that even if we don’t get the best case scenario, you’re still going to get some value out of this.'
Here's a look at Teklink's NPV-based business formula, slide 4 in their presso deck:
TekLink often recommends a Proof of Concept to put the numbers to the test. During his HANA 2014 presentation, Guetersloh walked us through a Proof of Concept that TekLink performed for a BPC on HANA evaluation.
In this case, performance was benchmarked against several systems: System 1 – BPC NW 7.5 on Oracle, BPC NW 10.0 on Oracle, and BPC NW 10.0 on HANA. BPC on HANA showed significant performance gains across six criteria, including Ad Hoc Reporting and Data Refresh:
Those numbers then get plugged into an overall Net Present Value calculator. From there, an informed decision can be made based on a risk-adjusted business case. There's a lot of meat on the HANA bone here; I hope to do a follow-on piece that digs into this business case method more closely.
Final thoughts
These HANA approaches are not magic bullets, nor is that their intent. But it's encouraging to see partners tackle the ROI challenge head-on. Firms like Dolphin and TekLink can keep larger firms on their toes (I won't name names, but you know who I am talking about) with advisory-type approaches that do not require a waterfall budget or mentality.
That said, I am still an advocate of giving customers potent HANA apps they can easily try and consume, perhaps in the cloud, and definitely with cloud-style pricing. Many of the HANA solutions on the table today involve bigger projects around the existing SAP application portfolio. The SAP Startup Program aims to change that, and to SAP's credit there was a strong startup presence at HANA 2014, but that's a topic for another time.
End note: For the curious, we released the Dolphin video in black and white because, well, stuff happens with on-site video. Nothing Den's post-production tricks can't handle...
Disclosure: SAP paid most of my T/E for attending SAP HANA 2014, which is a joint conference from SAP and SAPinsider. SAP is a diginomica partner as of this writing. Diginomica does not have a financial relationship with Dolphin or Teklink - we produced these videos on our own time/cost.