Value, and not TCO is the key driver for analytics in the cloud

As analytics giant SAS treads a fine line in its cloud and edge analytics strategies, 2020 will be significant for market expansion, deep use cases, and breaking into new focus areas.

analytics binary code network digital transformation
Getty Images

SAS’s iconic founder, Dr. Jim Goodnight, and his successor, EVP-CTO-COO, Oliver Schabenberger, are both men of few words. However, during the AnalyticsX 2019, at Milan which is possibly one of their largest global annual analytics conferences, recently, Dr Goodnight and Schabenberger, took stage and clearly outlined the direction that the Analytics giant was going to take, some big announcements and some tough decisions that they may have to take over the next year.

Here are excerpts from brief in-person conversations with them, on the side-lines of the 3-day conference at Europe’s largest convention center, MiCo. 

SAS is among the few companies that really charted a succession path quite early in time – you are grooming your second line too. What are the innate challenges you face?

Goodnight: Yes, in fact as is well known, I am grooming Oliver (Schabenberger) for the role of CEO. When I made this decision, I was and am distinctly aware of the fact that he understands technology and is a master at that, but what we are working towards is honing the understanding of “business of technology” – the strategies that encompass taking SAS vision forward. We are also building a healthy line of data scientists. But, these are pretty much a part of any large organization’s succession and talent journey. 

It is felt that the cost of SAS Cloud Applications is too high for many markets. So, how do you evaluate this perception, and do you have specific action points on lowering the prices for cloud apps, especially given the shift in SAS’s stance towards Open Source?

Schabenberger: You cannot talk of pricing in the absence of value, while you are also assuming that SAS applications are expensive. The platform is about using analytics to solve real business problems primarily for large enterprise. We are asking ourselves constantly in this marketplace about the value that we bring to our customers. If you think about Open Source, if you see, our attitude has changed completely – and we define too that in terms of the innovation value it opens up and not TCO, which won’t necessarily be the key driver.

Jim Goodnight, Co-Founder, Chief Executive Officer, SAS SAS

The fear of data warfare is driven more by hyperbole than reason – it is more exaggerated than what is possibly true. The focus must shift and is decisively shifting to harnessing data for outcomes and innovation.
Jim Goodnight, Co-Founder, Chief Executive Officer, SAS

We want to be tightly integrated to the cloud ecosystem, and we view our tools such as open source model manager as one of the ways to bridge the gap between the data generated and the quality of data that can accelerate cloud migration.

So, what does rewriting code for the Cloud really imply from a SAS and SAS customer perspective?

Schebenberger: Hyper cloud is going to the operating model. There will be a certain percentage of apps that will move and some that will stay on-prem.

 When you can rewrite for cloud native apps, one of the ways in which this can happen, which our future Viya platform version 2020 is factoring for SAS customers is containerization. There are a lot of benefits that will accrue from those. Having said that, while the move to the cloud is inevitable, not everything needs to be rewritten. Customers have to set the right KPIs and performance improvements when they review their cloud analytics strategy, to orchestrate the right model, and get tangible results. 

How is “Analytics on the Edge” shaping up and what are its limitations?

Oliver Schabenberger, Executive Vice President, COO & CTO, SAS SAS

We don’t acquire much at SAS. We have been successful in building technology we need. And of course, we also have third party products embedded in our solutions. I place a lot of emphasis, in my role, on strong partnerships. I also see a new technology landscape – driven by cloud providers – as being very dominant.
Oliver Schabenberger, Executive Vice President, COO & CTO, SAS

Schabenberger: Analytics on the Edge has been a focus areas of our IoT effort. It is part of the evolution that wherever the data is, the analytics follows. Part of the problem is when we try to move the data to the analytics, it is expensive we create data repositories outside of the governing cycle. With the increase in the data generated by edge devices, the operating model of sending all data to a central place or analysing does not work in all cases especially where you need real time decision, it doesn’t. Edge analytics uses different forms of analytics such as ML, when data has to move fast.

SAS is training few models on the edge, which are more at the back end – part of the reason is training is difficult - you cannot do a logistic regression or revisit the same record twice. We are more looking at aggregation, filtering scoring or inference of models – with the current form of AI, this model is becoming big; these can pick up massive amounts of memory. This can be complicated and lots of effort since we have to compress the model without it losing its veracity or volume of the predictions.

We spend a lot of energy in binary representation, and we have a very efficient scoring technology. We are working with a number of partners to embed our technology into their devices – this is a big partner play for us; GE Transportation is a good case in point; they are leveraging our Events Stream Processing (ESP) engine through EdgeLINC.  

We have made some progress in training models live – these will manifest into game changers eventually, when we have edge devices that do not have to depend on the back end.

While SAS does have the enterprise base, there is a distinct shift to newer verticals. How is SAS devising its balance between “Lights on Business” vis-a-vis new focus areas from a  customer acquisition standpoint?

Schabenberger: In a sense, lights on technology is also new technology. At AnalyticsX, we have talked about Computer Vision and NLP. These have existed before, but what has changed are perhaps the use cases. From a solutions perspective, our platform or software do not solve specifics or operate from specific industry IPs. Therefore, the thumb rule is if we cannot answer five core focus domain questions through our solutions – that relating to fraud, risk, customer intelligence, IoT and retail, then we do not enter that industry. Importantly, in 4 out of 5 of these focus areas, we have lines of businesses that report directly into me - just to show the emphasis. 

Traditionally, while SAS has invested heavily in R&D, it has not been a very M&A driven company, no significant acquisitions having happened in the last few years. Is SAS contemplating strategic acquisitions to boost its Intelligent Automation mission?

Schabenberger: We don’t acquire much at SAS. We have been successful in building technology we need. And of course, we also have third party products embedded in our solutions. I place a lot of emphasis, in my role, on strong partnerships such as Intel, HP and Cisco – the hardware and network providers, among others.

I also see a new technology landscape – driven by the cloud providers – as being very dominant. Years ago, when we moved away from mainframes to Mini and Micro Computing, we build Multi-vendor architecture (MVA) – which gave us the flexibility to run any OS, with a portable layer of code, with 95% running anywhere. We had a small team of developers who were responsible for connection between the application code and the OS. Today, MVA has turned into MCA or Multi Cloud Architecture. What we are trying to address with the cloud providers is – to what extent we are agnostic and how we are integrating with their technologies. 

There is so much debate around data warfare in the 21st century. How true or deep will this phenomenon actually be?

Goodnight: The fear of data warfare is driven more by hyperbole than reason – it is more exaggerated than what is possibly true. With so much legislation on the prudence and governance of data coming in place across the world, we don’t have too much to fear.

As far as enterprise is concerned, if we choose to be secure, we choose to be so, and design data and our networks to be secured. The focus must shift and is decisively shifting to harnessing data for outcomes and innovation.

Copyright © 2019 IDG Communications, Inc.

Download: EMM vendor comparison chart 2019