Thread regarding SAS Institute layoffs

"JHG wouldn't listen. He's the smartest guy in the room. That's why he's in this mess."
"I'm the billionaire here, not you."

Yes, and hunger and debt are great motivators. With pockets of your workforce still hungry and in debt, they seek to hunt. Allow them the opportunity to do so and share the spoils with the tribe.

"JHG wouldn't listen. He's the smartest guy in the room. That's why he's in this mess."

"I'm the billionaire here, not you."

"I’m in for a 1000 pounds if JGH is interested in forming an advisory council made up of now retired members of the SAS R&D brain trust"

Does that involve investing your own personal fortune in the business?

I always wished that the King would have walked disguised among his subjects and listened to their opinions, bypassing his Generals and others who would tell him what they thought he wanted to hear.

At some point, each new initiative took on a status of emergency and the pace of new initiatives grew rapidly. Each big change in direction would be accompanied by new management and reorganization. leaving work on the last big thing unfinished and unpolished. The frequent pivoting left so many things abandoned, produced vast waste, and created a good deal of confusion for the customers.

JHG wouldn't listen. He's the smartest guy in the room. That's why he's in this mess.

… unfortunately, he’s already “wasted” several billion

JHG won't waste money -- he'll just read this site :-)

… I’m in for a 1000 pounds if JGH is interested in forming an advisory council made up of now retired members of the SAS R&D brain trust.

"...Maybe the plan is to keep digging until he reaches the Beijing R&D Center...."

In for a penny, in for a pound as it goes at SAS. In this case it would be in for the Yuan.
Throwing good money after bad is another SAS-ism.

"And why that a layoff in that area [Viya] has NOT already happened is probably why many do not believe JG is serious about selling."

JHG has said something to the effect that when you've dug yourself into a hole, you need to know when to stop digging and get out.

Maybe the plan is to keep digging until he reaches the Beijing R&D Center.

How do you feel about adopting SAP's HTML5 version? That also was an exercise in "we have to make it here"

And DOJO HTML before that, and WebAF before that. All part of the "made at SAS" heritage.

Largely the same development teams, too, if IIRC.

"Attempting to impose heavy O-O semantics on C Via a homegrown system of macros, V-tables and pointer swizzling feels more like an academic exercise."

Reminds me of making AF/SCL O-O. What customer wanted that? I think that it's fair to say that current customers at that time that understood/used O-O numbered close to 0%.

“Perhaps salvage the Viya infrastructure and modernize Risk and Fraud as @4dat+1pH4BhK3 suggests.”

Given the symbiosis between the SAS business model and its technology, it is my personal conviction that this is the way forward for SAS — transitioning from being a classical analytics toolset/language vendor to an ultra modern ML/AI/advanced analytics-based enterprise software company. This should include integrating open source where it makes sense.

 @5rul+1pH4BhK3

That’s a clear answer. Thanks very much.

Viya is not selling to open-source users. If it can’t be made compatible for SAS users, the only choice is to cancel it, and move the staff onto more profitable products.

Perhaps salvage the Viya infrastructure and modernize Risk and Fraud as @4dat+1pH4BhK3 suggests.

@4hnp+1pH4BhK3

There are many complex organizational and technical reasons for why Viya is not fully compatible with all Version 9 use cases. These have been covered ad nauseam over various comments (sometimes not really matching the title of the thread) on several threads on this website.

Honestly, to attempt to achieve the goals of Viya while making it fully backward compatible with SAS9 Foundation along with other critical products would take 50+ hour weeks from the best developers in all of SAS history working for 8 to 10 years. At this juncture and time, no sane engineer is going to do that without a clear mandate and support from executive management along with a a huge raise and a sh-t ton of equity.

“It’s fair to say JG has a deep personal identity with the SAS language and architectural heritage.”

Thanks so much for the history. I know that the above statement is true. So why isn’t Viya compatible with SAS?

@4dax+1pH4BhK3

It’s fair to say JG has a deep personal identity with the SAS language and architectural heritage. LASR ( genesis ~2008) is a direct result of major SAS customer requirements pre-dating open source software that was up to the task. Although R had gained some ground in the single user analytics space, the Python ecosystem was nascent at best, Spark did not yet exist, and the major cloud players were only beginning to provide significant analytics products. In the whirlwind decade that followed, the overall analytics market rapidly shifted and changed. This included the birth of the “data science” job title and many new programs to mint such professionals.

So yes, coulda, woulda, shoulda, … yet the dynamics at play around 2008-2010, especially the founder’s radical commitment to proprietary software (SAS at that point was still recognized as the overwhelming leader in advanced analytics, based on “the family jewels” of proprietary, internal algorithms) with Oliver as his protégé, it should come as no surprise that JG would continue in this modus operandi.

Again, it’s easy to look back and say in retrospect “what if SAS had done this?”, yet circa 2008, company innovation culture and momentum was in conflict with the still unproven and very nascent open source, analytics/distributed, data management, etc. worlds. How many job listings existed for these things in 2008? SAS needed to move its architecture forward and building it ourselves is what we had always done. LASR was built by Oliver’s small team and its evolution CAS began about five years later with his larger team, which eventually became a division.

...while showing respect to SAS' heritage..."
Why?
Wouldn't it have been better if Viya had been developed completely indepedently of the SAS baggage like PROCs and Data Steps which may well be loved by hardcore SAS coders, but which are regarded as archaic and unintuitive by everybody else?

Viya needed to attract the new bread of analysts and data scientists in order to remain relevant and grow. Leave the loyal but aging hardcore SAS coders to continue doing what they do with SAS9, and build Viya as a completely unrelated and indepedent product line for the next generation.

If Viya had adopted open source for the language and distributed architecture, that would have freed up resources to focus on two additional layers that open source doesn't do well and SAS has expertise in...governance (model management, deployment, auditability, model explainability, etc), and industry vertical analytical apps on top of that.

If SAS had gone down that path back when LASR was first conceived, imagine how different things could be now.

@4dgy+1pH4BhK3

I think there’s a need to separate Viya as a product from as the Viya infrastructure. Viya as a product (actually it’s several products built on said infrastructure) cannot replace V9 and was never intended to. SAS has been in a multi-year regress due to the great Covid tech upheaval, the departure of Oliver and ~200 of SAS’ best R&D engineers (most of whom worked on Viya infrastructure for the previous 5 to 7 years) and most certainly the decline in relevance of V9.

With the right strategy and people, SAS could Innovate new ML/AI analytics-based enterprise solutions (starting with a major refresh of Risk and Fraud) on Viya infrastructure (Including some clever extensions to what is already there). It comes down to SAS executives having the will to reduce current staff who don’t have the skills to be part of this and paying to hire world-class engineers who do. Of course all of this is a big gamble and that’s probably why it’s not being done, or at least not more expeditiously.

"If moving forward with Viya is not the solution then what is?"

If there was a market for Viya, the market would already be established. SAS is giving Viya away for "evaluation". Little or no traction there.

Most V9 customers have little or no interest in Viya. They are working harder to get off SAS than continue with SAS post V9.

The scarcity of job postings seeking Viya candidates shines a dim light on the future of SAS post V9. So what is the future of SAS?
Pure speculation of course, but there are indications that the heirs of JG and JS will be the ones deciding the future of SAS.

@4sgf+1pH4BhK3

The problem with this logic is that V9 is a stale architecture that is not strategically optimal for moving the company forward. As explained many times previous, the basic Viya infrastructure: CAS + Go Microservices + Compute Server was the best effort the SAS brain trust could come up with could pass JG's capital expenditure/technical vision constraints while showing some respect to SAS' heritage (Data Step in CAS and/or Computer Server, PROCs that invoke CAS actions "under the covers", etc.).

Operating within the boundaries set around V9's replacement was an incredible hard task and one that took many R&D veterans everything they had to attempt. It's easy to "arm chair" QB now, but just try designing a modern replacement for V9 that deploys natively in the cloud while providing all the on-premise and other flexibility required by SAS customers.

As far as "V9 being the one making money goes" ... that revenue is in consistent decline with no sound way to modernize the software to reverse the trend. If moving forward with Viya is not the solution then what is? Build yet another new platform (even one based mostly on Open Source)? That does not seem realistic given JG's age, the competition, etc.

"The most recent major decision is to continue supporting both V9 and Viya. But SAS doesn’t have staff to support both. If they did, they would not be laying off."

One makes money and the other does not. Seems like that simplifies which one should be targeted for an upcoming layoff. And why that a layoff in that area has NOT already happened is probably why many do not believe JG is serious about selling.

It sounds like the entire company was a cascading series of thwarted plans and ignored advice. The irony of all ironies, is what appears to be the lack of economic and systems engineering analyses coming from a company that sells analysis tools.

Flash was a sound technical and business decision. At the time, HTML5 was immature, so many companies chose Flash. Once HTML5 matured, first Apple and then other companies abandoned Flash.

Our mistake was not in choosing Flash, but in clinging to it. The Photographer urged SAS to switch to HTML5, but he could not execute that switch. Perhaps he was not allowed, due to short-term business concerns.

If so, that would be consistent with the SAP decision. We had short-term business concerns, because of our partnership with SAP. But even from a pure business perspective, that was a terrible decision. Anyone technically competent could see that SAP’s code was going obsolete, so that our costs would outweigh our benefits.

The Forester aka Big German gets blamed a lot, because his decisions were recent, and perhaps the most costly. But between Version 7 and today were a series of poor decisions (Web Report Studio, anyone?). The Forester’s decisions were just penultimate in the series.

The most recent major decision is to continue supporting both V9 and Viya. But SAS doesn’t have staff to support both. If they did, they would not be laying off.

@4iie+1pH4BhK3

Another thread mentioned that was a business not technical decision to adopt the SAP framework. This again points to sub-optimal/poor executive management and a culture where a highly functioning CTO advised by veteran R&D architects is not the final say on such decisions.

How do you feel about adopting SAP's HTML5 version? That also was an exercise in "we have to make it here" efforts, right on the heels of transitioning all the products to Flash. Why Flash and not HTML 5 to begin with?

YEARS wasted. Loads of opportunity to improve products in the transition wasted. If that were a corporate sabotage effort against a competitor, SAP could not have played it better! Hook, line, sinker, rod, reel, and boat!

@4uhq+1pH4BhK3

I remember when all that was going down. About 1995 sound right? Yeah, it's coming back to me now.

At that juncture in time, I was involved in the design and construction of some lower level XOB-related services, that could also be invoked from C in the X and Y layers. A certain group of assertive developers insisted on their vision of XOB and it seemed to me the internal implementation was a constant moving target.

IIRC, the justification for not beginning to adopt C++ was that it had not matured enough to build SAS on all supported host targets. That's fair, but I'm not sure what XOB really provided other than maybe making the ODS APIs "friendlier"?
Attempting to impose heavy O-O semantics on C Via a homegrown system of macros, V-tables and pointer swizzling feels more like an academic exercise.

Imagine had SAS had the insight to get involved in the early formation of the C++ standard? We could've had influence and gained super valuable insight without having to build so many homegrown, workarounds like TK they come with a heavy price in terms of ongoing maintenance and the inability to attract top talent because nobody wants to waste their career on something for priory and now 25 years old.

I believe we were involved in ANSI C, given SAS' own C compiler Development heritage, which was likely another reason why there was little push to get to C++. Those were turbulent times, riding high on the fast growing adoption of MVA version 6, SAS' reputation as THE tech workplace of that era, Goodnight and Sall becoming billionaires, etc.

From my vantage point, there's a real trade off here because at that point (mid 90s) the micromanaged doofus R&D culture did not exist and developers were free to actually do research, intense first principles tech learning and innovation. This phenomena is one reason why SAS continued to grow so rapidly all through the 90s.

Unfortunately, the vision, maturity and effectiveness of management did not keep up with the SAS' revenue growth and technical prowess of that juncture in time.

There are people at SAS that despise this site but its refreshing that its allowed to continue without disruption. If you look at Glassdoor or Indeed most of the reviews are positive and I'm sure SAS makes an effort to keep it that way. It's difficult to post here because like the OP said, a lot of us spent many years of our career at SAS and now depend on that experience to get new jobs and promote ourselves in the workforce. I had a lot of great times at the Cary HQ - from getting an office in Q when it first opened to participating in the filming of Iron Man 3 and getting to see Robert Downy and Gweneth. There was such opportunity there and it was wasted on ego and the inability to see the writing on the wall. So I'm bitter.. but it's not because I'm rooting against SAS and the people there... it's because of the virus of incompetence and self-importance that have taken over the place.

@4lvn+1pH4BhK3

At my level, the O-O language was XOB, though I don’t doubt there were other mistakes at other levels.

A problem like Viya is not unprecedented, and it has a precedented solution. Why they don't apply that is a mystery.

@3cjh+1pH4BhK3

Do you include XOB, TK, the SAS Metadata Object Model, IOM (among other things a wrapper for COM, DCOM and CORBA) in your observations about homegrown object models? Just curious which object model you’re referring to WRT the performance penalty we suffered. Version 7 is a blur to me now.

By 2000 R&D should have been moving towards C++ and using host-specific assembly when absolutely necessary. Sadly too many Director level brain farts had already occurred.

@3ehq+1pH4BhK3

Near the top of your list should be SAS Version 7, based on our home-grown object-oriented language. At that time, the "smartest guy in the room" thought that no commercial language met our needs.

So we rewrote vast amounts of code -- but we were unable to optimize our new language. The result was that Version 7 ran like a snail in January, and no one wanted to buy it.

We never charged for Version 7. We moved our customers, as quickly as we could, on to Version 8. There's a precedent, if anyone wants to apply it :-)

@3erk+1pH4BhK3

In the second instance you must be referring to the hot-headed (now deceased— RIP) former head of R&D who was ceremoniously escorted off campus by an Executive Protection Agent after going apoplectic on his admin ~12 years ago. He returned to his country of origin, only to be later rehired by SAS in some kind of middle eastern role.

"...SAS managers tend to stay at their management positions for a very long time, almost forever. They don't ever seem to step down or step aside from their management positions even when they are not good or capable of handling their jobs...."

Then they hire loads of young Interns who don't have the experience or maturity to discern how bad the manager truly is. The bad manager serves as a poor example of "professionalism" for these employees, and the cycle of corporate dysfunction continues.

@3obf+1pH4BhK3 "...In some ways, it's probably easier to fail someone upward than it is to deal with their craziness."

Honestly, one effective way to deal with a bad manager is to do everything in your power to get that manager promoted. For one, the manager notices your efforts. Second, your new manager adds a buffer of protection against you and the bad manager.

Previous post:
Revenue losses may be better attributed to the "smartest guy in the room" phenomena.
"I have an idea that will revolutionize the world! I'm going to develop it with JG's blessing, and little market research. It will be a success because I'm so smart and know what's best." (Product falls flat). Rinse and repeat.

The "smartest guy in the room", I asume you are referring to is the Big German. He deserves a big chunk of the blame.

He was arrogant, and totally out of touch. I was in pre-sales and I witnessed him literally storm out of the room in a customer meeting because he didn't like what they were telling him - i.e. that they were struggling with his product. He was incapable of handling any feedback. It was embarrassing. We laughed about him with the customer later on...they were genuinely surprised by how obviously ill equiped he was for the job.

"It will be a success because I'm so smart and know what's best. (Product falls flat)"

Viya has to be near or at the top of list for the most expensive product to fall flat.

@3use+1pH4BhK3

What are the top five products that fit your criteria?

The term "Corporate Conservatorship" is narrowly applied. While there may not be a loss of revenue associated with these cases, there are substantial costs due to wasted time and labor.

Revenue losses may be better attributed to the "smartest guy in the room" phenomena.
"I have an idea that will revolutionize the world! I'm going to develop it with JG's blessing, and little market research. It will be a success because I'm so smart and know what's best." (Product falls flat). Rinse and repeat.

What’s the correlation between “corporate conservatorships” in Director-to-VP positions and the loss of $millions in revenue due to their incompetence?

@3kub+1pH4BhK3

Multiply your experience by the total number of duplicitous, power-seeking Managers + poopy Directors combined with their micromanaging and the prohibition of basis research by the rank and file over the past 20-25 years. Result: irrelevant products and ultimately declining revenue — which could have been in large part rectified by firing (or demoting to IC in some cases) ~50 people AND building a better management culture at SAS.