Thread regarding SAS Institute layoffs

"...Unfortunately, this type of reaction is probably a core 'feature' of SAS 'culture'. Its probably one of my biggest professional challenges to not succumb and become a slave to it. Close second is building up the courage to watch a corporate training video...."

The comment you responded to was in reference to how SAS Customers feel about SAS BECAUSE of a lack of responsiveness. It wasn't about SAS culture. I'll leave it at that.

@4vao+1nHNoFH4 "At some point, people see they are not being listened to or taken seriously, and they give up and move on, never to return. This is why SAS can't be saved."

Unfortunately, this type of reaction is probably a core 'feature' of SAS 'culture'. Its probably one of my biggest professional challenges to not succumb and become a slave to it. Close second is building up the courage to watch a corporate training video.

Well this is one (thankfully retired) SAS R&D Gray Beard who agrees with @4jjw+1nHNoFH4. Believe me, some of tried to keep things consistent, stay focused on basic R&D to build on SAS’ core strengths, etc. By the late 90’s we were often countermanded my micro-management, internal apoplectic hotheads, the corporate politically motivated and the proliferation of silly products.

Ineffective management has taken three decades to foment.

@4vao+1nHNoFH4

"Coming to SAS after 20 years of C, C++, Java, etc that inconsistency drove me nuts. But it didn't seem to keep the users away."

Yes, you are correct. SAS failed to look at the products holistically and SAS failed to develop any type of internal or UI consistency. Big egos at work and all that. The one thing you will not hear from any of the SAS Grey Beards on this board is "You're right. Oops! We screwed up." Instead, you will hear defenses, rationalizations, and justifications for why things are inconsistent.

From what I understand, these inconsistencies also drove the common users nuts and started to keep them away. Do you remember the internal campaign to make things "Look, act, and feel like a suite"? That statement must have only applied to the UI layer, although it should have applied throughout the tech stack.

The defenses, rationalizations, justifications for inconsistency, and other foul ups only work for so long. At some point, people see they are not being listened to or taken seriously, and they give up and move on, never to return. This is why SAS can't be saved.

“
DS2 was specifically designed for better integration with SQL + databases in an OO language paradigm that supports threading."

Again, the majority of customers aren't programmers, and don't understand (or maybe don't want to understand) OOP.
“

I think that very much depends on how you understand customers across the history of SAS . I agree that the number of highly skilled programmers using SAS pales in comparison to languages like Python, Java, C++, etc..

Professionals develop necessary skills over time. Not have a computer science degree, or not otherwise learning programming in a more formal way, does not mean that people cannot pick it up as a matter of course, and become quite competent. This is true of SAS users, and also many programmers over the history of SAS R&D who did not possess computer science degrees, yet solved complex problems with beautiful code because they possessed the basic intelligence and drive to learn.

It’s very much a matter of perspective. How long have you been involved with the SAS user base and has your primary focus been traditional Base SAS users OR users of more modern tools like Visual Analytics OR the more vertical products like Fraud and Risk?

Again, they were very good reasons for developing DS2, 20-25 years ago like there was a solid reason for developing CASL 10 years ago. Neither are intended to be replacements for the original data step nor procs which I believe are supported natively in Viya via the Compute Server.

That said, likely the biggest faux pas SAS committed WRT our user base, was not continuing to develop Enterprise Guide as a desktop native OS application. AFAICT SAS Studio has still not caught up to what EG does. I was involved in a very large Viya POC where this came up frequently.

@3zbe+1nHNoFH4 Yes, sorry. I edited too many times and messed up the wording in the end.

"Again, the majority of customers aren't programmers,"

The ones who care about consistency of language syntax would be, though, right?

Of course that's never been a SAS strength - even the oldest of the PROCS have very different syntax amongst themselves (vars vs tables, options in parens vs. not, etc.)

Coming to SAS after 20 years of C, C++, Java, etc that inconsistency drove me nuts.

But it didn't seem to keep the users away.

"DS2 was specifically designed for better integration with SQL + databases in an OO language paradigm that supports threading."

Again, the majority of customers aren't programmers, and don't understand (or maybe don't want to understand) OOP.

"Given all this negative sentiment, what accounts for the announcement of an 8% increase in Total Operating Revenue through the first 6 months of 2023?"

Don't believe everything you read.

Especially when it comes to "the numbers" at SAS.

@3rni+1nHNoFH4 Do you mean hard to imagine any scenario that WILL be good?

“If you no longer work there, what difference does it make to you whether it sinks or floats?”

It hurts, because we worked hard, and trusted our managers, and hoped to create products of lasting value.

It’s not that we hope things will change. It’s more like a train wreck that we can’t help watching.

If we hope anything, it's that the train wreck happens in slow motion, to minimize damage to the friends we left behind.

I am agreeing with @3bid+1nHNoFH4; I am just phrasing things differently. When I was a graduate student many decades ago, it seemed to me that a few of my oldest and most distinguished professors had some pretty crazy ideas about many things including teaching graduate students. In my mind, I coined this phenomenon “old-professor syndrome.” Many decades later, I learned about “Nobelitus” (you can Google it), the idea that some laureates embrace strange ideas in later life. I see Nobelitus as the same as my old-professor syndrome. Someone who is both smart and exceptionally successful in life, someone who does not have to answer to others, eventually starts to believe that every idea he/she has must be brilliant, and any dissent, anyone who is not a sycophant, is an id--t. I believe the SAS CEO suffers from this (as I believe many other posters in this forum do even though we choose our words differently). Look at some of the R&D EVPs he promoted—in particular the ones whose first names began with vowels. The first two were sycophants who had no business at that level of leadership. Who in the company besides the CEO thought those appointments made any sense? The third EVP in that series, the forester, while brilliant, had the emotional maturity of a three year old. So again, he should never have been put in that kind of position. The CEO created a great company, one with a “golden age,” but he has no one to blame but himself for its subsequent downfall. Is it too late? Sadly, probably yes. It pains me to say that because I loved my career at SAS. Say there is an IPO (although I do not believe there will be one). Say he wants to sells less than 50% of the stock. Who would buy any of the stock with the current octogenarian CEO running the company? Which I still think means that some day the entire company will be sold. Sadly, it is hard to imagine any scenario that will not be good for current employees.

Why do people hold out hope for this company? If you no longer work there, what difference does it make to you whether it sinks or floats?

Assuming some hardworking people are going to save it, "if only the right leadership were in place" is some sort of hero-worship myth. If SAS were to return to profitability, do you think the culture would return to 35-hour work weeks, 2-hour power lunches, afternoon tennis, and 30 years of loyal employment?

I hope to be taller, thinner, and younger, too. But none of those are happening, and SAS will not return to any growth position. Wishing it were so is a foolish child's game.

“
Then you decide to create a new language with a new syntax - DS2. How well was that adopted?
“
DS2 was specifically designed for better integration with SQL + databases in an OO language paradigm that supports threading. All things considered it’s actually very innovative and accomplishes things that the traditional data step cannot do. DS2 has a decent size, loyal user base.

Traditional MVA SAS was never able to exploit parallelism very well — and several attempts were tried. The Threaded Kernel (TK) Is an evolution in the core SAS architecture that enables parallelized/concurrent programming. DS2 is built upon this TK. The traditional data step is primarily MVA-based Although by order of JG was forced into CAS, where it is a less than ideal fit And for certain use cases is actually slower running on parallelized threads then the same code/data running sequentially in 9.4. Such are the pitfalls of parallel programming for parallelism’s sake alone.

If management believed 8% increase in Total Operating Revenue was sustainable, they would not be doing layoffs.

When I worked at SAS, there were people who could, and did, create new products that were profitable.

Some of these people are still there. As @llv+1nHNoFH4 said, "We are still populated with talented and motivated people who want the company to succeed and are willing to do the work. With the right leadership and appropriate staff, the company has a profitable future."

It's that last sentence that contains all the trouble :-(

The Wikipedia entry for Viya references the SAS Help page: "SAS sells Viya alongside SAS 9.4, and have not positioned it as a replacement for SAS 9.4."

The Help page explains that "The CAS language is very similar to the SAS language." :-)

https://en.wikipedia.org/wiki/SAS_(software) https://documentation.sas.com/doc/en/vdmmlcdc/8.1/whatsdiff/n0m306a3p9spz7n1ew3j9scjiqjf.htm

Given all this negative sentiment, what accounts for the announcement of an 8% increase in Total Operating Revenue through the first 6 months of 2023? It seems like a good number of customers are still satisfied?

@3kak+1nHNoFH4 "How many other unique, niche vanity products are in the SAS portfolio?"

Last time I looked, SAS had 148 products on our website. Many product pages looked abandoned, which isn't a good look.

The previous comment describes the creation of new languages and syntaxes that are inconsistent with previous versions and what users already know. This is a symptom of a larger theme describing why SAS is in trouble. How many other unique, niche vanity products are in the SAS portfolio?

"I am [fill in the blank], the great SAS innovator! I will create a new [product, language, whatever] in my own image, and it's bound to be great! Everyone will want it, because I'm omnipotent and know what's best for all!" [Then said new technology fails spectacularly in the market].

If you want to understand the madness of SAS, read "The Power Broker: Robert Moses and the Fall of New York" by Robert Caro. Same dynamic, same character, same outcome.

"Most of the customers are not pleased with the CAS action syntax. It is highly inconsistent and does not following the SAS 9 PROC syntax that our customers are used to."

You have tons of customers that know DATA step, PROC, and MACRO. Most of them are non-programmers as stated before, but they're getting their work done.

Then you decide to create a new language with a new syntax - DS2. How well was that adopted?

Then you decide to create a new language with a new syntax - CAS Actions. How well was that adopted?

They don't know who their paying customers are, and spent tons of money developing languages for the mythical, magical new customers that never materialized to any great degree (probably because they only try to sell to existing customers instead of finding new ones).

"The glory days are long over at SAS."

The "Old SAS" has left the campus and is not returning. Don't wait for it. If that's your reason for staying then you risk getting burned.

No. Dr G should have sold a couple years ago when he had the chance. He was greedy and foolish for believing his company was worth more.
It’s his hubris and poor leadership the last 15+ years that put the company in its current state. It is sad because I used to view SAS as a wonderful, quirky company unlike any in the industry (worts and all). In hindsight, the decades of not nurturing strong executive leadership and surrounding himself with sycophants, are the root cause of his own making. After not having a real successor ready to take over when he was ready to retire, he started running through a series of potential successors. The best hope for it was Oliver, like him or loath him. Now, you have a feeble octogenarian trying to just keep the lights on until a sale or IPO (I don’t think it will actually happen).
After the botched sale, I decided to leave the company for a competitor because the IPO was clearly a knee jerk reaction to the sale making news. And while my new company is not perfect, it has competent leadership and coherent strategy. The glory days are long over at SAS. All companies are flawed, and some have proven successful at reinventing themselves. But it requires new executive leadership and competent mid level management to do that. SAS hasn’t cultivated either. So, the best you can hope for is a sale/IPO forces the new owner/board to clean house and make significant changes. Only then can SAS have a chance. But that’s easily 5 years away. If it happens, it might be a worthwhile company to work for again in the future… only time will tell. Good luck to all of you who are still there.

"...For me, it always comes back to SAS dropping the ball with universities..."

This frequent comment puzzles me. How did SAS drop the ball with universities? I thought they gave the software away free to universities. How's that dropping the ball? People do have free choice, and they simply no longer choose to use SAS.

Lets use marriage as an analogy. Imagine you married someone that looked like the ideal partner on paper. After a few years of marriage you find that they have some really bad habits that are negative and costly for you. You ask them to adapt and change, but they won't. Instead, they give you lip service and tiny feature adjustments in an effort to placate you. But there are no substantial changes made to meet your needs. Do you stay in that marriage and continue to get strung along in perpetuity? Only if you need to.

In my opinion, this marriage analogy is one of the reasons why SAS is failing. Installations and upgrades were lengthy and problematic. Customer issues, at least for the products I supported, were not being resolved sufficiently. There were lots of efforts to cook up new grand ideas, but no substantial change on things that matter. It was more of the same, just like a bad marriage.

Once other potential partners became available, the divorces increased. These new partners are more attractive, more responsive, and less costly to maintain than a continued marriage with SAS. At this point in these relationships, SAS has a negative reputation, and no amount of educational outreach would have resolved that. None. As I stated earlier, people have the freedom of choice. They are no longer choosing SAS.

CAS, Viya, whatever new scheme is being cooked up, are simply examples to illustrate the dynamic I'm talking about. In my opinion, this is more of the same grandiose drama without substantive changes. People are wise to it, and moving on.

Re "...the company receives an end-to-end software solution that is professionally supported and documented...."

I have heard some 1st hand stories from tech support types about how difficult their jobs are when it comes to attempting to support large corporate customers. Even getting logs from them is a difficult process, so I can see why a customer would feel compelled to seek answerers & solutions outside two bureaucratic systems.

"...the company receives an end-to-end software solution that is professionally supported and documented...."

Har-de-har-har-har! Oh, that's rich!

@2krm+1nHNoFH4 You raise a good point about our world-class department on technical support. This used to invaluable to our customers but there has been a paradigm shift which makes tech support no longer an appealing factor for customers to stay with SAS.

1. The open source community has grown in leaps and bounds and there is tremendous amount of online groups and forums available to get quick questions answered and lookup historical posts for solutions.

2. The customer horizon is no longer 10 to 20 years to keep their code running and maintained. The customers are changing technology constantly and they are fine if their code is good for 4 to 5 years. The current customers are more agile and adaptable and they are expecting to migrate to new and modern software every few years.

@2kmg+1nHNoFH4 "if we have to retrain our staff, why not just hire employees and new grads with R and Python experience and use the open source, and just skip the whole SAS train."

For me, it always comes back to SAS dropping the ball with universities.

The CAS usage is not as high in practice like it is reported. SAS has stopped selling SAS 9/* products but rather they sell Viya products only but they bundle the SAS/* products for "free". Lot of customers have not migrated to Viya but due to "clever" marketing, it appears like more customers are using Viya based on the licenses sold as shown on the books. They were sold Viya licenses but continue to use the SAS 9 products.

Most of the customers are not pleased with the CAS action syntax. It is highly inconsistent and does not following the SAS 9 PROC syntax that our customers are used to. So, they have to retrain their staff to rewrite the SAS 9 programs with the CAS actions. Plus there is no parity with SAS 9 functionality preventing them from doing the migration.

In fact, customers are rethinking if we have to retrain our staff, why not just hire employees and new grads with R and Python experience and use the open source, and just skip the whole SAS train. No new analytic functionality is being added to the SAS 9 PROCs since they are in maintenance mode; we are basically giving our customers a way out of SAS.

This whole layoff makes me sad, particularly the way it was handled. At least the Viya comments are educating me.

Viya's performance is apparently no longer superior. And SAS is unlikely to "innovate new analytical methods" when most of the statisticians took the buyouts.

So I'll ask: if those were its two main selling points, is Viya still viable? Does it still make strategic sense to devote resources to the MVA->Viya conversion?

"The idea at the time was that SAS provided and would continue to innovate new analytical methods, not available in open source OR for dataset sizes/performance requirements that R, Python, etc. could not handle... Advances in the Python ecosystem, other open source data technologies (recently DuckDB) and the proliferation of powerful laptops that can run analytics against datasets sufficient for most users erodes the need for Viya/CAS."

"Whereas if the company chooses to pay for SAS, the company receives an end-to-end software solution that is professionally supported and documented."

This was true for most of SAS's history. Unfortunately, it is no longer true. With the waves of retirements over the past 5 years--some voluntary, some not--many parts of SAS no longer have technical support. Everyone familiar with certain aspects of my work have retired. Users only recourse is to post their questions to SAS communities. Unfortunately this is true in many areas not just mine. Thinking about it makes me sad.

One point that has not been raised in discussions is the technical support that is provided by SAS to customers at no additional cost. Unlike most companies, SAS Technical Support is a career-track organization instead of just a high-turnover entry-department. In our umbrella for example, the median tenure of a support engineer is over 20 years experience.
Yes, a company can choose to use free software to accomplish a specific analytical task. The company has to consider what happens when the person who wrote their code leaves. Whereas if the company chooses to pay for SAS, the company receives an end-to-end software solution that is professionally supported and documented.
Might that point serve to offset some of the leadership mistakes, and thus keep SAS viable?

@2nbn+1nHNoFH4 "So many things to comment on, but I'll limit it to this.."

I for one am interested in hearing your comments on the other things.

"SAS's customers don't want to use open source frameworks and languages to interact with SAS"

They want/wanted SAS to continue developing EG as a Windows desktop app (MacOS would be great too) with native Viya capabilities.

“
Court a Python developer who is not a SAS customer and tell them, "You can pays us a bundle of money and then run PROC REG from Python". They'll look at you like you have 7 heads (that's being kind) and say, "Why? There are free Python packages that can do everything the PROC REG can do, and more".
“

Open Source Client language support in Viya was intended to allow existing and new users to invoke CAS actions against datasets ( i.e. distributed across multiple CAS Worker nodes) too big for SAS9 scenarios OR where Viya/CAS runs as a cloud service such that users pay fractionally for invoking specific CAS actions. The idea at the time was that SAS provided and would continue to innovate new analytical methods, not available in open source OR for dataset sizes/performance requirements that R, Python, etc. could not handle.

These considerations along with the data pipeline/flows flexibility of interacting between Python and elastically scalable CAS services was thought to be one of the stronger Viya value propositions. In some cases it has been a win for customer, though “hindsight is always 20/20” applies. Advances in the Python ecosystem, other open source data technologies (recently DuckDB) and the proliferation of powerful laptops that can run analytics against datasets sufficient for most users erodes the need for Viya/CAS.

@gph+1nHNoFH4 is spot on.

So many things to comment on, but I'll limit it to this:

"SAS's customers don't want to use open source frameworks and languages to interact with SAS"

That should have been apparent to anyone who has interacted with core customers. Most of them are not programmers. They are researchers, scientists, retail and manufacturing people, and so on. They are not interested in being programmers so they are not interested in open source (unless it's threatening their jobs if they don't embrace it). I see some embracing it but it is because they want to be able to get another job after their organization pulls the plug on SAS (or attempts to pull the plug).

I never understood the selling point of SAS + open source. Who is the customer and what is the selling point?

Court a Python developer who is not a SAS customer and tell them, "You can pays us a bundle of money and then run PROC REG from Python". They'll look at you like you have 7 heads (that's being kind) and say, "Why? There are free Python packages that can do everything the PROC REG can do, and more".

"With the right leadership and appropriate staff, the company has a profitable future."

Having the right leadership has been a perennial problem at SAS.

The hapless clowns at the top are not examples of the right leadership.

“ But then Dr. Goodnight promoted him to COO …”

By that point, JG was looking to offload the parts of SAS he did not want to directly deal with at the executive level — sales was probably at the top of his list!

OS did not want to relinquish being CTO. He was already being groomed to be second in command (learned this from a low-visibility yet high authority) and eventually take over. Think about how many ways that could’ve gone if JG brought another executive in to be COO who was highly successful in the role. My understanding is OS took the COO position on the condition that he would remain CTO in a dual CTO/COO role. Most of the fun was in the CTO role, yet there power perks were in the more boring COO administrative executive chair. Doing both was too much responsibility for any individual — but JG loves the bargain. Two executives for the price of one!

Being CTO for a company with a tech stack as complex and diverse as SAS is an all-consuming 60-70 hour/week job if it’s done right … and that’s IDEALLY with a trusted direct reporting group of master architects, mathematicians and computer scientists who are world-class experts in areas the CTO is/cannot be deep in. As CTO, OS managed R&D division heads and many fell short of this ideal. Add the dual COO role and his effectiveness of CTO had to have gotten diminished.

This is very sad to me because I agree with the comment that OS was SAS’ best hope to superintend efforts in evolving Viya to successfully replace SAS9.

We failed to take open-source seriously early on. SAS could've helped define defacto standards and been/be a major contributor, shaping a new business model by doing so.

10-15 years ago JG and the powers that be were not interested because they didn't really believe that open source could ever catch SAS. We were riding tall, Fortune #1 best company to work for two years running. 37% share of the advanced analytics market -- more than our next six competitors combined. What could possibly go wrong at SAS?

We were still hiring some of the best and brightest PhD/postdoc scientific math types in 2005. Then the big cloud players figured out that highly integrated big data and analytics was one of the next big plays. And most of them already had corporate campuses and perks that rivaled or exceeded SAS'. Just add rapidly appreciating stock and higher salaries --> Shazam Shazam, I'm moving to Mountain View instead of Cary as soon as the ink on my PhD diploma certificate is dry.

Meanwhile, back at the plantation, the troops are getting complacent -- settled into the din of corporate perks, and advertised 35 hour work week. Many of the longer term, highly skilled/experienced Developers were working well beyond that 35 and suffering burnout from a decade of cranking out version 9 releases on top of a progressively more unwieldy infrastructure Stack (not only in the build/test system but the product itself -- Java versions, various Web server, implementations, the SAS Java versions, various Web server, implementations, the SAS Metadata Server and mid-tier. Pretty much a house of cards that was temperamental to deploy and very hard to diagnose customer problems in.

It is difficult, but not impossible, to compete with open source. We did not, because we:
> Dropped the ball with education.
> Failed to innovate.
> Failed to penetrate vertical markets.

This did not have to end so soon.

@1dth+1nHNoFH4

"The grassroots appeal to hardcore scientists was what made it into a big success. That generation of folks has since retired or moved on."

This reality could have been delayed if SAS didn't drop the ball with education. Regardless, opensource would eventually challenge and disrupt the SAS business model.

And for all the new employees and low maturity Sr. Director(s):

"All the king's horses, and all the king's men could not put Humpty Dumpty back together again."