Extracted from a thread began on @OP+1jgxdnf0h
207 replies (most recent on top)
Former direct report of OS. Accompanied him on various covert partner engagements as well. Oliver treated his inner circle very well, yet generally kept it small. He remained objective and would tell you straight to your face if he felt you were off. He was rarely one for small talk.
We had lots of fun/humorous moments on the CAS team, even sharing photos and anecdotes from personal and family events that occurred outside of work hours. Things moved at a very fast pace and in my experience, there was definitely a sense of needing to prove oneself regularly. The standards were high and f-ups required immediate remediation, and sometimes “natural consequences”, like a good old-fashioned public shaming.
I could definitely see some of the “bold and beautiful” people or those manifesting mediocrity as not having a strong like for OS. He could be very exacting, yet definitely energized those who came to get things done! Under Oliver, the CAS team felt like membership in something special, though he required we always be helpful to those outside the team working to learn CAS or build actions. He trusted his team members to attend in numerable meetings across R&D and Solutions as representatives of CAS in our common work.
a lot of the comments about "good" or "bad" leader are basically just along the lines of
- this person's introverted, prickly, bad social skills, strong technical fits my personal style
or
- this person's style doesn't fit my personal style.
good leaders have to have way more emotional intelligence and communication skills than that - part of the "diversity" part of "DEI" is about being able to lead all kinds of people different from you in many ways (it's not just about race or gender or age etc). sounds like this leadership wasn't able to do so. walking away from your own people trying to talk to you or not being able to talk to customers or yelling at them is a much more fundamental failure, though.
"As a 30 year veteran at SAS, I never heard anyone say anything good about Oliver's leadership style."
I heard (directly in 1 case, indirectly in several others) that people loved working for Oliver. My own experience (I was not one of the anointed top-tier guys tapped for CAS work) was that he could be extremely personable but had bouts of extreme prickliness (like the time someone marked one of his defects as a THUP (testing-holdup)). And he was an extremely hard worker capable of almost unbelievable amounts of productivity on hard problems (coding and documentation).
"We now talk to the suits, instead of to the people actually using the software."
The problem isn't mainly in talking to suits. It's a good idea to add that conversation. The people using the software are aging out (SUGI Sam has likely left the workforce) or being displaced by users of other software, and we rarely talked to the users of the rapidly emerging competition/replacements. That is/was the bigger problem. Perhaps lack of awareness of the need to even do so the even bigger problem. People say OS couldn't talk to customers, but correcting that (critical) fault alone wouldn't be enough. The innovator's dilemma gets you ironically because you're focused on own customers, and unaware of everything else happening.
"Others writing elsewhere have described SAS as “where innovation goes to die.” That’s an extreme statement, but I can’t disagree."
My belief is that there was building up to a point. After that point, it was destruction.
Of the companies acquired by SAS, what percentage of them did better AFTER acquisition? My theory is that they were acquired for their promising technology, and then hamstrung by JG through efforts to integrate past technologies. But I could be wrong.
My preceding post reflects what I saw over decades at SAS: an inability to get ideas accepted on their merits.
We had no shortage of good people, because for so long, we were one of the best local employers.
But good people would come, rise in the ranks for a few years, then see their ideas blocked by politics. Those who left in frustration, by definition, tended to be those talented enough to find better jobs.
Others writing elsewhere have described SAS as “where innovation goes to die.” That’s an extreme statement, but I can’t disagree.
As @nd+1jk94kx0x wrote, the politics grew worse, and the company less customer-oriented, in recent years. But these trends are not specific to Viya; they began decades ago.
‘Who were “they who were not supported?” What were their grand ideas?’
Most of the people I had in mind left SAS before Viya and its predecessors were conceived. So their ideas did not relate to this thread.
Several of these people left SAS in frustration over their inability to get their ideas supported. (Dr. Mindle, of the preceding post, may have been one.)
The only one who did not leave SAS did leave R&D. When he could not get his ideas supported, he put them into JMP.
As a 30 year veteran at SAS, I never heard anyone say anything good about Oliver's leadership style. No first hand experience with him so have no idea if that is spot on or not.
It is possible that everyone I heard that from is wrong. Not taking or making bets. Life at SAS was great for more years than it was not. Thank you Jim for the good years!
"SAS had a long history of “listening to users” via the annual “SUGI Sam” feature polls, with results announced at SUGI->SGF"
The death of SUGI, and the staggering gasp of SGF, and now eventually SAS Innovate, is the transformation of a more technical, user-focused conference to a marketing event.
We now talk to the suits, instead of to the people actually using the software.
I called it building "see". The sites were grand, but the pancakes sure tasted funny.
Thanks OP for the historical reflections, word of mouth is sending old timers like me here : )
SAS's failure to me sums up to the abolishment of customer obsession: 100% of SAS customers are in the Third World of AI/ML[1]. 100% of SAS customers want to migrate to a modern day future proof AI/ML/Analytics platform that's built upon best-of-class open source components. 100% of SAS customers' talents speaks mainly in Python.
Thinking from customer's shoes, there are still a lot SAS can do to survive and thrive:
- SAS are quite a few steps behind state of the art AI/ML, but SAS customers are miles behind. With feet already on the customer grounds, SAS may help them mature. "Let's learn and grow together!"
- SAS could invest in and own the emerging in-demand areas such as causal analytics [2].
- SAS may retreat and focus on providing Pythonian AI/ML/Analytics components in statistics, econometrics, forecasting, OR and quality control... Abandoning Data Step and Viya the platform all together.
But...
I don't know why I'm still so emotionally attached to SAS, maybe that's love : )
[1] https://www.drmindle.com/the-three-worlds-of-ai-ml/
[2] https://www.drmindle.com/ab-analytica/
Part 2a — Jet Setting and a Bifurcated Campus
SAS had a long history of “listening to users” via the annual “SUGI Sam” feature polls, with results announced at SUGI->SGF. JG/OS took it to an executive level as they jet-settled around the world. In the same timeframe, the posh EBC came online in Building C, sporting a fresh cohort of well-heeled, often stunning occupants, ready to mesmerize existing and potential customers with polished talks on “strategic partnerships”.
JG closed his office in building R, departing halls punctuated with T-shirted R&D geeks, some of whom did not always practice the best personal hygiene. His destination, beautiful new executive digs in Building C, corporate residence of the “bold and the beautiful”— those now tasked with interfacing directly with customers, increasingly from the C-Suite or Government oversight roles. R&D or Tech Support geeks in print T-shirts, cargo shorts, and hiking boots would trek over for lunch appearances in the basement café. The normal visual stimulation provided by both architectural and human elements of Building C was much gentler on the senses.
So why mention this? What does it have to do with what eventually became CAS and Viya?
Perhaps others can fill in some of the blanks, but from the OP author’s viewpoint, it represented a finality in the divide between the developers building SAS technology and the customers we were building in for. It also cemented the growing class of director level politicians within R&D and a more rigid hierarchy of micromanagement — along with increasingly intransigent DevMgrs. In short SAS, the company that made its first $Billion in annual revenue with a home spun, flat Org, almost anti-corporate mindset, and added its second $Billion by scaling that, was now pivoting to become much more like IBM in the name of making its third $Billion.
The the battle for pockets of internal control accelerated to a new level while simultaneously setting the stage for mediocrity to “grow in the dark” within R&D and solutions development groups. Perhaps this is what @ky+1jk94kx0x means, prompting me to ask:
Who were “they who were not supported?”. What were their grand ideas that were technologically superior to what OS did in his **HPA->LASR->CAS progression? Were these exceptionally intelligent people simply not politically connected enough in the new era of SAS corporatized culture?
**technologies that JG either had a direct hand in the genesis of, or green-lighted after taking OS on more than one international “analytics roadshow” where are the duo met with large international SAS customers in the financial sector. I distinctly remember those trips in the internal media coverage they got back around 2008.
… TBC
@me+1jk94kx0x —. PRECISELY! 👏👏👏
Many tech companies in SV empower their Principal and Distinguished Devs with Director and Senior Director authority . They are peers at these respective management levels and can only report to a peer or higher level. Thus having proven and trusted technical Mojo, Principal/Distinguished (P/D) don’t answer to line level managers like they do at SAS. P/D have direct access to VPs and regular meetings with SVP‘s and the broader technical brain trust of the company. It makes all the difference in the world when it comes to innovation and getting new initiatives in step with a cohesive vision.
Also, to be clear, historically, many Principals at SAS would be titled lower at Senior or Staff Engineers elsewhere. FaceTime
Yep. This is my recollection of those years as well. Remember the skip-level meetings Oliver was having with his direct reports' direct reports? He started doing that when it became clear to him that his direct reports were lying to him or, we'll say, "misrepresenting the truth".
And unfortunately you are exactly correct, the real power in R&D at SAS isn't the CTO. It's the Development managers who report up to the CTO. They're the ones who turn a mandate into a suggestion and they are experts at managing expectations. In fact, that's the most essential skill of SAS middle management, paired closely with "don't fly to close to the sun".
As it happened to Goodnight, and then to Armistead, and then to Oliver, and now to Bryan. The CTO always arrives with an ambitious plan and leaves having seen their ambitious plan nibbled to death by moppets. First a little here, then a little there...
No modular "building block" Viya architecture because microservices? It was never clear. Versioned APIs are sacrificed on the altar of technical complexity. V9 compatibility recedes forever into the future because the edge and corner cases are too hard to solve. Every little excuse and failure is another nail in the coffin of the ambitious plan, over years. All they have to do is wait it out.
Why not bring those "vanity projects" into mainstream R&D? Because their technical lead constructed a detailed plan explaining why it wouldn't, couldn't, work. The plan might not be on paper, but it exists and they are constantly evaluating everything they know about the ambitious plan and trying to figure out ways to protect their vanity project that won't get them fired. Whatever they say, they explain why the ambitious plan would cost more and could never succeed. Why do they do that? Because their vanity project is their chance to show everyone else what they've been doing wrong all this time, to do it their way, to make a name for themself. Or because they don't want someone else making their job any harder.
'The buck stops with OS because the orders came from him. Ultimately, perhaps that is part of the reason Oliver departed?"
Maybe it was JG's way of testing OS to see if OS had the mojo necessary to get other divisions walking lock step with OS?
Weak leadership skills could have been OS's Achilles Heel...
The buck stops with the person who gave orders to The Forester, The Photographer, and The Yeller.
However, the comments about SingleStore are on point. Most VCs are just sharks. JG is much kinder, and more loyal, to his employees, and is insufficiently appreciated for it.
If we ever get this thread back onto Viya, and the reasons for it, I’ll be glad. But whatever the reasons, the decline of SAS will be gentler under JG than under anyone else.
"failure is the collective fault of all of the other division leaders."
Nope. The buck stops with OS because the orders came from him. Ultimately, perhaps that is part of the reason Oliver departed???
OS would make statements that set the strategic direction for how things were supposed to work moving forward. For example, all solutions must convert to using CAS server. I took these as mandates. Others took them as mild suggestions.
So, failure is the collective fault of all of the other division leaders.
"Many people in the world were better qualified to be CTO. But he was the best we could get without paying more or providing equity."
In my opinion, it didn't matter who you had as CTO. JG wasn't going to relinquish control of anything. Whomever took that position would be stymied by JG's control efforts, and rendered ineffective.
Please remember, JG also allowed the clown with the dirty pictures to run R&D.
“How many of them were smart enough to garner JG’s support? … Why isn’t the fruit of their intelligence and hard work now saving SAS?”
Because they weren’t supported.
I’ll agree that the Big German was intelligent, and he worked hard. And I don’t know the dynamics of S2, except to agree that most VC’s are cutthroat.
And I will admit the BG was in an impossible situation, competing against Open Source and huge public companies.
But we never should have been in that market.
Old former SAS customer here. This site is a good way of remembering the past.
I started using SAS in 1980 and loved it.
The first time i was around Oliver S was Las Vegas SGF. Oliver S was in the opening session and he talked on and on and on, su-king the life out of the audience with his techy mumbo jumbo. My boss was there too. She asked me to explain his spiel in plain English. My answer was "He did not say anything related to how we are currently using SAS". Embarrassing yet true . A year later we started pulling back from using SAS. My younger colleagues fresh out of college never saw the value of SAS. Their voices were louder and there were more of them every year. And fewer SAS users. So relieved to be retired now.
Back to lurking. Good luck everyone!
… “dozens of smarter people, inside and outside SAS” …
Let us guess. Many of those “smarter people” at SAS were among the spoiled children maintaining the status quo! How many of them were smart enough to garner JG’s support for building better technology than OS did? How come he prevailed for so long and they did not? Why isn’t the fruit of their intelligence and hard work now saving SAS?
Also, most people denigrating OS’ tenure and end at S2 are just making sh-t up and have no idea about insider dynamics there. A couple of us here do. Sadly, S2 is not exactly on a rapid growth trajectory. The founders and top talent have left and they are a long way to a double digit ranking
https://db-engines.com/en/ranking
The big difference with SAS as a private company, JG has historically been much more benevolent with layoffs. S2 has become a place driven by Silicon Valley and cutthroat politics under VC rule.
@ke+1jk94kx0x That’s hilarious. I had almost the identical experience. He just walked away.
He was not socially skilled, and not terribly bright. I worked with dozens of smarter people, inside and outside SAS.
He had no “wealth of experience and skill… that few people in the world possess.” Singlestore certainly didn’t think so; they let him go after two years.
Many people in the world were better qualified to be CTO. But he was the best we could get without paying more or providing equity.
He was intelligent, and he worked hard, everyone agrees on that. And it was impossible to compete against the public companies and Open Source.
But why were we in that market?
We should have built something they were not building. A really smart person would have seen that.
I met Oliver at SGF during a social. He was standing in a crowd not talking to anyone and no one was talking to him. So I walked up and introduced myself.
Me:: Hi, i am John Doe and I work in the yada yada division doing such and such.
OS: My name is Oliver. He turned away. So, I walked away. First impressions can be lasting impressions. That can be good. Or not.
Many of us were/are, years prior to OS and after as well … within and without SAS
Once he trusted your competence, he gave plenty of freedom to innovate and even advise him on options for optimal technical directions
I don’t understand, then, if he and you/they were so smart, capable, and trusted, why could you all have not been successful all by yourselves?
There's no doubt that OS was highly intelligent and probably had many other good qualities, and there's no doubt he had some significant flaws. That's true of any successful leader....just look at Elon Musk, Mark Zuckerburg, etc.
But keeping up with the hyperscalers, with the limited capital of a privately owned company, is pretty much impossible for anyone.
Even if OS had been the world's most brilliant and charismatic leader (which he obviously wasn't, not even close), without access to the capital markets of a publicly listed company, it was an impossible task. As recently as 2023, SAS was making a big deal about spending "a billion dollars on AI". That's laughable when the hyperscalers were spending 10x, 20x, 50x that.
If JG wants to make a difference, he'd be better off shutting down the whole company and turning to philanthropy, because SAS R&D spending is just a complete waste of money at this point.
I spent many intense years working for and with OS. Once he trusted your competence, he gave plenty of freedom to innovate and even advise him on options for optimal technical directions. He was tough, yet fair and spent considerable effort building his team up.
RIGHT ON! Perhaps some of those same spoiled children are the ones hating on OS here.
"After he was promoted to CTO he couldn't drag the rest of the company along with him in his wake, but no one should have expected that in the first place. The same things that make you a developer's developer don't make you a good C-level executive."
I was in R&D, but not a Developer. OS would make statements that set the strategic direction for how things were supposed to work moving forward. For example, all solutions must convert to using CAS server. I took these as mandates. Others took them as mild suggestions. The products I worked on that were in early development did NOT convert to CAS server. They continued as they had previously.
I asked the lead stakeholder "Why are you not converting to CAS, per OS's mandate? Isn't OS the new boss now?" I was met with smirks, shrugged off, and told to continue to work on the new vanity project as if OS had never spoken.
Perhaps OS didn't have what was needed to rally the troops. I argue that many of the troops weren't people that could be rallied. Spare the rod and spoil the child they say. Well, these were spoiled children, folks whose behinds never served as a back stop for the high-velocity leather.
How do you manage that unruly crowd of Montessori children? How do you manage people who've never experienced natural consequences? You can't. It didn't matter how good OS was at management, he was dealing with a bunch of spoiled, entitled children who knew they'd never be spanked. They just waited OS out, since they knew they'd face no consequences. SAS is in decline, in my opinion, partly because it fostered a culture of spoiled, entitled children with no experience in reality or accountability.
OS’ career and life choices are his business. Hate on him all you want but I’d love to see you toe to toe in a conversation with him concerning the depths of modern analytics and data management — especially if the math is involved.
"He brought a wealth of experience and skill to S2 that few people in the world possess."
Statement above seems like BS. Rather than upping his career post SAS, OS flopped back to Va Tech.
“SPDS, TKTS, several silly failed solutions and perhaps HPA, LASR, Hadoop 💩, and CAS consumed 25+ years of development time and $Billions — all while the cloud hyper scalers and OSS changed the real game and captured a massively accelerating data and analytics market while SAS barely kept up, only to ultimately fall behind. Such was the opportunity cost of doing things “the SAS way”.’
Thank you. That’s as good a summary as i’ve ever seen.
Sheesh, with your technical abilities and powers of discernment why aren’t you CTO at a major, successful company?
IIRC, TKTS never gained significant traction and got “parted out” as a set of internal services and ultimately landing as the Federation Server. SPDS was OK, but sill a “parallelized bag” hung off the side of MVA after attempts to add asynchronous I/O to host and core failed.
In the grand scheme, SPDS was not a significant source of revenue ($12M annually?) and its chief architect’s often surly demeanor made OS look a most congenial person. Taken in total:
SPDS, TKTS, several silly failed solutions and perhaps HPA, LASR, Hadoop 💩, and CAS consumed 25+ years of development time and $Billions — all while the cloud hyper scalers and OSS changed the real game and captured a massively accelerating data and analytics market while SAS barely kept up, only to ultimately fall behind. Such was the opportunity cost of doing things “the SAS way”.
OS brought much needed advanced analytics expertise to S2. There is more to his departure from there than meets the eye. Silicon Valley/VC politics played a significant role. Several high-value tech people also departed S2 in that time frame.
Be careful of projecting latent SAS drama on OS’ next act. He brought a wealth of experience and skill to S2 that few people in the world possess.
- SAS has been slowing going to he-l in a hand basket for the past 25 years now.
Hmmm, that's not my take from this thread. Mine was that Oliver destroyed SAS culture without benefit (maybe harm). So maybe the last 15, just when he took over.
- Blah summary SPSD -> TKTS -> HPA -> LASR -> CAS
SPDS/TKTS were engines not core architecture. Quite nifty actually and SPDS was a quality piece of software. BTW, SAS was at 15÷ at that time and grew after SPDS.
- - All attempts at solving the long standing problem of parallelizing data with SAS.
The first four support your case and IMO were reasonable tries. Well, HPA and LASR were poorly implemented but the effort was understandable.
CAS was demonstrably a horrible choice and cost the company as a whole.
- - Apparently JG should have put someone else (@hb+1jk94kx0x perhaps?) in charge
I think you now understand the point of this thread.
"The same things that make you a developer's developer don't make you a good C-level executive."
OS is a good example of the Peter Principal. His post SAS job fizzled. After that he went back to his comfort zone. Academia. An unusual career path for someone in his prime. "A man has got to know his limitations". If that applies to OS, good on him.
I want to think OS is a much better person than he was at producing wildly desired software.
Former CAS core team member here. Loved working for Oliver and still consider him my best manager in 30+ years at SAS. I think the dual role of CTO/COO was too much for one person, even someone with Oliver's discipline and drive.
Some of the "being underpaid" came form the reality of salary compression impacting long-term SAS employees. One could attain and maintain the top-of-their game by working 10+ extra hours/week for years, only to get minor annual raises/bonuses while those much younger, hired 20 years later into your division/group exceeded your base.
This is reality industry-wide yet harder to overcome when equity is not also given. It does not take a genius to recognize that a significant problem with productivity and effectiveness at SAS correlated to hiring a cheaply as possible.
After 2005 or so, try hiring top-tier disturbed systems and analytics devs for what SAS paid. Most went to FAANG or hot startups. Could this correlate to not designing and building competitive modern data management and analytics products that kept reasonable pace with the grown of these markets in the past 20 years?
Hmm. To my recollection the core CAS team was the team of developers that rose from the ashes of LASR and Oliver's first, brief flirtation with being a "manager" under CB before it was made official and he was promoted directly to Director, past Manager and Senior Manager. Every person on the core CAS team was an extremely talented developer, and many of them _loved_ working for / with Oliver. Some of them were almost fanatically loyal.
I often wonder if their opinion has changed in the years since, but no one can deny that Oliver inspired the people that were eventually sorted into ASR, or that they were a good team. So he had successfully "[organized] enough hard workers to produce competitive software". After he was promoted to CTO he couldn't drag the rest of the company along with him in his wake, but no one should have expected that in the first place. The same things that make you a developer's developer don't make you a good C-level executive.