Thread regarding SAS Institute layoffs

Fair point: some V7 work was saved. I can assure you that much was discarded, because some of that was mine 😔. But if all of Viya is eventually discarded, that’s much worse of course.

Viya could become a salvage project. There were many bright people working on it, so surely there are valuable parts to be saved. If that were done, Viya might have a better ROI than V7.

If that were done 😔.

I don't think there is a meaningful comparison between V7 and Viya. I don't know what parts of V7 were thrown away (I'm not doubting it, I just don't know), but in the Applications Division (later advanced analytics), we had a huge effort to go from V6 8-character all upper-case variable names to the (unfortunately named) VALIDVARNAME=V6|V7|ANY. It has been so long that I don't recall other details, but we did a lot of work that went straight asis* into V8, which sold well. Even with V7 being a flop, a lot of the work got used in V8, so IMHO it is not comparable to a billion-dollar effort (or whatever it cost) to make Viya while simultaneously trying to ki-l the goose laying the 3B a year golden egg.

Yes. Version 7 was so bad that we gave it away for free.

Then we threw much of its code away, and quickly shipped Version 8.

Back then, SAS had no strong competitors. So the market forgave us. It's different now.

Has SAS ever had anything with a worse negative ROI than Viya? Serious question.

The Vegas story suggests that Viya was from the beginning a niche product, attracting a small market segment willing to pay for fast cloud performance.

Others have written that JG wanted “One SAS”, that is full compatibility with V9. This would have produced two versions for two market segments, with an easy migration path.

If that was JG’s vision, it makes good business sense. Why it has not been achieved is not clear. It’s been said that OS refused compatibility; but he’s been gone for years.

The compatibility problem is difficult, but it doesn't have to be completely solved. If SAS could claim 80-90% compatibility, and treat the edge cases as a consulting opportunity, most customers would be happy with that.

"V9 SAS gets disrespected by people in this group"

Disrespected mostly by those too stupid to realize that V9 paid their salaries and benefits better than any other SAS offering. That is a fact that the Viya sycophants could never grasp much less truthfully utter.

Sounds like cutting off one's nose to spite his face, eh?

"Presumably, users were there because they used V9."

No presumption necessary. Using V9 is exactly why they were there. I was there also, as an employee having many years of SAS experience via usage and can certainly relate to their(2 tall guys) ad nauseam uninspiring droning away. When they were done, I left with a feeling of exhaustion instead of being energized.

Hind sight is always 20/20. Viya was a big gamble against open source and time has proven Viya to be a failure.
Simply because open source is thriving and Viya isn't.

SAS leadership should have realized MUCH sooner that they should have folded Viya because time kept proving over and over that Viya was not the answer that the market was seeking. Hubris can be and was the downfall of SAS. There is no better single word to define that than "Viya".

Inflection and tipping points do occur. The Viya spiel at the opening night at Vegas SGF was the beginning of SAS circling the drain.

I vividly remember the day Viya was announced. It was at Global Forum in Las Vegas. They waited until the end of the long opening session. Then the two tall guys sat on the stage and droned on seemingly forever. People were getting up and leaving in droves. It was Las Vegas! There were things people could do! Oblivious to that, the tall guys just sat there droned on and on and on. It showed their arrogance. Presumably, users were there because they used V9. Some might have had a need for what Viya promised, but probably not most. I know V9 SAS gets disrespected by people in this group, but it remains a powerful and useful system, even if it is a bit dated. It is such a pity that the tall guys did the whole “one train is leaving the station for the last time” thing instead of nurturing a 3B a year revenue stream while simultaneously trying to develop something new that customers actually wanted with value added beyond free software.

"SAS is working to fix Viya’s problems, but not investing enough to fix them in a competitive timeframe."

Insufficient Viya sales to right that boat. Sadly, the Viya game clock has expired.

@suwz+1uhfEpsR

Spot on reasoning IMO. It's the unfortunate reality

“To buy time for what?”



1. To fix Viya’s problems?

2. To innovate in some new direction?



3. To shop the company around and find a suitable buyer?


SAS is working to fix Viya’s problems, but not investing enough to fix them in a competitive timeframe. SAS lost so much engineering talent in recent years, and has not replaced it.

SAS' insular culture has stifled innovation for many years. The company is a prime example of the Innovator’s Dilemma. “Too slow to pivot” is too kind; they’re unable to pivot.

That leaves Door Number 3.

@rowl+1uhfEpsR "SAS's response? Viya Workbench. And and endless supply of executive Trustworthy AI webinars to buy time. To buy time for what? You know. In your heart of hearts, you know."

My heart of hearts does not have much smarts. For what is time being bought please?

SAS was founded by academic nerds. They had egos and hubris, too.

But JHG, raised in his Father’s hardware store, learned how to take care of customers. That’s the difference.

As @rowl+1uhfEpsR says, in your heart of hearts, you know what comes next.

“So how does this relate to the SAS we’ve been using all these years?”

The best answer to that question is to recall who was the brain child of Viya. An academic nerd having no real world business experience, coupled with an astronomically big ego and hubris, and worst of all... an unwillingness to listen to customers. That is how we got from way back then to now.

It's the Innovator's Dilemma. Read Clayton Christensen's book "The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail". Read the case studies in the book. SAS is just another case of too slow to pivot.

Successful companies face a dilemma when dealing with disruptive innovations (cloud & open source in 2010's and now Gen AI in 2020's).

SAS's focus instead has always been centered on meeting the needs of their most profitable customers. This lead them to ignore early-stage disruptive technologies. By the time those disruptive innovations matured, it was too late for Viya to compete.

SAS's competitors are releasing SQL functions that return prompts from LLMs (https://www.databricks.com/blog/2023/04/18/introducing-ai-functions-integrating-large-language-models-databricks-sql.html).

SAS's response? Viya Workbench.

And and endless supply of executive Trustworthy AI webinars to buy time.

To buy time for what?

You know.

In your heart of hearts, you know.

Viya was and, to all appearances remains, a classic “Fire Ready Aim” proposition. SAS upper management never settled on a consistent response to the very reasonable question, “So how does this relate to the SAS we’ve been using all these years?”

Other than some niche such as pharmaceutical and the 3rd worlds of AI/ML practices that rely on low/no code, SAS' days of being a 'platform' provider has gone. Enterprises start with "what cloud provider to use?" That's Azure, AWS or GCP. Then "what data platform to use?" That's Databricks or Snowflake. The answer to these two questions already determined what AI/ML platform to use. e.g., AWS --> SageMaker, GCP --> Vertex AI. SAS comes to mind only when they need a comprehensive time series forecasting library. "I have to buy all of Viya to get that? No thanks."

Perhaps some humor can be extrapolated from the following with regard to the current state of Viya and SAS in general:

https://youtu.be/0SfHKR_7NjM?si=mSLdH6trEMzhaJIx

Pay special attention to the first thing Joliet Jake says to the Mother Superior.

@5htg+1uhfEpsR

My (limited) understanding is that Viya’s problems are threefold:

1. It does not compete on price with Spark et. al.

2. The products offered on Viya are not compelling to customers.

3. For SAS9 customers, there’s no easy migration path.

I wish more knowledgeable people would weigh in on this thread. Whenever Viya is discussed, I learn something.

I expect the next 18 months will be telling. At this time, SAS can claim, as BH has, that Viya’s sales are increasing. That”s plausible, from a low base. But it’s not sustainable, unless the problems are addressed.

"Looks like the ” quiet part out loud” thread has fallen victim to admin discretion. Perhaps they are just pruning and it will return in a redacted state?"

HIGHLY doubtful

"World class Product design/management and a fresh crop of A-list engineers could build successful niche products atop a select set of existing Viya components. Would take some tweaking but there is a lot of good infrastructure there."

That is positive wishful thinking!

Realistically speaking, your idea requires an investment above and beyond the fruit that Viya has thus produced. The unfortunate reality is that most of the Viya fruit is rotting on the ground because customers simply do not like the taste of it. If the founders are committed to selling, they are not going to have much interest in spending money for water and fertilizer to invigorate a dying Viya tree. The best hope is for a buyer to see the potential of your plan...a beat case scenario. And the problem with that scenario is that most buyers will only want to pay for the declining v9 revenue stream.

Bottom line, the founders need to adopt the same mindset as the typical buyer. That likely won't be easy change for them to come to accept, simply because they have not thought that way for decades. If they can not or will not have that change of mindset, then SAS won't sell before they pass away. And once they pass away the heirs will have an expeditious estate sale and SAS as we know it will be forever gone.

The remainder of this year will be very telling. If little changes are made, expect that estate sale to become an increasingly likely reality. In crude terms, SAS is at a po-p or get off the pot moment.

Looks like the ” quiet part out loud” thread has fallen victim to admin discretion. Perhaps they are just pruning and it will return in a redacted state?

This notion has been repeated many times on threads over the past year or so. World class Product design/management and a fresh crop of A-list engineers could build successful niche products atop a select set of existing Viya components. Would take some tweaking but there is a lot of good infrastructure there.

Simply a matter of the chief executive having the will to exit the current ineffective resources and replace them with competent modern skilled professionals who are not stuck In the delusion of Shangri-La thinking. The existing bloated sales and marketing model will remain fundamentally broken until SAS starts producing products the market wants.

Anyone who says CI360 is a shining star has never been forced to use it. It looks like a relic from the heyday of AOL and the UX is even worse. If it was an animal, a kind soul would put it out of its misery.

Some of these comments need a /s

No idea now who’s joking, maybe everyone

"SAS CI360 vs Adobe vs Salesforce...

No contest! CI360 all the way! 😆"

True, CI360 leaves those b-ms in the dust.

Let's make sure the CI360 team gets big bonuses!

SAS CI360 vs Adobe vs Salesforce...

No contest! CI360 all the way! 😆

"CI360 is one of the shining stars of SAS"
hahahahahah....so funny!

"Have you used CI360?"

Agreed, CI360 is one of the shining stars of SAS!

Literally everything at SAS is created in a vacuum with no market research beforehand. Have you SEEN the software? Have you used CI360?

we weren't ever told much so had to speculate. going from obsolete architectures to cloud native big data tech? ok made sense from a tech perspective. certain use cases like detecting fraud on big data in real time needed that leap? ok, makes sense. but most other v9 use cases and general bi/analytics ones? my new employer is using all legacy bi tools. they're all meh. something way faster, for much bigger data, way more expensive? more than what we need. if the plan was/is to let these kinds of sites use a small slice of big data tech at an affordable price ... maybe there would be something there. if that was part of the plan, it's a well kept secret. if it's not, why give up on so many opportunities and only go for certain ones, then lose at that, and still not adjust? the entire thing is baffling.

pipeline good!

fire bad!

"The best metric of Viya's success is for the market to embrace it. And that has not happened so far. Is it realistic to expect that to change?"

About as realistic as expecting to look younger as you age. But hey, we are told that the the pipeline is good.

Was it for internet scale companies (not addressable market)? V9 customers (not fully compatible)? Something else?

What are some capabilities that somebody thought would be desired? Better? Cheaper? Faster?

"Doesn’t Viya provide quite a bit of compatibility with legacy/SAS9 then?"

If the answer was yes, customers would be flocking to Viya. The best metric of Viya's success is for the market to embrace it. And that has not happened so far. Is it realistic to expect that to change?

"So there was significant, new capability in Viya that was unavailable in SAS v9."

Unfortunately, almost none of the new capabilities were of interest to most of our V9 customers.

Significant new capability like running advanced AI/ML algorithm/models in CAS natively on Nvidia chips. Anybody else think that could have been a huge win had the right engineers and technical leaders remained working on it?

Five years ago SAS had a half dozen PhD-level ML engineers who are among the best on Earth. They all left and some are likely making 3 to 4x what they were making at SAS because they work for Nvidia now.

@1bik+1uhfEpsR
"But was it true that Viya initially just a framework that customers had to upgrade to (and pay more for) in order to get new functionality in the future? That it offered nothing new initially?"

No. Not sure where you'd have gotten that idea.

Creation of SAS v9 statistical procedures largely ceased in favor of developing new CAS analytic actions. So there was significant, new capability in Viya that was unavailable in SAS v9.

What is wrong with Viya? The best answer to that question is another question.

What else was better than Viya to motivate customers to walk away from SAS?