Thread regarding Teradata Corp. layoffs

A frank question: What's the new direction?

Just in time data and analytics on cheaper, flexible, faster platforms that have excellent tooling support.

A frank question: What's the new direction?

Mega majors who can and do have an EDW (ours or from the competition) are shifting directions

&sas-institute has no posts, should I assume they are doing just fine?

People commenting on SAS + Teradata as having a market beating offering must be smoking something strong. Are you guys in Colorado?

SAS is in exactly the same boat as TD.

• A customer base that is traditional, old, dying and is resistant to any change.

• Newer, more agile technologies that are far better, cheaper and provide equivalent experience,

• Lack of traction with the younger parts of the organization.

• Caught in the headlights reaction by throwing everything including the kitchen sink at the customer hoping something will stick.

• Attempt to get invited at the cool kids table by buying up sub-standard, borderline useless companies and shutting them down 12 months later.

• Complete inability of the middle to top management to comprehend the winds of changes.

People beating the drums of integrated data forget that it is possible to integrate data in places other than the precious EDW. No one here is questioning the value of integrated data, they are questioning the value of the EDW. These are not synonymous.

"an EDW is not RDBMS"

Mind blown...

So True:

Sales director, Account Exec, IC, SA, EP, other SME and few people talk but not much value added.

#gold

The guys who think data has to be integrated all the time are living in fools paradise. The reality is the clients have marts and they also have platform burning issues. We keep pushing the agenda of EDW which is not real anymore. The biggest problem with Teradata is our messaging. It all depends on who shows up onsite. One day our CTO will say to the client we should do integrated EDW even though we have screwed up large number of integration projects in the past. On other days someone will show up from PS and say we should do TVA. On another occasion think big guys show up and say everything should be on Hadoop and by the way we are vendor neutral. The reason we are in the mess today is because we didn't invest good money into our future. We took close to a billion dollar and invested it into buying applications and other acquisitions and today no one will give us what we paid for those products and acquisitions. If we had invested the good money in something valuable we would have been in better shape.

The other part is ground realities. We refuse to face realities at client site and keep throwing every other product at client. When they need apple we give them oranges, grapes and watermelon. We have to listen to the client and solve their problems. We have to operate like a small company and start with smaller teams. Right now we have too many cooks trying to cook a simple meal. We have meetings at client sites where large number of Teradata folks show up. Sales director, Account Exec, IC, SA, EP, other SME and few people talk but not much value added.

@FwMkxJd-1ycq

"an EDW is not RDBMS"

A very important point. Also, I 2nd the rats nest that is a SAS implementation as I've seen them. Talk about duplication!

"an EDW is not RDBMS"

A very important point.

Don't look far, but SAS already has a real threat on their hands with Microsoft competing on all fronts with Power BI (vs. SAS VA) and Revolution R (vs. SAS Base/Enterprise Miner). Microsoft is also building a cloud analytics strategy with Azure ML. This is a clear mid-market play, with the 2 biggest lures: low cost and ease of adoption with easy integration with Microsoft's Windows and Office stack -- which EVERYONE has.

So imagine in this fight, Teradata's differentiators are not clear to a mid market customer.

Exactly the point about differentiation I've made in the other post. SAS has its own problems because their core value prop (a comprehensive analytical engine) is eroded by bad implementations, unhappy customers, and high cost. The culprit: data marts! I can't tell you how many out of control SAS data marts I've seen in my lifetime. Irony is: had Teradata and SAS combined their messages, and then incorporated Hadoop early on, this would have been the dynamite offering to beat in the market.

Don't look far, but SAS already has a real threat on their hands with Microsoft competing on all fronts with Power BI (vs. SAS VA) and Revolution R (vs. SAS Base/Enterprise Miner). Microsoft is also building a cloud analytics strategy with Azure ML. This is a clear mid-market play, with the 2 biggest lures: low cost and ease of adoption with easy integration with Microsoft's Windows and Office stack -- which EVERYONE has.

So imagine in this fight, Teradata's differentiators are not clear to a mid market customer.

Exactly the point about differentiation I've made in the other post. And yes, SAS has problems in their space because they also had a core value prop which is a comprehensive analytical engine, but has been eroded by bad implementation, high cost, and a long list of unhappy customers. The culprit: data marts! I can't tell you how many useless out of control SAS data marts I've seen in my time. The irony in this case is had SAS and Teradata cooperated more, and later incorporated Hadoop as technology shift, this could have been the best market offering by far.

This is missing the point entirely. Mega majors who can and do have an EDW (ours or from the competition) are shifting directions. The industry as a whole is in a long-term decline and has been for more than 4 years.

Actually the point I'm making is differentiation - and my view is the crux of the discussion is how TDC can continue to differentiate. From an outsider looking in, Integrated Data, Analytics, and Hardware have been differentiators for TDC in the past - but is it enough to keep the company afloat these days? Remove those differentiators then the only other thing TDC can be known for is "expensive" - and that's not a good brand story in a big data era.

Calling the whole thing merely a bandwagon is trivializing the issue.

To your other point, there is nothing that stops various Big Data technologies to be integrated. I would argue that advanced practitioners are in fact doing exactly that. Getting an integrated, fluid and very flexible view of the data , it just happens that they don't need an EDW to do it.

Trivial or no, it is the truth - most clients thought they could Hadoop their way into datawarehousing to save cost, and many still fail.

On the other hand "Integrated Data" is a core value prop of TDC - and even with the change in technologies the value prop holds true. Companies with integrated data have a competitive edge to those that do not. The balancing factor has always been time-to-market/agility - in the quest for integration if it takes too long or costs too much, negates that competitive edge.

Can TDC continue to sell the value of Integrated Data? Even better - can it be had at a cheaper price? If you look at the Horton and Cloudera web sites - this is their main play actually.

But you actually unearth a further point that even Teradata needs to realize - Data Warehousing and RDBMS are not the same thing in this value prop. Data Marts and Data Warehouses are a key difference - you achieve more incremental value with more integrated data in an EDW vs. a Data Mart, but an RDBMS (even a parallel one) is no longer the sole way to achieve a good EDW. Emerging hybrid architectures already include Hadoop as part of an EDW - but the goal is still the same thing - have more data accessible in one place. The opposing approach is fit for purpose data marts - which only get away with cost and agility as the main driver - but again this is an inferior strategy from an integration standpoint.

As an aside, this is just like analytics and data warehousing are not the same thing - although the strongest use case for data warehousing is the efficiency and extent of the analytics that you could do on integrated data vs. small sets in a data mart - again going back to the previous point.

2) The Hardware first focus: Trying to force our clients to buy our appliance if they want a big data implementation.

This is arguably the biggest paradigm shift Teradata needs to re-evaluate - and the most painful. The close integration of Teradata appliance and software is a another key differentiator vs. competitors from a performance standpoint - they just work so well. Take the Teradata database out of the box and you narrow the edge vs. other MPP offerings - it's just another database.

But as with the thought that an EDW is not RDBMS, same is true here - an EDW is not an Appliance either, but you could still argue that there are still pros and cons to a highly bare metal implementation vs. something in the cloud - and again it is not one-size-fits all, but especially true for mid market - the motivations can vary now. Many smaller companies do not need a super computer in their building, they just need a way of accessing and visualizing their data faster - because the competitive pressures are different.

Now when integrated data, analytics, and hardware performance aren't as catchy, what other thing can TD sell? Expertise. I can say I've worked with generally good people from TDC - they know their stuff. But this is something that may have just been a fluke - the company doesn't sell its people as much as its hardware, and this could be the biggest opportunity loss of all. Since the layoffs are the main purpose of these threads then that's the rub. When good talent walks out of those doors, you don't get that back. And with that talent goes innovation, and customer relationships. Maybe the competitors like SAP, HP, IBM, Oracle, even SAS could buy them, but the magic isn't there because those vendors have their own brand issues. Finally the new guys like HDP and CDH are still in push mode, they have a lot to prove. Maybe the fallouts from TDC could join them, but it will be a very different world.

Somehow the industry is losing something with this development, but maybe it's just another wave of change for technology.

@FwMkxJd-dfj: Great comment.

A few points -

customers are jumping on the Big Data bandwagon and open-source primarily as a cost-management motive (hey its free) but are missing the point about the value of integration (even a good data mart is at least partially integrated) as well.

I have frequently heard this argument. This and the "It is just a blip" should be patented as the TD go to market strategy.

This is missing the point entirely. Mega majors who can and do have an EDW (ours or from the competition) are shifting directions. The industry as a whole is in a long-term decline and has been for more than 4 years.

Calling the whole thing merely a bandwagon is trivializing the issue.

To your other point, there is nothing that stops various Big Data technologies to be integrated. I would argue that advanced practitioners are in fact doing exactly that. Getting an integrated, fluid and very flexible view of the data , it just happens that they don't need an EDW to do it.

@FwMkxJd-dfj

You are so right that Teradata should have put together their own Hadoop Distro. I pull my hear out thinking about the money our leadership has spent the last 4 years buying proprietary software vendors in an open source space! This is so stupid, they should have just spent those millions building the knowledge base internally and working up our own Hadoop offering. After re-thinking things after that post, I now see the following strategic failures as the reason for TD's current downturn:

1) The strategy of buying their way to Big Data credibility, instead building it up internally.

2) The Hardware first focus: Trying to force our clients to buy our appliance if they want a big data implementation.

Nice thread. Just on:

SAS is the biggest and most notorious criminal here - and it's had a love-hate relationship with Teradata for years.

Don't look far, but SAS already has a real threat on their hands with Microsoft competing on all fronts with Power BI (vs. SAS VA) and Revolution R (vs. SAS Base/Enterprise Miner). Microsoft is also building a cloud analytics strategy with Azure ML. This is a clear mid-market play, with the 2 biggest lures: low cost and ease of adoption with easy integration with Microsoft's Windows and Office stack -- which EVERYONE has.

So imagine in this fight, Teradata's differentiators are not clear to a mid market customer.

However, before any of that happens, there are 2 critical factors. The absolute first thing is to concede that we don't know the needs of the consumers who are further ahead in the analytics chain, and if you think you do, it is because you are guessing. We need to remove our blinkers. Success depends on processes. Processes even though might be thought of as abstract, in reality is a function of people at the top. Our middle-to-top management simply does not comprehend the truth. They need to go.

Before anything, full disclosure, I am not a TDC employee. I used to be a contractor/counterparty, but I have worked closely with the company's associates and technology for years.

I like the second part about management needing a revamp - same happened to Apple years ago when it was tanking under Gil Amelio. Jobs came back, fired the board, rationalized the product line, and shifted the focus back to innovation and drilling down to customer requirements.

On the other hand, I disagree with the idea that the company doesn't know what customers need. I think it's the opposite: customers are jumping on the Big Data bandwagon and open-source primarily as a cost-management motive (hey its free) but are missing the point about the value of integration (even a good data mart is at least partially integrated) as well as disciplines around data management that impact the quality of analytics - this is Teradata's message. TDC in fact has a good story to tell, but just got hit by the cost-reduction bus and needs to assess where to go.

Some strategy challenges I can say:

• Do you continue to focus on your previously profitable niche market? The mega majors who need an EDW simply because they are too big? Or do you go down to mid-market where arguably cost is a bigger motivation, hence cloud, data marts, and agile development rule? I think the jury is out on the need to expand revenue - hence play in the mid market, but the company is a bit half-assed about the execution. Even the decision to support Horton and Cloudera came slowly. I actually was previously betting on Teradata developing its own Hadoop flavor years ago and if played well it would have locked in the client base to a trusted name, but there was too little investment and too late for that.

• On the other hand, while supporting Horton and Cloudera is smart, there is a fundamental problem with this: both HDP and CDH do not have any qualms bad-mouthing Teradata as passe, and the trend towards virtualization and cloud servers have already destroyed the market for a Hadoop appliance before it even started. Teradata's classic response has always been: "But the performance...?" - but to a mid market company this is a non-issue.

• Applications? Yes Applications - but not just any applications. It has to be Analytical Applications. This is closest to the core of Teradata's experience and existing technology. Notice in this space, TDC also relies on partnerships: notably SAS, Fuzzy Logix, Rev-R, Tableau etc. Problem the same as the Hadoop space: these partner vendors also have no qualms shooting down an EDW requirement, because it prevents them from selling quickly. They'd rather position a quick and dirty (and eventually useless) data mart which their apps can run on. SAS is the biggest and most notorious criminal here - and it's had a love-hate relationship with Teradata for years. Teradata should a) buy or merge with one of these guys, or b) create its own analytical application offering/s, or c) scale down its core EDW offering to accomodate small, agile, marts that can support analytical apps then scale up from there.

• Aster: The argument for analytical apps or a scaled down EDW with analytics would have been great for Aster - but as it played out, Aster was positioned as an analytical engine, but in a form factor that required yet another appliance! The perception was muddled further when Aster was recently put on Hadoop, which then sounded like a SQL-on-Hadoop offering, which was good since it is a common problem with Hadoop users, but then Teradata made a bet on Presto - which was the same thing but free! So as a customer when you think about what exactly Aster is, you get three possible results: it's a SQL-on-Hadoop engine, it's a database with some machine learning algos, or its a data mart appliance with built in analytics. They are all actually good positioning plays, but I think marketing 101: customers will find it hard appreciating a product that is everything in one. We live in a point solution world unfortunately (as Oracle, IBM, and SAP have learned long ago...). This could just be a tweak of messaging though and it could go a long long way.

• Teradata Consulting - this is a story could almost be told but never is. A lot of what customers appreciate with many Teradata engagements is that the people come with experience dealing with other customers, and also have an appreciation of the industry that the customers are in. Many TDC consultants have a business background and bring that with them on engagements. Unfortunately, the consulting resources are quite few, never enough, and cost a fortune for a customer to retain. There is knowledge transfer at the end, but I think TDC is missing an opportunity to build a sustainable services offering. All TDC consulting engagements usually come with or from an appliance deal - and this locks TDC out of a big lucrative market where Accenture, IBM, and SAS are playing (and ripping off people). Seriously think about this Teradata if you want to evolve and survive.

@FwMkxJd-pfs , @cranton:

I am old enough to remember when the BlackBerry was the enterprise smartphone of choice and it had capabilities which the newly released iPhones and Androids simply could not match.

There is nothing wrong with people realizing the benefits of implementing big data right now, it's just reality that the EDW is necessary to get the most out of it.

This is a false premise and i believe the reason why i say we have blinkers on. I include myself in this.

We have spent our careers at TDC and cannot fathom an eco-system without an integrated warehouse. However, from the case studies of our largest customers (including eBay) it is clear to me that the EDW is now a small, fast diminishing and in many cases completely irrelevant part of their data architecture. We have to moderate our expectations accordingly or attempt to play in the new space.

"Big Data" is just like "User Friendly", "Web 2.0", "'the Linux Revolution" and a host of other buzzwords

This is a dangerously hubristic notion. If Big data was simply a buzzword it would not be having the impact it is on our top-line and stock price. We would not have lost colleagues and friends continuously over the last 12 months plus.

I would also argue that Web 2.0 and The Linux Revolution are in fact real.

\/ below mine.

You can compare TD to Blackberry if you want, but the reality is the industry analysts at the time called Blackberry inferior, whereas TD is considered best in class. There is nothing wrong with people realizing the benefits of implementing big data right now, it's just reality that the EDW is necessary to get the most out of it. People have the crazy notion that just maybe big data will replace the EDW itself, which doesn't even make sense.

"Big Data" is just like "User Friendly", "Web 2.0", "'the Linux Revolution" and a host of other buzzwords touted by grasping CIOs that have been whipped up by industry publications. They call a big meeting touting the savings if all the company desktops were to go Linux, order some hardware, do some tests cases, but nothing ever really comes of it. The same thing is happening in lots of big businesses across America who are rushing to get a big data cluster stet up as soon as possible to analyze some paltry crap that doesn't really constitute a use case.

Again, the lesson is that big data is special and only for unstructured data, and the majority of companies that are trying to do it to squeeze out some amazing new insight, will only really use it for cold storage and possibly ETL. Both good use cases, and in fact what TD should be focused on.

@FwMkxJd-jjn

"I wonder why the rest of the market cannot see this. The trend-line for the entire traditional RDBMS market has been down for the last 4 years. This cannot be an aberration.

Your theory that that this is a temporary blip is a fallacy, HBR calls this the BlackBerry argument. For a long time, Blackberry executives would turn up at earnings call and simply tout the superiority of their platform and say "the people" are wrong for preferring Android or iPhones. It will come around. We all know how that turned out."

Except that in reality BlackBerry was an inferior product. All industry insiders said so and so did the market. In this case industry insiders say the EDW is superior to a Data mart.

You are right that we should not be lecturing about big data to people who already have a big data implementation under their belt, as far as that goes, but you're saying that a DWH is dead is illogical. Ebay, Amazon, and netflix all use those insights from their unstructured data in conjunction with their Data warehouse. Let's also mention those all have petabytes of unstructured data directly related to their core offering, something very uncommon in the real world of business operations.

@FwMkxJd-xsx: Thank you for contributing. It is very informative and hopefully we can keep the discussion civil and fruitful.

it still doesn't address the fact that a data mart is a second rate solution.

I am not even contending that the data mart is a first rate solution, my point is that for a large number of problems for a large number of customers a second rate solution is perfectly acceptable.

these things are cyclical and there will be a return back to normalization

I wonder why the rest of the market cannot see this. The trend-line for the entire traditional RDBMS market has been down for the last 4 years. This cannot be an aberration.

Your theory that that this is a temporary blip is a fallacy, HBR calls this the BlackBerry argument. For a long time, Blackberry executives would turn up at earnings call and simply tout the superiority of their platform and say "the people" are wrong for preferring Android or iPhones. It will come around. We all know how that turned out.

you should be suggesting that TD focus on the EDW and ignore big data, however you contradict yourself on the point.

I agree with you. In fact, I mention two paths, one is either go down the Big Data route and do it properly or stick to the EDW while providing sane revenue projections and curtailing expenses.

saying that customers are further ahead in the analytics chain because they have a big data implementation is a joke

I believe the point here is that we do not know the analytic needs of customers who have one or two Big Data implementations under their belt. Further or behind is a very subjective thing. The fact is we do not know how to solve their problems. We have blinkers on.

Show me one successful company that built it's success on a big data platform?

Do Google, Apple, Netflix, Amazon, eBay etc. count?

Lets take the case of eBay, it is one of our flagship customers. What is their spend on the EDW vs. the spend on the Big Data platform over the last 5 years? Where is the momentum? There is a reason why eBay is aggressively contributing in the Big Data space and enhancing their capability in this space.

Furthermore, a well-integrated RDBMS is a completely separate thing to an EDW as i am sure you are aware of.

@FwMkxJd-xsx: Thank you for contributing. It is very informative and hopefully we can keep the discussion civil and fruitful.

it still doesn't address the fact that a data mart is a second rate solution.

I am not even contending that the data mart is a first rate solution, my point is that for a large number of problems for a large number of customers a second rate solution is perfectly acceptable.

these things are cyclical and there will be a return back to normalization

I wonder why the rest of the market cannot see this. The trend-line for the entire traditional RDBMS market has been down for the last 4 years. This cannot be an aberration.

Your theory that that this is a temporary blip is a fallacy, HBR calls this the BlackBerry argument. For a long time, Blackberry executives would turn up at earnings call and simply tout the superiority of their platform and say "the people" are wrong for preferring Android or iPhones. It will come around. We all know how that turned out.

you should be suggesting that TD focus on the EDW and ignore big data, however you contradict yourself on the point.

I agree with you. In fact, I mention two paths, one is either go down the Big Data route and do it properly or stick to the EDW while providing sane revenue projections and curtailing expenses.

saying that customers are further ahead in the analytics chain because they have a big data implementation is a joke

I believe the point here is that we do not know the analytic needs of customers who have one or two Big Data implementations under their belt. Further or behind is a very subjective thing. The fact is we do not know how to solve their problems. We have blinkers on.

Show me one successful company that built it's success on a big data platform?

Do Google, Apple, Netflix, Amazon, eBay etc. count?

Lets take the case of eBay, it is one of our flagship customers. What is their spend on the EDW vs. the spend on the Big Data platform over the last 5 years? Where is the momentum? There is a reason why eBay is aggressively contributing in the Big Data space and enhancing their capability in this space.

Furthermore, a well-integrated RDBMS is a completely separate beast to an EDW as i am sure you are aware of.

@thelayoff_td

I know of quite a few (more than 5) truly integrated DWH from first hand knowledge, and I've seen what a multiplier it is for a company once they undertake the effort. Doing it right and maintaining it with professionals yields measures that can be used across the entire organization, and there is no substitute for doing it right.

While you have a point that big data may be a shorter way to a data mart, it still doesn't address the fact that a data mart is a second rate solution. Please feel free to continue spouting off about how people are going to do it anyway, as clearly they are, but these things are cyclical and there will be a return back to normalization one the execs realize what a PITA it is trying to reconcile all of the conflicting reports their differing departments come up with.

The hardest thing to sell to a company is an integrated DWH, but once they see one that is done correctly, they never go back to the rinky dink data marts that only serve one department, can't be tied into the data from other departments. Have you ever been inside of a company that is having to re-issue SEC statements because they were erroneously sourced from a slap-dash data mart? I have, and they are building out a real integrated EDW the next year every time.

Using your own logic, you should be suggesting that TD focus on the EDW and ignore big data, however you contradict yourself on the point. The reality is big data lakes still serve the most purpose when they are integrated with the EDW data that is the core of the business. ie: All that sentimental data is interesting, now show me that demographic segment of our current customer base so we can mail them this offer.

And by the way, saying that customers are further ahead in the analytics chain because they have a big data implementation is a joke. Show me one successful company that built it's success on a big data platform? Amazon you think? They use that big data information in conjunction with their EDW! Again, that's a case where they used the clickstream and other big data to enhance their core offering. ENHANCE. Read and understand: all of it centers around a well integrated RDBMS.

@FwMkxJd-fxo:

TDC becomes a "me too vendor"

This is correct and at most customers that's what we are because our Big Data vision is exactly that, a me-too vendor. We are a Johnny come lately

We either need to commit to it fully or vacate the space,

@FwMkxJd-mno :

So is putting apps on an EDW a good thing or not?

This is a non-sequitur. Your question and the rest of your post presents the EDW as the center of the data architecture with the Apps as an also ran.

I posit that Apps are the centerpiece No one cares if they run on an EDW, a filesystem, a row-store or a column store. That is almost immaterial.

However the TD pitch has always been: get integrated data right, great analytics follows,

My contention is that this has always been more a puritanical and possibly even delusional approach. The number of customers who have real integrated data warehouses can be counted on 1 hand. In practice, most customers have data marts.

What has changed is these data marts can now be implemented on other cheaper, more efficient and far more agile technologies. It a game-changer.

@FwMkxJd-mno :

So is putting apps on an EDW a good thing or not?

This is a non-sequitur. Your question and the rest of your post pre-supposes the EDW as the center of the data architecture with the Apps as an also ran.

I posit that Apps are the centerpiece. No one cares if they run on an EDW, a filesystem, a row-store or a column store. That is almost immaterial.

However the TD pitch has always been: get integrated data right, great analytics follows,

My contention is that this has always been a zealous and possibly even delusional approach. The number of customers who have real integrated data warehouses can be counted on 1 hand. In practice, most customers have data marts.

What has changed is these data marts can now be implemented on other cheaper, more efficient and far more agile technologies. It a game-changer.

#gold post

Success breeds a belief that "our differentiators" need to be considered and baked into the overall sales strategy, subsequent products etc. Case in point, our efforts with the UDA, RIA, Aster, Loom, Rainstor,Teradata Listener and so on. More often than not, these differentiators act as constraints and lead to situations where you're fitting a square peg into a round hole.

There's an opposite problem: TDC becomes a "me too vendor" - just jumping on each and every bandwagon (Hadoop, open-source, in-memory, etc.) - and has no differentiator other than price. Oracle and IBM are on this side of the fence.

I'm confused with this. You criticize Ellison for venturing into ERP, and yet you started with the triviality of leveraging an EDW into other areas. So is putting apps on an EDW a good thing or not?

For all practical purposes, Teradata's current business model is: Convincing people to use platforms they shouldn't be using and then Selling the victims ongoing hacks and services to work around the limitations of the model.

I would disagree - I think the model is more of Teradata being an expensive option when there are open-source alternatives, but in all cases it IS a platform the customers SHOULD be using.

However the TD pitch has always been: get integrated data right, great analytics follows, which is now running sharply in contrast with the Tableau and SAS folks who are just bent on exposing any data at the lowest cost, so they can run visualizations and analysis on them. The truth is there is room for both paradigms in any organization, but TDC tends to be puritanical about its own stuff.

And unfortunately, TDC having an existing list of named customers who "swear" by the technology helps propagate this.