Teradata Corporation (TDC) Q3 2013 Earnings Report, Transcript and Summary

Welcome to the Q3 2013 Earnings Call. My name is Donna, and I will be your operator for today's call. [Operator Instructions] Please note that this conference is being recorded. I will now turn the call over to Gregg Swearingen. Mr. Swearingen, you may begin.

Good morning, and thanks for joining us for our third quarter earnings call. Mike Koehler, Teradata's CEO, will begin today's call; and then Steve Scheppmann, Teradata's CFO, will then provide more financial detail. Our discussion today includes forecasts and other information that are considered forward-looking statements. While these statements reflect our current outlook, they are subject a number of risks and uncertainties that could cause actual results to vary materially. These risk factors are described in Teradata's 10-K and other filings with the SEC. On today's call, we will also be discussing certain non-GAAP financial information, which excludes such items such as stock-based compensation expense and other special items, as well as other non-GAAP items such as free cash flow and constant currency revenue comparisons. A reconciliation of our non-GAAP results to our reported GAAP results and other information concerning these measures is included in our earnings release and on the investor page of Teradata's website, which can be found at teradata.com. A replay of this conference call will also be available later today on our website. Teradata assumes no obligation to update or revise the information included in the conference call, whether as a result of new information or future results. I'll now turn the call over to Mike.

Good morning, and thanks for joining us today. This morning, I'll be providing comments focused on our top line revenue results for Q3 and for the full year, comments on some of the headwinds we are currently facing and the longer-term view of our business. Then Steve will provide more color on the financials for Q3 and for the full year. Total revenue of $666 million in Q3 was up 3% and up 5% in constant currency over prior year. This was below the 10% constant currency growth we had expected for the quarter, as well as for the second half of 2013. Looking at our revenue results by region. First, the Americas reported Q3 revenue of $409 million, which was up 7% and was less than the low-double digits growth we had expected. The U.S. grew revenue 8%, which was in line with our forecast. We had expected an increase in revenue outside the U.S., but had a decline, which is what triggered [ph] our shortfall in the Americas. However, given that our Q3 pipeline and the number of large opportunities was up from what we had going into Q2 in the U.S., we thought we had an excellent opportunity for upside to help mitigate risk elsewhere. Unfortunately, as we moved further into the quarter, we had some customers and industry verticals report economic challenges. This caused some opportunities to not only get pushed out of the quarter, but also out into 2014. Looking ahead to Q4, the Americas has another good pipeline. The pipeline has been growing sequentially each quarter this year, but we are more cautious as to how much of it will close given what happened in Q3. As a result, we have lowered our revenue expectations for Q4 in the Americas to be roughly flat versus prior year. Turning to our international region. Q3 revenue of $257 million was down 2% and flat in constant currency. We were expecting revenue growth to be mid- to high-single digits in constant currency. EMEA's revenue in Q3 grew 10% and 9% in constant currency and within that, Europe's 13% revenue growth was in line with forecasts, but Middle East and Africa had a revenue decline of 19%, which was more than we had anticipated. In addition, APJ had a revenue decline of 21% and 10% in constant currency versus prior year. We had expected APJ to be roughly flat in constant currency versus prior year's Q3, but missed mainly due to shortfalls in China and in Japan. We knew APJ would be challenged going into the quarter by the tough prior year comparison, when revenue grew 16% in constant currency, and in China when revenue grew by more than 50% in Q3 last year. Regarding Q4, we are looking for international revenue growth to be low-single digits in constant currency. We should continue to see good revenue growth in EMEA, but also some revenue declines in parts of APJ. In summary, Q3 revenue came in as expected in the U.S. and Europe, which grew a combined 10%, and 9% in constant currency. And we did lower than expected outside of the U.S. and Europe, where revenue declined 18%, and 10% in constant currency. Total new customer wins were strong in Q3, only down 2 from Q3 of 2012, which was the second highest Q3 ever recorded. Q4 revenue should be approximately flat as reported and up 1% in constant currency. For 2013, revenue is also expected to be flat as reported and up 1% in constant currency. The major contributor to our reduced revenue guidance for 2013 was the number of data warehouse opportunities that have moved out into 2014. With a large amount of that happening in the U.S., where the pent-up demand in our user base that we expected to see in the second half, has not materialized yet. And to a lesser degree, due to reductions in revenue forecast in the second half for the emerging markets, as well as for our Consulting Services. Our win rates remain strong and consistent with previous years. We are aware of the speculation that new and lower-cost technology offerings are impacting Teradata's revenue growth. We believe it is a factor, but a small factor. When cost becomes a top priority in corporations, many of our customers invest more time and resources to optimize their current analytic environments and to also evaluate potentially lower-cost alternatives. We have dealt with environments like this before. These evaluations can have a short-term impact on Teradata by causing delays in purchases. In the event that some of these alternatives do get deployed, we often end up replacing it a couple of years later. Hadoop is a completely different story than the lower performance and perceived lower-cost data warehouse and data mart alternatives. Those of you who were at that Teradata global users group conference last week, probably heard from many of our customers that they are looking at or using Hadoop already. We advocate and support the use of Hadoop. It plays a key role in our Unified Data Architecture because there are specific workloads that Hadoop can handle better than traditional relational databases, including Teradata. Hadoop ingests and stores data very cost-effectively and handles workloads, such as the simple transformations in ETL. On the other hand, Hadoop does not address the mission-critical, complex, business-analytic workloads, which Teradata provides to our customers and excels at. Based on the work we are doing with many of our largest customers, we believe that the likely impact of Hadoop on Teradata is minimal. What we have found is that ETL consumes about 20% to 40% of the workload on their Teradata data warehouses, with some outliers below and above that range. We think that 20% of the 20% to 40% ETL workload being done on Teradata is a good candidate for moving to Hadoop. This means that on average, 4% to 8% of the total workload on Teradata data warehouses could potentially move to Hadoop. These estimates are also in line with the handful of customers that have actually moved some of their ETL workloads from Teradata to Hadoop. However, Hadoop and Big Data are providing benefits to Teradata today. The discovery analytics done with Hadoop have proven difficult and costly to do. With Aster and our Unified Data Architecture, we are enabling mainstream companies to do Big Data analytics quickly and cost-effectively, and doing it with existing workers and existing SQL skill sets. And as a result, our Aster revenue, which more than doubled in 2012, is on track to more than triple in 2013. Furthermore, a lot of the new Big Data, such as click stream, sensor or archival has proved to be a strong fit with our 1700 Extreme Data Appliance, which costs $2,000 per terabyte, works with SQL and is highly available, while delivering extreme scalability into the hundreds of petabytes. And some of the new insights and data, both from Hadoop and our 1700 are landing in our EDWs and adding new workloads there. Net-net, we believe that Big Data and Hadoop are both benefits to Teradata over the long term. Looking at the longer term for Teradata, we have great opportunities for growth and we have a strong competitive position in each of the markets we serve. First, analytics should remain a top priority for corporations. There is a large and growing market for data warehousing and an even faster growing market for Big Data analytics. We are the leader in the data warehouse market today and plan to be, going forward. We are well positioned for Big Data analytics with Aster and our 1700 Extreme Data Appliance. Our Unified Data Architecture, along with our Unity software, consulting and support services will enable Hadoop and other various analytic platforms to work together. And we will continue to add IP for Big Data analytics and Hadoop going forward. Second, the Integrated Marketing Management market, or IMM, is projected to grow high-teens each of the next 4 years. We have a strong competitive position with our IMM portfolio and are the leaders in each of the 3 key components of IMM: Marketing resource management, Multi-Channel Campaign Management and Digital Messaging Center. In addition, we see increased synergy coming from IMM with Teradata and Aster Analytics as companies try to gain competitive advantage through data-driven marketing with their customers. Third, everything in Teradata is available on-premise and now in the cloud. Last week, we announced our Data Warehouse as a Service or cloud offering. We have had a handful of mid-market customers in production for the past year or 2, as well as some large corporations such as Procter & Gamble. And last week, we announced that Netflix is a Data Warehouse as a Service customer. In addition, we announced our Discovery as a Service with Aster, as well as Data Management as a Service for Hadoop. Longer term, we believe we can get incremental growth with our Data Warehouse as a Service, in particular with the mid-market, and with our Discovery Analytics as a Service, and by providing immediate capacity to customers for new analytic demands, along with test and development and disaster recovery. Longer term, we also expect Teradata's business model to have a higher percentage of recurring revenue. In 2013, our recurring revenue will be approximately 40% of our total revenue. Procuring revenue includes our applications, our data warehouse cloud offers, maintenance, subscriptions and our managed services revenue. The remaining nonrecurring revenue is comprised of our Consulting Services, without the managed services portion, which is estimated to be at 22% for 2013; and our data warehouse product revenue, without subscriptions, at 37%. With more of our revenue shifting to recurring revenue and having a large Consulting Services business that operates with a backlog, our revenue should become more predictable and less lumpy over time. In closing, one question we have been getting asked quite frequently is if we can grow double digits again. And the answer is yes, we believe we have the opportunity to grow double digits. First, our recurring revenue will grow double digits in 2013 and we should be able to continue that longer term. Second, our Consulting Services has shown the ability to grow mid- to high-single digits most years. Third, our big data analytics and Hadoop related business is the fastest-growing business we have in Teradata today, and we will continue to add to it. And last, we are confident in our data warehouse competitive position and our ability to take market share in a market that is projected to grow high-single digits annually for the next several years. You add it all up and, longer term, we have the opportunity to grow double digits. With that, I'll now turn the call over to Steve. Steve?

Thanks, Mike. Welcome, and thank you for joining us. Product revenue of $306 million was the same as reported in the third quarter of 2012, up 1% in constant currency. For the first 9 months of the year, product revenue was down 8% as reported and in constant currency. Services revenue of $360 million was up 6% from the third quarter of 2012, up 8% in constant currency. For the first 9 months of the year, services revenue was up 8%, up 9% in constant currency. Within services revenue for the quarter, Consulting Services revenue was $200 million, up 3% and up 6% in constant currency. And Maintenance Services revenue was $160 million, up 9%, up 11% in constant currency. Year-to-date, Consulting Services was up 7%, up 9% in constant currency. And maintenance revenue was up 8%, and up 9% in constant currency. During my discussion today, except where otherwise noted, I'll be addressing margins and expenses on a non-GAAP basis, which excludes stock-based compensation and other special items. Product gross margin in third quarter was 62.7% as compared to an unusually high 70.3% in the third quarter of 2012. When product revenue is flat between years, it is difficult to cover the increased fixed costs, such as an incremental $5 million of FAS 86 amortization versus the prior year quarter. In addition to the incremental FAS 86 amortization, the year-over-year decline in product gross margin was due to a larger mix or number of floor sweeps, which drives higher mix of lower margin product revenue due to the larger hardware versus software components, as well as product cost incurred for future capacity-on-demand at some of our larger customers. Each of the 3 components contributed approximately 200 to 300 basis points decline in product gross margin year-over-year. Capacity-on-demand or COD, is extra capacity typically purchased by the customer within 12 to 18 months. COD enables customers to turn on capacity quickly when needed. However, this can create a timing issue as product cost is recognized upon delivery. Therefore, eroding margins in the current quarter. When customers activate the COD, we record the revenue associated with the added capacity and the gross margin is recovered. In Q3, there were a larger number of customers adding COD than the number of customers who are activating COD that had been installed previously. As a percentage of total product revenue, our 2000 Series appliance revenue in Q3 2013 was only about 10% versus approximately 12% of total product revenue in Q3 2012. For the full year, we still expect the mix of our 2000 Series appliances to be in the historical range of 10% to 15% of total product revenue. Services gross margin in the quarter was 48.3%, up 330 basis points from the 45% in Q3 2012, primarily due to a higher mix of maintenance revenue versus consulting revenue, as well as the higher margin rates in the Consulting business. The improved consulting margins are a result of better utilization of our consulting resources. As a result of the product and services gross margin effects, total gross margin was 55% in the third quarter, lower than the 56.9% in the third quarter of 2012. Turning to our operating expense profile. SG&A expense of $170 million increased $7 million from Q3 2012. SG&A increased primarily due to higher selling expense, largely from the addition of sales territory during the last 12 months, offset by reduced variable incentive-based compensation expense. Let me highlight our incentive comp structure. Everyone in Teradata has a portion of their compensation that is variable. For the non-sales and service associates, it is based on corporate results. When our results are higher or lower than planned, there is an elasticity to our performance-based incentive compensation expense. Clearly, our revised 2013 full year guidance is now under our 2013 plan. And as a result, our overall expense for incentive-based compensation for 2013 is lower than originally assumed and lower than 2012. Correspondingly, if results are better than our plan, higher variable incentive compensation expense somewhat moderates the overperformance on the operating income side of the equation. Research and development in the quarter was $38 million, down from the $40 million in the third quarter of 2012, largely due to lower-variable compensation expense. Total R&D spend in the third quarter, which includes R&D expense plus additions to capitalized software costs from the cash flow statement, less capitalization of internal-use software, was approximately $59 million. This compared to approximately $60 million in Q3 2012. Year-to-date, total R&D spend was $178 million or approximately 21% of our product revenue. As a reminder, these capitalized costs, when amortized, are then added back to the income statement as product cost of revenue, which reduces product gross margin. To provide further color around this activity, we have capitalized approximately $75 million in 2012 and are forecasting to capitalize approximately the same for 2013 and '14. We amortized or expect to amortize $43 million, $51 million and $67 million, respectively, in each of those years. Amortization is expected to increase approximately $16 million in 2014 from 2013. As a result of all these items, operating margin for the quarter was 23.7% compared to 25.6% in Q3 2012. On a GAAP basis, our effective tax rate in Q3 2013 was 25.8% versus 27.3% in Q3 2012. Our non-GAAP effective tax rate for the third quarter was 26.6%, compared to 28.3% for the same period in 2012. This rate differential was mainly driven by the U.S. federal R&D tax credit benefit, which is included in the Q3 2013 effective tax rate, but was excluded in Q3 2012 as the credit had expired at this time last year. We expect our 2013 full year effective tax rate to be approximately 25.5% for GAAP results and approximately 27.5% for non-GAAP results. In terms of earnings per share. Our Q3 GAAP EPS was $0.59 compared to $0.60 in the prior year. Adjusting for stock-based compensation and other special items, which equated to $18 million or $0.11 per share, our non-GAAP EPS was $0.70 per share in Q3 compared to $0.69 per share in the same period last year. Turning to cash flow. Net cash provided by operating activities was $64 million in Q3 2013 versus $107 million in the third quarter of 2012. After $35 million of capital expenditures for software development, property and equipment versus $40 million in the third quarter of 2012, we generated $29 million of free cash flow versus the $67 million free cash flow generated in Q3 2012. However, for the 9 months ended September 30, 2013, free cash flow was $347 million, a $5 million increase from the same period last year. Moving on to the balance sheet. We have $862 million of cash as of September 30, 2013, up $36 million from $826 million as of June 30, 2013. With respect to accounts receivable, day sales outstanding was 75 days as of September 30, 2013, compared to 74 days at September 30, 2012. Total deferred revenue was $400 million as of September 30, 2013, which was up $18 million from September 30, 2012. Deferred maintenance and subscriptions revenue continue to grow at expected rates. Assuming the currency exchange rates at the end of October, we still expect 1 percentage point of headwind from currency for both the fourth quarter and the full year. Turning to the guidance we updated a few weeks ago. We expect reported GAAP revenue for the full year to be approximately the same as the $2.665 billion reported in 2012. Adjusted for currency, revenue is expected to increase 1% year-over-year. In terms of EPS. We expect GAAP EPS in the $2.25 to $2.35 range, which translates to $2.70 to $2.80 on a non-GAAP basis, which excludes stock-based compensation expense and special items. In closing, although we were disappointed with our results in Q3, and our need to lower full year guidance, we believe we are very well positioned in terms of our competitive position and our continued technology leadership. And with that, operator, we are ready to take questions.

[Operator Instructions] Our first question comes from Ed Maguire from CLSA.

Yes. I appreciate your discussions on the change in use of -- workload usage. Have you been able to quantify how much the adoption of the COD program and the shift of roughly 4% to 8% of these ETL workloads may be having on your overall growth rates looking forward? And also, have you also been able to quantify the delays in decision cycles as your customers are really figuring out what they're going to do with Hadoop in context to the broader architecture?

Ed, it's a couple of different questions here. So I'll address them one at the time. On capacity-on-demand, that is basically just a timing issue. So we've been installing capacity-on-demand with customers for several years. And the way it works is, most quarters, we have a balance between those customers activating it and turning it on, and we recognize revenue from that with the amount that we're putting out there, new capacity-on-demand to other customers, within a quarter. So unfortunately, what happened this quarter is we had more capacity-on-demand going out versus the amount that we were turning on. So this isn't related to ETL or Hadoop or anything else like that. That's just -- its own separate variation had occurred in the quarter with capacity-on-demand, and the biggest impact was on our product margins. I wouldn't look at it as a revenue impact. Although if more turned on capacity-on-demand, yes, we'd have more revenue. As it relates to ETL and the impact that's having on Teradata, as far as our results in Q3 and what we're looking at in Q4, it's very, very, very minimal. So we've had a handful of customers that have actually moved some ETL off of Teradata to Hadoop, and it's just a handful, and the amount of capacity that was freed up on Teradata is very, very minimal. As we look at ETL workloads moving off of Teradata to Hadoop, we actually advocate that it makes sense that some of the simpler transformations that we're doing in Teradata go to Hadoop, we expect that to be going over a period of time. So we don't anticipate we're going to run into a quarter where, all of a sudden, all of this ETL work's coming off the Teradata and going into Hadoop. It should be smooth over a period of time, most probably years, a couple of years. The last part of your question, Ed, was -- if you can help me recall.

Just the decision cycles as your customers have been trying to understand what, really, what the different use cases are for Hadoop and that impact on their furthering their commitments to Teradata?

Okay. The impact of Hadoop on our customers' decision cycles, or basically the impact on the extension of sales cycles with Teradata. It has had an impact. I wouldn't categorize it is a huge impact, but as customers evaluate and look to understand what is practical to be done in Hadoop versus what isn't, that does take a little bit of time. Now this has been going on for several quarters -- well, more than quarters, years. And part of the Unified Data Architecture and the benefit of it, is it maps out logically what are the workloads that should be in Hadoop, what are the workloads that should be in Teradata and other environments in the analytical ecosystem. And that's been going on for quite a while. So it continues to be an education process and it continues to take some time. We've made a lot of headway in the U.S. with our customers who have adopted the UDA and that's the -- and basically, the impact is not increased. So we advocate Hadoop and it has a role and it has taken a period of time.

Our next question comes from Brent Thill from UBS.

A number of other tech companies have had similar issues in Asia Pacific, and I'm just curious if you broke down the macro versus execution, and specifically, what you're seeing in Japan as it represents a big bulk of your Asia Pacific revenue?

In terms of Japan, Brent, I don't look at execution as kind of a deal-closing kind of thing. It's more of a business development type of thing. And it's the execution around developing the market and doing development specifically with customers and creating demand in the market and with those customers. And we've been struggling in Japan a number of years. And one of the reasons why we merged our EMEA and APJ regions is to get critical mass -- more critical mass of subject matter experts. These are industry consultants, business consultants, subject matter experts around the various offers we have, use cases and on and on and on, and to create more critical mass, just not in Japan, but across Asia Pacific. So where we're at is we're 3 quarters of a year into landing more resources and building more resources to develop the Japan market. And the good news is we're making progress. However, the progress is not something that we're going to see results short term, this is going to take a while. And like I said, we are making good progress there.

And just back to your comment on returning to 10% growth. Everyone is looking at '14 in our models, I know you're not giving guidance, but when you say longer term, how do you think about the recovery given you're going to have fairly easy comps going into '14?

The -- so when you look at 2014 -- well, first of all, longer term, without a doubt, we have the opportunity to grow double digits. I mean, it's right there laying in front of us. You look at the market growth rates of the markets we're in, they're top priorities, their strategic, high value to customers and we have a strong competitive position in all these markets. So I mean, it's sitting right there, right. So then you get into the short term where we're at. And short term, we're having some struggles, with the macro, with a lot of those types of things. So when you look at 2014 on the more positive side, we've built up a recurring revenue that's growing double digits. So when we exit 2013, we've got 40% of revenue that's growing double digits. That's a good tailwind. On the headwind side, our Consulting Services business is on a trend of slower growth. And when we exit the year, it's going to be single digits and lower-single digits at the tail end of the year. That said, we have the opportunity, of course, to build our backlog and continue to improve that trajectory when we get into 2014. As we sit right now, it's a little bit of a headwind, if you will. So the big variable is what will be our data warehouse hardware-software product revenue. Before I go there, the other thing is, of course, our Big Data analytics and Aster and all of our Hadoop-related things is becoming a meaningful number. It's more than tripling this year. And I think it's closer to the -- almost quadrupling, right. So we got that going into 2014. So the real variable is the data warehouse product revenue. We had a lot of opportunities, a lot. 20 to 30 major opportunities that we're working on in the third quarter, data warehouse opportunities that fell out of Q3. And some of these we were working on, because we're working on all of them, were upside to Q3 and were sitting in Q4, we've had a ton fall out into 2014. We're not saying they're going to surface in the first quarter, but they're actively being worked. And we did not lose one single one of these. So this big drop in our revenue guidance, the vast majority of it is data warehouse product revenue. We did have a decline in our forecast for Consulting Services revenue, but the big drop are these data warehouse opportunities. And we continue to work 'em, we did not lose one of them, none of them turned into little data mart or turned into this or that. They are still there. The question is, we don't know how long and what it's going to take to win the data warehouse product revenue resurfaces. I have to say, as we enter next year we have 600 sales territories, where we had 385 just 4 or 5 years ago. And the emerging markets have been a headwind for us. And emerging markets, we've had extremely high growth in these markets in 2012 and back in 2011, and some of these the first half of 2012 -- or '13. It's been a headwind. Eventually, that's going to turn into a tailwind. But at the end of the day, the variable is the data warehouse product revenue. And at this juncture, we can't -- we don't have a good idea where we're at, unfortunately, until we get to January next year on that.

Our next question comes from Wamsi Mohan from Bank of America.

Could you maybe address 2014. Looking at it, as opposed to products and services and maintenance, through the lens of geographies? It sounds like you had large push outs into '14, particularly in the U.S., where you were expecting to see double-digit growth at the back half of '13. So should we not expect that '14, should rework to some number actually higher than that? And then how worried are you about the sustainability of the strength that you have seen in Europe on a constant currency basis as you go into '14?

When we look at it by geography, I'm trying to be careful with how I answer this. When I look at the prior year comparables and I look at the performance of some of the geographies this year, it makes me want to jump to -- we have a great opportunity to grow everywhere in every geography, and I don't think we want to go there yet. I do think that opportunity is there. When we look at the Americas and in the U.S. -- well, also outside the U.S., and the number of deferrals we've been having, it's been a year now and we had expected the pent-up demand to come here in the second half. At this juncture, I'm hedging a little bit here. I would think we would have the opportunity for some good growth next year in the Americas, just based on that. I can't factually say I can see it. In Europe, I think this is a good question because we have had strong growth there now for 2 or 3 years. And sometimes I've seen, over the years here at Teradata, we get on a long run, sometimes we can run into a timeout after we've had a very strong growth over a number of years. When we look at Europe though, I do see, in terms of the market opportunity and our expansion there and new customer wins, Eastern Europe, Russia and Middle East, Africa, which has been a headwind for us the last 1, 1.5 years. I think we're okay there.

And a quick follow-up. What percent of your installed base is now using capacity-on-demand?

Well, Wamsi, I'd say it's predominantly our larger customers. And on a percentage basis, it's hard to quantify, but predominantly our larger customers.

Our next question comes from Phil Winslow from Crédit Suisse. Philip Winslow - Crédit Suisse AG, Research Division: You commented on some of the verticals that you've been expecting to see good business from, it really slimmed down. And as you did mention, you said you saw a healthy pipeline in Q4. I was just curious what were those verticals that did slowdown? And as you look into Q4, how are you kind of thinking about those verticals in terms of your guidance. And then what are you hearing from the customers there about sort of when they could start spending again on those EDW projects?

Phil, as we got further into the quarter, we had specific customers, as well as some industry verticals that reported challenges with their own financial performance and measures they were taking. The 3 verticals -- 3 of those verticals were the retail, travel and transportation industries. So that was one set. And that's -- I'm talking about here in the U.S. On the other side, those that are doing well -- financial services has been very strong this year and was strong again in Q3. A lot of it was customer specific, but when our customers are feeling pain, we feel the pain. So that's basically what happened here. Philip Winslow - Crédit Suisse AG, Research Division: And also just as a follow-up to that. I know, obviously, the majority of your business, the vast, vast majority comes from existing customers, but I wonder if you could comment on just your new customer acquisition and how that trended?

So the new customer acquisition in the third quarter was very strong. It wasn't a record. It was within 2 of the second highest we've ever had in the third quarter. Year-to-date, it's good. It's not a record, but it's strong. The number of new customers were -- or the types of new customers we're getting are very good. So we're getting a lot in the Global 3000, Fortune 500, Global Fortune 500, relating to customers in some new industry segments that are targeted for growth, growth segments, as you will. Some major utility wins we've had in the third quarter and during the year. We're also getting into a lot of the public sector with helping cities manage traffic and optimize railways and everything else like that. So on the new customer side, it's been -- we've had a very, very good year. The one thing I want to add is, we talked about the 95% of our revenue comes from our user base. And I think sometimes what gets lost is these new customer wins -- you win them, they're in the user base. And those new customer wins have a tendency to grow at a faster rate over the next 3 or 4 years versus ones that have been a customer of Teradata for 10 or 15 years. It's not that the customers that have been with Teradata for 10 or 15 years aren't growing, there's always new data types. We got customers working with sensor data, manufacturers working with it. We got B2C companies working with new data types, coming from the web and social and all the different new channels. However, these new customer wins give us a nice -- it's a higher growth rate for the first 3 or 4 years, and they're very important to the business.

Our next question comes from Katy Huberty from Morgan Stanley.

Given the Global 3000 customers are dealing with constrained budgets and looking for a cheaper alternative, wondering what you're doing to improve your sales focus and reach to help offset this trend, assuming it continues. So specifically, are you investing in more mid-market sales teams? Are you building out a partner ecosystem to improve reach? Are you compensating your salespeople differently for leading with Hadoop or leading with Aster Data? Is there any change in strategies such that if EDW remains sort of flattish growth into next year, you can drive growth in other areas?

Katy, if I kind of segment this thing a little bit, when you look at our major data warehouse customers in the Global 3000, there's a good portion of that, or a majority of that, that really advocates, understands in detail the cost of an Enterprise Data Warehouse versus putting in data marts or whatever. So we don't get impacted much there by the alternative technologies. Then you run in to another subset of users that, in environments like this, an easy Band-Aid to get capacity that might be an initial outlay of capital that's smaller, will put in data marts. This happened to us before in 2008 and 2009. And it's happened previously to that. The good thing is, we have our 2000 Series appliances. And we, actually, compete very well there and we land 2000s. However, over time, when the macro turns around a little bit or the budgets turnaround, if you will, we go back and collapse some of these Band-Aids that have been put out there. In terms of what actions we're doing and everything else like that, the mid-market we've been investing in and putting in territories in the U.S., and we've had some good results. We've done it in a couple of the other major countries around the world as well. The Data Warehouse as a Service, where we have a handful of customers already in the mid-market in the U.S., we'll take that globally. It's a very easy -- it's much easier for those types of customers to implement a very sophisticated data warehouse, minimizing the resources that they would need to acquire that they don't have. So these are all areas that, in and around our core data warehouse business, that we're focused on the increase growth. So there's the Data Warehouse as a Service, there's the Aster Data and Big Data and Hadoop piece of our business that's growing very rapidly. So these are all things that we've got going and that we're focused on.

And will the Aster, Big Data, Hadoop piece that's growing triple digits, will that exit this year close to 10% of product revenue or is it not that big yet? So that's starting on the run rate?

It is not that big yet, but it's becoming a sizable piece of our revenue.

Our next question comes from Bhavan Suri from William Blair. Bhavan Suri - William Blair & Company L.L.C., Research Division: You've have a number of quarters now, you've had good customer wins. Q3 last year, Q1 this year and again Q3. I guess my first question is, when you look at those customers and what they've purchased, are they on track to become sort of EDW customers or do these look like sort of the 2000 Series customers or are they sort of Aster customers. How would you, if you could categorize them a little bit, when you think about the longer-term opportunity there, how that will work, how they fit?

Bhavan, can you repeat the first part of that. We had some distortion here on our end. Bhavan Suri - William Blair & Company L.L.C., Research Division: I'm sorry. Sure. So you've had some good customer wins Q3 last year, Q3 this year, Q1 this year, sort of new logos added were pretty healthy. Of those logos added, what I was trying to understand is how many of those were EDW purchases initially? How many were sort of appliance purchases? Just to get some color of sort of when you look at those customers, how many of those could be equivalent to a large multiyear, multimillion dollar-type deployments longer term?

Okay. The majority of them, Bhavan, are EDW customers, when we saw a new customer. So it's kind of not platform-specific. We might have 40% of the customers start their EDW with a 2000 class data warehouse, but they're using our logical data model, they're integrating a couple of subject areas and so forth. Just like you would do with the 6000 class. And I can't give you a precise answer but generally speaking, these new customers will behave in that they integrate a couple of subject areas, regardless of the class of platform they're on, and they'll get an ROI from it. And then we're working with them on the subject area, the next subject area, the next use case and so forth and so on, and they grow. So a directional answer I can give you is, the new customer wins that we've had over the last year, to your point, which have been very, very good, have been helping to offset, not growing in a lot of our current user base and has added to our results over the past 12 months. Bhavan Suri - William Blair & Company L.L.C., Research Division: Okay. And then as you look at the pipeline, you know you've commented on sort of the positive sequential improvement in the Americas and the strength in Europe. One of the things you started talking late last year was sort of the large deals dropping out, sort of guys adding new subject areas or mature -- expanding with the mature data. And you've sort of seen that pick up a little bit. And now it feels like a lot of these large deals are -- a few of these large deals have been pushed out. If you just look at the large deal portion of the pipeline, big deals over like $10 million, has that improved sequentially or is that sort of flat sequentially or has that declined sequentially?

The number of large deals -- when you look at it more than $5 million, they have been improving sequentially, right, the number of large deals and it goes from Q1 to Q2 to Q3. Actually, with the number of deals that fell out of Q3 and Q4 and gone into 2014, they're actually down a bit sequentially. The large deals more than $5 million. But I want to be clear, they're down a bit, not a lot. And the deals that fell out, once again, they're being actively worked. It's not like the customer says, no, I don't want to do this, it doesn't make sense to do it or I'm buying something else. These customers basically are saying, not this year. That's it. And we'll continue to actively work them. So they are down sequentially, not a lot, but they are down. But once again, in aggregate, we've got a pot of deals now moving forward that -- it's getting bigger.

Our next question comes from Jesse Hulsing from Pacific Crest.

Mike, when you look at your customers that have adopted UDA, and maybe throw in those that have done some ELT offloading, how has spending trended in that group? Has the overall pie grown or shrunk less relative to other customers over the last year? Maybe some color around that will be helpful.

Jesse, I don't know if I can draw a correlation between the customers that are working with UDA, working with Hadoop and Aster and so forth, and as it relates to their overall spending. So I think it's a little bit of a mixed bag. Jesse, please, I want to make sure I got your question right. I think you're asking is, how does spending look with those customers versus customers that aren't doing that?

Yes. The customers that have adopted Hadoop -- I guess more directly, yes. The customers that have adopted Hadoop versus those who haven't, how is spending trending in that group?

Once again, I think it's a -- I can't draw a direct correlation. One new customer, when we had -- not in the quarter, but a couple of quarters ago, had a huge Hadoop environment, no Teradata, and now has a big Teradata environment. But I don't think there's a correlation between the big Hadoop environment and then purchasing a Teradata environment. We have customers, Teradata customers with Teradata environments. The large customers here in the U.S. that we're very familiar with, it's something like 60% of them either haven't [ph] Hadoop implemented in, these are the major customers, where they're looking to implement it. And I don't think there's a correlation between those that aren't doing it right now and those that are doing it.

Great. And as a follow-up, you've mentioned pretty strong growth in Aster and your Hadoop appliance. Given that it's a little surprising that the total appliance revenue is still kind of in that 10% to 15% revenue range. How has your non-Aster appliance business been trending and how do you think that will trend going forward? What's your outlook for that business?

Okay. The Aster business is not in the 2000 class product that we report. Basically, we were tracking the 2000 platform when we released it 4 or 5 years ago. We're basically looking at it as a way to see if there's cannibalization occurring with our EDW. And the 2000, if you will, there's potentially an overlap with what we're doing with Teradata 6000 class. And that's evident with -- a lot of our new customers put in 2000 and then are doing an EDW like thing. So the 2000 class was close to the 10% of revenue this year. Now, that said, the Aster and the Aster platform is growing rapidly, as I said. Also, our 1000 Class Extreme Data Appliance is really picking up and growing fast as well. So there is an increased focus in resources, if you will, around Big Data and Big Data analytics right now, and we're participating and benefiting from it.

The next question is from Raimo Lenschow from Barclays.

The question I had is like, can you talk a little bit about the competition from vendors that are supposed to be not so great anymore, but seemingly having a comeback. I'm thinking about Vertica and Greenplum. There's been some noise about some interesting deals that they have won recently. Can you talk about that in a competitive environment with those players.

We really don't see much of any of these, Raimo. They can land in a data mart or in a department or an organization somewhere, but we really don't see that much of them. I think you may be alluding to, we may run into one head-on or something in a new customer type of situation and we are -- we have an extremely high win rate, it hasn't changed, it's running 80% to 90%, same as before. But we can run into a situation, but there aren't many. I just want to say, we don't compete with them much, but you can run in a situation where we may lose and -- but for whatever reason, right. Customers have different requirements. They could be performance related, they could be availability related, they could be price related, they could be whatever. And most of the new customers that we win, by the way, a majority of them, we lost them before, once or twice over the years. So this is uncommon. That's why I want to be clear, these alternative or newer kind of options, by and large they represent an opportunity to do a small subset of work at an initial low price point for capital expense over time. We normally will come back and replace them, but it just isn't much of an impact on us, I mean, globally. Thank you. That concludes the call. I want to thank you all for joining us here today. And I hope you all have a great day. Thank you.

Thank you. Ladies and gentlemen, this concludes today's conference. Thank you for participating. You may now disconnect.