Cloudera, Inc. (NYSE:CLDR) Q2 2019 Earnings Conference Call September 5, 2018 5:00 PM ET
Kevin Cook - VP, Corporate Development, IR
Tom Reilly - CEO
Mike Olson - Co-Founder, Chairman and Chief Strategy Officer
Jim Frankola - CFO
Michael Turits - Raymond James
Chad Bennett - Craig-Hallum
Greg McDowell - JMP Securities
Tyler Radke - Citigroup
Kash Rangan - Bank of America Merrill Lynch
Karl Keirstead - Deutsche Bank
Josh Baer - Morgan Stanley
Good afternoon. My name is Sheryl, and I will be your conference operator today. Welcome to the Cloudera Second Quarter Fiscal 2019 Quarterly Results Conference Call. All participant lines have been placed in a listen-only mode to prevent any background noise. After the speakers' remarks, there will be an opportunity to ask questions. [Operator Instructions] Please note this conference is being recorded.
Your host is Kevin Cook, VP, Corporate Development and Investor Relations. Kevin, you may begin your conference.
Thank you, Sheryl. Good afternoon and welcome to Cloudera's second quarter fiscal 2019 conference call. We will be discussing the results announced in our press release issued after market close today. From Cloudera with me are Tom Reilly, Chief Executive Officer; Mike Olson, Co-Founder, Chairman and Chief Strategy Officer; and Jim Frankola, Chief Financial Officer.
During the course of this call, we will make forward-looking statements regarding future events and the future financial performance of the company. Generally, these statements are identified by the use of words such as expect, believe, anticipate, intend, and other words that denote future events. These forward-looking statements are subject to material risks and uncertainties that could cause actual results to differ materially from those in the forward-looking statements. We caution you to consider the important risk factors that could cause actual results to differ materially from those in the forward-looking statements in the press release and this conference call. These risk factors are described in our press release and more fully detailed under the caption Risk Factors in our annual report on Form 10-K, our quarterly report on Form 10-Question, and our other filings with the SEC.
During this call, we will present both GAAP and non-GAAP financial measures. Non-GAAP measures exclude stock-based compensation expense and amortization of acquired intangible assets. In addition, we provide a non-GAAP weighted average share count for fiscal 2018. These non-GAAP measures are not intended to be considered in isolation from, a substitute for, or superior to our GAAP results, and we encourage you to consider all measures when analyzing Cloudera's performance.
Additionally, our commentary today and the guidance we provide are under existing accounting standard ASC605. For complete information regarding our non-GAAP financial information, the most directly comparable GAAP measures, and a quantitative reconciliation of those figures please refer to today's press release regarding our second quarter fiscal 2019 results. The press release has also been furnished to the SEC as part of a Form 8-K. In addition, please note that the date of this conference call is September 5, 2018, and any forward-looking statements that we make today are based on assumptions that we believe to be reasonable as of this date. We undertake no obligation to update these statements as a result of new information or future events.
Now, I'll turn the call over to Tom.
Hello everyone. Thank you for joining us to discuss our second quarter fiscal 2019 results. As I will detail shortly, in Q2, we made substantial progress in our product and our go-to-market initiatives, delivering strong financial results in the quarter, and accomplishing many of our goals for sustained success in our market. I am pleased with our execution in the quarter, and look forward to continued improvement in performance as all of our initiatives are fully implemented. It is encouraging to see the changes we're making being validated by customers and partners. In addition, the secular tailwinds in our market remain intact, and demand is strong across our solution set.
Total revenue for the second quarter was $110 million, representing year-over-year growth of 23%. Subscription software revenue grew 26% year-over-year. Operating cash flow was negative $24 million, bringing year-to-date operating cash flow to positive $1 million. One of the measures of our transition that I'm happiest with is the sum of new customers beginning with annual recurring revenue above $100,000, and existing customers graduating to greater than $100,000 of ARR. These customers best reflect our target market and support our business model. We increased this number by 30 in the quarter, exceeding our expectations for a total of 568 customers with more than 100,000 of ARR at the conclusion of Q2.
Recapping our transition plans very briefly. We have made transformative investments in two areas. First, we have refined our go-to-market model to reduce customer acquisition costs and to sustain high net expansion rates. And second, we're investing in innovative and differentiated technology in machine learning, analytics, and cloud to expand our competitive moats. I'll spend a moment to provide an update on advances in each area. Let me begin with our go-to-market strategy. We made very good progress this quarter. At the highest level, we're exhibiting more discipline and focusing on our target market. Recall that we refined our target market to approximately 5,000 named prospects in existing accounts that display the type of consumption and expansion characteristics representative of our best customers.
Next, we have reorganized our field around two new selling roles, account managers who exclusively land new customers in our target market, and those who exclusively focus on expansions and our largest customers driving their success. The early indicators are encouraging. On the new customer acquisition front, we acquired more than 60 new customers. In banking and financial services, we won 16 customers, including Bank of Baroda, Abu Dhabi Islamic Bank, and [Zions] [ph] Bank. In manufacturing, we added seven customers including Tata Steel, Waikai Power, and Leonardo. In the energy space, and up-and-coming vertical, we added five important customers, including Husky Oil, Saipem, and [Gulf Energias] [ph].
In the public sector, we continue to have success both domestically and internationally across intelligence, law enforcement, defense and civilian agencies, adding eight customers this past quarter including Statistics New Zealand, and France's Ministère de la Justice. Impressively, the focus on new customer acquisition in our target market resulted in a more than 30% increase in our initial deal size for the quarter. It is too soon to call a trend here, but the Q2 results suggest a positive impact from better customer targeting and our sales force reorganization.
With respect to existing customers, greater focus and execution discipline in Q2 translated into renewal rates that were above recent averages in a strong expansion bookings quarter. We now have 69 customers exceeding $1 million of annual recurring revenue, representing half of all our software revenue. Also, our in-quarter net expansion was higher than expected for a reported net expansion rate of 128%.
Many of our Q2 expansion transactions reflected typical customer journey, illustrating our land-and-expand strategy. It is notable that most of these customers are moving towards higher value use cases employing machine learning, artificial intelligence, advanced analytics, and often on public cloud infrastructure. For example, Lufthansa Technik is one of the world's leading providers of civilian aircraft maintenance and modification services. Lufthansa began using Cloudera to integrate and analyze data from across their business to improve operational efficiencies. They have since expanded use of the software to run their Internet of Things analytics platform with Cloudera on Microsoft Azure. And they're using machine learning to help airlines predict and prevent maintenance issues, improve flight safety, and reduce costly unplanned repairs. With this new use case Lufthansa Technik is spending over 200% more than they were previously.
ANZ Bank, one of the largest banks in Australia and New Zealand, has run analytic workloads on Cloudera for several years. Today, the bank is using Cloudera's machine learning capabilities for cyber security, supporting identification of anomalous behavior, and assessing its likelihood for being malicious. In addition, they've recently adopted Cloudera Data Science Workbench to accelerate the delivery of new machine learning applications. This quarter's expansion represents a 60% increase in ANZ's annual payments to Cloudera.
Samsung Electronics America, one of the world's largest IT and consumer electronics companies, built a 360 degree view of their customer base using Cloudera. They have expanded their initial use case in a number of ways in order to develop personalized customer interactions and more effective marketing campaigns. With their most recent expansion, this organization has more than tripled their annual payments to Cloudera.
Our strategy is to focus on the needs of the organization that are striving to get insight out of data versus solely managing data. We accomplish this by first leading our customers in machine learning and artificial intelligence adoption. Second, we are disrupting the data warehouse market; and third, we are helping our customers capitalize in cloud adoption. The customer journey is highlighted in the case studies that I shared were all driven by our capabilities in machine learning, analytics, and cloud, reinforcing the importance of our investments in these key industry trends.
I am very pleased with our refined target market, our selling rules, and our success in driving the customer journey. I am also pleased to announce that we have concluded our search for a new sales leader. We will be making an announcement shortly as the start date is being finalized. This executive has the strategic and operational experience to further execute our go-to-market transition and scale our business to match the market opportunity.
Now, Mike Olson, our Co-Founder, Chairman, and Chief Strategy Officer will discuss some exciting new product announcements. Mike?
Thanks, Tom. Hello everyone. As you may know, we made three major product announcements in August, the introductions of Cloudera Data Warehouse, Altus Data Warehouse, and Cloudera Workload Experience Manager or Workload XM. These announcements are significant for Cloudera and the industry as they represent several years of development work and a big leap forward with our analytics business.
We've had great success in helping customers perform self-service business intelligence and SQL analytic workloads with the high performance Apache Impala SQL engine. It's a big deal for customers to be able to run traditional SQL analytics as well as cutting-edge machine learning from the same platform and on the same data without making copies.
Data Warehouse offloading has been one of the most popular use cases on our platform. This was mostly in organic process. We delivered the broad platform, but didn't focus expressly on data warehouse workloads. Nevertheless, our customers and the Gartner Peer Insights Customers Choice Survey selected us as one of the top five data warehouse vendors. With our recent Cloudera 6.0 release, we made substantial progress on scale and performance, and have concentrated on improvements to enable data warehouse workloads in particular. We have more than five years experience supporting SQL analytics with Impala, and that investment coupled with our hybrid and multi-cloud strategy makes a deliberate move into the $16 billion data warehouse market, an obvious choice for us. Here's what we are introducing.
Cloudera data warehouse is a modern data warehouse for self-service analytics. Let me define modern data warehouse and why it's important in this world of exploding data and the Internet of Things. It's a cloud native architecture that can manage 50 petabyte workloads, deliver sub-second query performance on hundreds of thousands of daily reports, and scale to hundreds of computer notes. On each of these metrics, it beats legacy systems by at least an order of magnitude.
Our modern data warehouse is also hybrid. Hybrid is the new normal in data warehousing. Hybrid enables enterprises to optimize how their business works; public cloud for elasticity and scale, multi-cloud for redundancy and choice, and on-premises for performance and privacy. Traditional data warehouses and first generation cloud data warehouses have typically enabled only one cloud choice, theirs. Cloudera data warehouse helps people work with secure data, whether it's in a public cloud or on-premises. We deliver a hybrid cloud data warehouse that works where enterprise works, with the agility, security and governance needed by enterprise IT, and the self-service analytics demanded by business people and enterprise data professionals.
Cloudera is further expanding its hybrid cloud data warehouse offerings with the availability of Cloudera Altus Data Warehouse. This is our modern data warehouse delivered as a service, and it's available on either Microsoft Azure or Amazon Web Services. The all new Cloudera Altus Data Warehouse provides on-demand agility, with hybrid cloud flexibility and performance that first generation cloud data warehouses simply weren't designed to deliver. It maintains lineage and history for transient workloads, critical for governance and compliance, and it eliminates the need to copy data into proprietary object stores, simplifying governance, and making it easier and more secure for people to collaborate and experiment on shared data sets in real-time analytics workloads.
So, as I stated, we've been doing SQL for years. The very best data warehouse offerings on the market though must do more. Traditional data warehouse providers offer real-time workload management for performance and tuning technology. This capability represents a big portion of the value and revenue for these vendors. Last week, we announced our own workload analysis and performance management solution, Cloudera Workload XM, to accompany our modern data warehouse. This is a new SKU and an upsell opportunity for our sales force. Workload XM is an intelligent cloud service that provides guided self-service workload analytics for visibility and control over analytic workloads through their entire life cycle. It affords customers the visibility they need to efficiently migrate, analyze, optimize, and scale workloads running on a modern platform, reducing migration risk, speeding up troubleshooting and improving uptime and resource utilization.
Together with Cloudera Altus Data Warehouse -- I'm sorry, with Cloudera Data Warehouse and Altus Data Warehouse, Workload XM delivers the full featured solution our customers require. Tom said earlier, our intent is to disrupt the data warehouse industry. All of this development is designed to ensure that we gain more share as customers begin to upgrade their legacy data warehouse and data mark technology. This migration is happening now as IoT exploding data and a demand for machine learning and artificial intelligence solutions means that enterprises are rapidly outgrowing their legacy data warehouses. Our modern offerings are well-timed and tailored to compete directly in this major market shift.
I look forward to your questions on our solutions and on this opportunity. First though, Jim will review the financial results. Jim?
Thanks, Mike. Hello everyone. We had a strong quarter across the board, especially concerning the key initiatives of our transition plan. Subscription software revenue was $93 million, an increase of 26% year-over-year. This represented 84% of revenue up from 82% in Q2 of fiscal 2018. In total, revenue was $110 million for the second quarter, representing 23% growth over the year ago period.
Given the go-to-market transition, we want to share more detail than we typically do, and provide additional early indicators of progress. The transition is not complete, but Q2 suggests that we've taken some positive steps. We have steadily grown annual recurring revenue per customer. In the first-half of the year, 67 customers have either started in or have grown to be more than $100,000 of ARR, bringing the total to 568 customers in this class, representing 93% of our software revenue.
Million dollar plus customers represent another significant milestone. As of Q2, we had 69 customers with more than $1 million of annual recurring revenue, representing 51% of software revenue. We are also seeing growth in new customer ARR. In Q2, new customers started at approximately $87,000 versus a historical average of $65,500, an increase of 33%, showing the benefits of heightened focus on our refined target market.
Finally, as Tom mentioned, the 128% net expansion rate in Q2 was better than expected due to improved renewal rates, and increased focus on customer success. Collectively, these measures best reflect our ability to both acquire targeted customers and advance customers along the journey towards increasingly attractive unit economics. As I review the remainder of the income statement, note that unless otherwise stated, all references to expenses and operating results are on a non-GAAP basis. Historical non-GAAP results are reconciled to GAAP results in the previous release issued earlier today.
I would like to highlight the steady progress we are making in terms of operational efficiencies. We are seeing the benefits of our technology being applied to our support organization, our go-to-market processes, and to other functions. Together with appropriate expense controls, these results increased confidence in our path to positive operating cash flow, specifically margins and expenses improved in nearly every dimension.
Total gross margin for Q2 was 76% compared to 73% last year. This was driven by subscription gross margin of 87%, up from 85% a year ago. As to operating expenses, sales and marketing expense was $52 million for the second quarter or 48% of total revenue. This compares to 55% of total revenue in the year ago period. Research and Development was $31 million for the second quarter or 29% of total revenue, improved from 33% a year ago. G&A was $13 million for the second quarter or 12% of total revenue improved from 13% of revenue last year.
Overall operating loss was $13 million in Q2, representing a negative operating margin of 12%. This was a substantial improvement of more than 16 percentage points compared to the year-ago period. Loss per share was $0.08 in the second quarter based on 150 million weighted average shares outstanding, compared to a loss per share of $0.17 in the second quarter of fiscal 2018. Please review the financial statement tables in today's press release for additional information regarding historical and forward-looking stock-based compensation expense and shares outstanding.
Now, turning to the balance sheet and cash flow, we exited Q2 with $458 million in cash, cash equivalents, marketable securities, and restricted cash. Operating cash flow for the second quarter was negative $24 million. Through the first-half of this fiscal year operating cash flow was a positive $1 million, as compared to negative $18 million in the first-half of fiscal '18. As a reminder, Cloudera's cash flows are seasonal, with collections highest in fiscal Q1 and Q4. Capital expenditures were $3 million in the quarter. Total deferred revenue was $284 million at the end of the second quarter, up 23% year-over-year. Short-term deferred revenue was $254 million, up 31% year-over-year.
I will conclude by providing initial guidance for fiscal Q3, and updated guidance for fiscal 2019. We expect Q3 total revenue to be between $113 million and $114 million, representing approximately 20% growth compared to Q3 of last year. With subscription software revenue in the range of $96 million to $97 million, up approximately 24% year-over-year. Net loss per share is projected to be $0.12 to $0.10 based on 152 million weighted average shares outstanding. For fiscal year 2019, we expect total revenue to be between $440 million and $450 million, representing approximately 21% growth. With subscription software in the range of $372 million to $377 million, up approximately 24% year-over-year.
Net loss per share is projected to be $0.53 to $0.50 based on 151 million weighted average shares outstanding. We expect operating cash flow for the year to be approximately negative $35 million. Consistent with the objectives we established at the time of our IPO, we expect operating cash flow to be positive in fiscal Q1 2020, and for the full-year fiscal 2020 as well.
I will now return the call to Tom for some concluding remarks.
Thank you, Jim. It was a very good quarter, and we achieved or exceeded many of our goals. I am increasingly confident that we are on the right path with the initiatives we are executing in our go-to-market and our product strategy. The team remains focused on positions Cloudera to capture more of the long-term market opportunity as reflected in machine learning, artificial intelligence, and analytic workloads moving increasingly to the cloud.
As Q2's performance indicates, with proper focus and continued product innovation we are doing just that. The potential of our three new data warehouse offerings is particularly exciting. With a modern architecture for on-premises deployments and being cloud-native for public cloud infrastructure and Platform as a Service implementations, we believe that we have the right set of solutions for the next phase of the data warehouse industry.
What's more, with Workload XM we are further differentiating our offering by ensuring optimal performance, reducing downtime, and improving utilization across the complete lifecycle of analytic workloads. I look forward to updating everyone as we seek to disrupt this very large market. The team and I remain grateful to our customers, our employees, our developer community, our partners, and of course to our investors. Thank you all.
Operator, we're prepared to begin the Q&A portion of the call. Thank you.
[Operator Instructions] Our first question comes from the line of Michael Turits of Raymond James. Please go ahead. Your line is open.
Hey, guys. Looks like a very strong quarter all around. Two questions, one, can you talk a little bit -- and Jim, you referred to ways in which you were getting operational efficiencies for adding customers and expanding customers, and we did see OpEx below our model, so if can drill down a little bit on how you're accomplishing that would be great. And then I've got a follow-up.
Yes. The operational efficiencies come pretty much across the board. In our support area, we're using out technology to do predictive and proactive maintenance. We've identified the customers that need the greatest amount of help from us that are having a lot of support calls and have programs in place to make them more successful, thereby also driving our support margins up. When we go to the sales line, where we see a big increase, we're starting to see the effects of the go-to-market changes. When we describe these changes, we looked at the data that we were collecting over what drove successful expansion of customers and what sort of behaviors inside of the company were aligned with that, both in terms of the sales approach focusing on new customers versus focusing on expanding customers, as well as the technical support that we provide to those customers. We are seeing the benefits of those organizational changes both in terms of our cost structure and in terms of our ability to renew and expand customers.
And then, Tom, the last quarter you discussed how you were competing in cloud, and the fact, again, you're relatively new there. I don't know if you can quantify for us what the progress is there whether it's in terms of total number of customers, win rates, or just qualitatively in terms of how you're competing against the native cloud offerings like [indiscernible]?
Thank you, Michael. So with every passing quarter, we are competing more often in the cloud, and we're seeing our competitive win rates improve as our sales force builds the skills. Today we have roughly 26% of our customers that have moved workloads to the cloud. Invariably most of these customers are working in a hybrid fashion. So they're using our capabilities on-premise, they're using it in the public cloud, and often in a multi-cloud fashion.
Our competitive differentiators are just that. We deliver enterprise features that large enterprises require in the cloud. We do that in a hybrid fashion, in a multi-cloud fashion. And importantly, leveraging SDX, we give multifunction. Now with Altus Data Warehouse, we can bring to our customers data warehouse capabilities and machine learning capabilities against shared data whether they're doing that on Amazon or Azure, and they can move workloads back on prem.
Great. Thanks, Tom.
Thank you, Michael.
Your next question comes from the line of Chad Bennett of Craig-Hallum. Please go ahead. Your line is open.
Great, thanks. Echo nice job on the quarter. Maybe one quick clarification for Tom, and then a follow-up; Tom, I think you said in your prepared remarks with respect to net expansion in the quarter was higher than maybe the net expansion trailing four quarters, I guess did I catch that right?
I'll let Jim answer that more accurately.
Yes. So net expansion rate for the quarter was 128%, which is an average of the last four quarters. It's a trailing 12-month measurement. It is lower than it was last year, but quite frankly, higher than we expected due to good execution in the quarter.
Okay, and then maybe follow-up, again for Tom. Tom, can you maybe talk about, or Jim for that matter, on I think you believe -- obviously you've raised guidance. You believe you're through the most difficult part of the go-to-market changes in transition. Can you give us a sense for how confident you are that we're past that, I guess? And then maybe talk about; you've introduced a lot of new products in the Data Warehouse new offerings are really exciting, I think. How much of those new products are -- what's kind of a penetration rate into the base or how much are really being reflected in kind of the current run rate? Thanks.
All right, Chad. So as far as our transition, when we announced the transition, these transitions take time to see the full results. I said the expectation that would take roughly four quarters or through the end of our year. The good news is the bulk of the disruptive activity is behind us. And I've seen our changes take route even sooner than I thought. We have work to do in our second-half, but I'm confident with the changes we have in place. And completing all the changes we want to drive by our Q4.
With respect to our Data Warehouse offerings, as Mike shared, we have been working with our customers in analytics for well over five years. We have roughly 800 large enterprises that we have gained tremendous expertise. Our recent work was in Performance in Scale their C6 release, our integration with the cloud data stores, that's three in ADLS, building out our PaaS as a service offerings, and introducing our Workload XM capabilities, all which has us very well positioned to compete in this market. And so we have great expertise with many customers, and we think this will be one of the higher growing parts of our business portfolio.
Great. Nice job, guys.
Your next question comes from the line of Greg McDowell of JMP Securities. Please go ahead. Your line is open.
Great, thank you very much. And it's great to see the progress being made. I wanted to first specifically ask about the new customer initial ASP. I think you said it's up 33%. I was just hoping you could expand a little bit on some of the drivers for the new customer ASPs, and why they're up so much, whether it's -- are customers adopting more Cloudera upfront in terms of workloads or users or more SKUs or is it just more confidence that they're going to be using you guys as their standard platform and expanding quickly. If you could just touch on that a little bit, and then I have one follow-up.
Thank you, Greg. This is Tom. So I'm very excited about us seeing the increase in the new customer ACV. I think there's three reasons driving that. So first is we are really refining our target market in making sure that our field is pursuing opportunities that we believe will fit into our longer-term, not only land but expand model. Second, we've introduced specific new account logo hunters who focus on landing, so they're building betters skills at articulating our value to new customers. And we've elevated our selling beyond IT to line of business around machine learning and analytic use cases. And when you get to the line of business we're seeing higher ACVs in starting points. So, historically, I think we had too many IT evaluations in the platform, and now we're very focused on business outcomes.
Great. And then one quick follow-up for Mike, just expanding a little bit on the three major product announcements, which I agree are going to be very interesting to follow the progress. But I was hoping you could expand, like if some Cloudera customers, as you talked about at the analyst day, or use Cloudera for ETL offload or data warehousing offload. So how should we think about maybe the existing Cloudera customer base using these new products? I mean do they migrate their existing agreements to some of these data warehouse offerings or do they just expand their existing agreements. I was hoping you could you help me think through the mechanics of Cloudera customers that already use you for data warehouse type solutions.
Yes, Greg, let me dive in. So first of all, this isn't a wholesale replacement of the existing platform. This is an evolution of what they've already got. We've improved scale, we've added additional capabilities. We're just delivering more and more complete data warehouse services. We think we should be able to capture a much larger share of the data warehouse workloads in these large accounts. Think of IoT, think of long-term collections of transactional and other data. Data that was just too big to fit into legacy platforms, those are a big forward opportunity for us. And we'll see, we believe, traditional data warehousing tools and analytics running on data at much larger scale on this modern platform.
I'll point out as well; most of our big customers are running hybrid deployments. They want to be able to adopt not only in their datacenters where they're data warehouses have run for a long time, but they want to apply those skills in the cloud. And they want to be able to move them from the datacenter to the cloud and among the clouds, so delivering hybrid and multi-cloud capabilities allow us to compete for new and emerging workloads, and to capture those workloads as they move off premises. Legacy systems weren't designed to handle either the scale or the hybrid nature of those deployments. And we think we're well positioned to capture a good share of that market.
Your next question comes from the line of Tyler Radke of Citigroup. Please go ahead. Your line is open.
Hi there. And I apologize if my question has been asked already; I'm jumping around calls here. I was just hoping you could talk about just where we're at in terms of the return to strong net expansion rates back to where you want to see them. Are we past the peak point of disruption? And what's your expectation on timing in terms of when you think strong bookings will rebound?
Tyler, thanks for the question. This is Tom. So first off, our transition is to achieve two things. To improve our customer acquisition costs and to drive sustained high net expansion rates. Historically they've been in the 120% to 150%. So that's the transition we're executing on to achieve that. The disruptive part of our transition is behind us. And I'm pleased in our second quarter of that transition we had strong results, i.e. early indicators that these changes are having affect. And that's very rewarding for, not only me to see, but for the whole company to see. We do have more work to do. So one of our results is we're introducing more outcome-based services to help our customers through their journey, which will drive those expansion rates.
We're doing that via our own services, but also through our partners. We're implementing those capabilities in the second-half. But I expect most of our transition work to be done by our Q4, and hopefully gives you a sense of where we're at.
Yes, that's great. And this is a follow-up. Could you just update us where you're at just in terms of the maturity of the -- kind of the new go-to-market approaches that you outlined around analytics, cloud, and in machine learning, just where we are in terms of maturity or even like a revenue contribution?
I don't think I'm prepared to break out the revenue contribution. But for the broader audience, we're shifting a bunch of our R&D and innovation work into three areas. Machine learning, it's our intent to lead large enterprises to machine learning and artificial intelligence. We're investing in the data warehouse market, and in intend disrupt -- aggressively disrupt the data warehouse market. And third, we are capitalizing on cloud adoption by our large enterprise customers. In each of those areas we're making significant progress. This call we dedicated to three major releases in data warehousing, Cloudera Data Warehouse, Cloudera Altus Data Warehouse which is our managed service offering for data warehousing, and Workload XM. That's a reflection of how quickly we're shifting to helping our customers get insights. We're making great progress in machine learning and data science. We will be giving some more updates probably in our next call.
And then with cloud adoption, we're very excited that the use of cloud by our customers is the fastest growing part of our business. We have 26% of our customers moving workloads into the cloud and running workloads in the cloud. Given that our customers are large enterprises, the majority of them are working in hybrid environments. And increasingly we're seeing multi-cloud deployments by our customers who want to avoid cloud lock-in, and they value the portability. So these three areas we're doing extremely, extremely well. They're all underpinned by our core platform. And we also released Cloudera 6.0 which dramatically improves our performance and scalability.
Your next question comes from the line of Kash Rangan of Bank of America Merrill Lynch. Please go ahead. Your line is open.
Hey, guys. Thanks for the update and congrats on the stability that you're seeing in the business. Tom, I had a question for us. A couple of quarters back you talked about how the company -- the sales force had been particularly successful in getting a deep degree of success across a narrower section of the funnel. And that that necessitated a change in the sales approach to have a dedicated farmer organization versus a hunter organization, and therefore you went through the GM unit -- a GM structure for the different products. I'm curious where do we stand today with respect to broadening out the funnel and making this a broader market opportunity, and that maybe there is certain things that needed to have been worked in, they have already been worked through, in order to make this move of a mainstream because if you are to build, I am sure that this is the case, a long-term enduring $1 billion plus revenue company. One would assume that this traction is more broad-based and not just deeper across narrow set of customers that seem to get to value proposition. So what is happening out of more fundamental funnel entry point that will cause this market to be more mainstream for the likes of Cloudera and maybe even others? Thank you so much.
Thank you, Kash. You covered quite a bit of different topics there, but let me see if I can address it this way. Right now, we want to have greater discipline in how we go to market, and who are the customers we target, and how we accelerate their maturities in their journey. We have six stages of maturity that we measure, and we are looking at outcome-based services and relationships and technology capabilities that advance some through that.
Now, we also believe that we can introduce more capabilities to expand the size of our relationships with any existing customer. So we have done that with our new data warehouse offerings. And as our customers want to move data warehouse workloads to the cloud, we now have an offering to capture that. Our general cloud offerings allow us to grow that relationship.
One of the other things in building out a -- bringing discipline to our target market also requires us to invest in our channel. And I didn't cover that in our earlier comments, but we are also putting in place our channel to allow us to address a broader market at a lower acquisition cost and support cost. And so, we put in place our global distributors, we are making investments in our channel to be able to provide better support and services to a broader market. And then ultimately, our cloud capabilities, while our cloud capabilities today are geared towards capturing our large enterprise's desire to move workloads in the cloud, we are designing that also at the right time for us to be able to move down market and address the market in a larger way. So Kash, I hope that gives you a sense of how we are looking at broadening our market through the channel, through cloud, but our near-term focus is to stay disciplined and targeted into large enterprises.
Completely -- yes, I understand the near-term imperative, balancing that versus what needs to be done long-term. I completely appreciate that, thank you so much guys.
Your next question comes from the line of Karl Keirstead of Deutsche Bank. Please go ahead. Your line is open.
Thank you. Maybe this question is suited for Mike. I am just interested in your -- what I think is bigger focus on disrupting the legacy data warehouse industry, and I have got two questions on that. One is that market is fairly dominated by the likes of Oracle, Teradata and IBM with Netezza and other offerings, and I am just wondering whether you think any -- one of those vendors might be more vulnerable for displacement than others. And then secondly, I know Snowflake is also looking to disrupt the legacy data warehouse industry, and I am just curious maybe that's a long topic to address on a call like this, but are there a couple of albeit differences in your approach or perhaps the answer is the opportunity is so big that there is certainly room for a couple of vendors to go after it? Thank you.
Thanks, Karl. So I will answer those in the order you asked. So first of all, clearly this is a $16 billion market, there are some monster players in it; Oracle, IBM and Microsoft, decades in the market. We bring the market a modern platform, and I will highlight some of our strengths there. You asked whether any of those guys is ready for disruption, certainly as [indiscernible] noticed that Netezza is end of life, right, there is a great opportunity to go capture those workloads, modernize that infrastructure, help enterprises that have been using Netezza, move those workloads into the cloud if they want to do that, and give them much larger scale and powerful analytic program. We are running a program that we call Step-Up, and precisely at those Netezza migrations.
You call out also interesting new market entrants, and Snowflake is a great company and great product; really four sharp differences that I would hit for you. So number one, Cloudera is hybrid. Our customers want to run on-prem and in the public cloud we let them do that, cloud only vendors like Snowflake don't. We are zero copy. And that sounds like nerds, but look, if you have got data in the public cloud, you are storing it in Amazon S3 buckets or in ADLS, you don't want to copy it into a proprietary store where it's locked up and only available to Snowflake. We operate natively on those stores. So you don't need to make multiple copies of your data.
Closely related, most of our customers want to do more than one thing with their data, yes, they want to run petabytes, SQL queries over structured data, but they also want to train machine learning models, they want to analyze that data using a whole bunch of statistical tools. We've got a multi-function platform that lets them do that. And fourth, remember, we're aimed at large enterprises. Our customers need secure, compliant, governed, managed systems, they've got to be able to set and enforce security policies and track usage, not only over their SQL queries, but also over their machine learning and their statistical analytic capabilities applications, and we're able to deliver that shared data experience across that spectrum of workloads. So I think we have got a strong offering in the market. We're going to see a lot of customers who need to step off of their legacy platforms. IoT data is arriving way too fast and at a much larger scale than those vendors can handle. We think we're going to win our share.
Got it; very comprehensive answer, thank you, Mike.
Your next question comes from the line of Mark Murphy of JPMorgan. Please go ahead. Your line is open.
Hi, good afternoon. This is Matt Cass [ph] on behalf of Mark Murphy. Thanks for taking our questions. Can you talk about the level of your sales rep attrition or retention you've seen given some of the changes you've been making, and how amenable have those reps been to changes in any comp plans? And then, separately, so with artificial intelligence and machine learning in their very early stages, have you noticed any meaningful mistakes in the market with respect to how some companies have approached the deployment of artificial intelligence and machine learning? And how are you making sure that your customers hit the nail on the head when they deploy AI and machine learning?
Okay, Matt. Hey, you get two of us answer it, so you have Tom, I'll take the first part about the sales forces, and Mike will cover the second part. So, with any transition, you're going to -- it's difficult to take a company through transition, but I'm very pleased, one of our strengths and our culture is we have great transparency and lots of communication. And so, we have been regularly communicating with the company on where we're at in the transition, what to expect. And I'm quite pleased actually with our success of retaining our best talent. Our attrition rates are higher than they were a year ago, but that was a year of our IPO, but if you look over the past three years, our attrition rates are better than what we've had on average. So that's very positive.
And when it comes to the field, Q1 was very disruptive for the field because roughly 40% of accounts change hands and territories change, but we're able to do it by using our own data direct our sales force where they get the best returns and where they should be spending their time, and it's resulting in their success. So I think the sales force is very supportive of where the transitions we're making because their success is our success. And then, we've been investing in their skills to sell into a line of business around machine learning and data warehousing. And so, we've got strong enablement there to help them as well.
Hey, Matt. This is Mike Olson. I'm going to dive on the second part of your question. This call, we talked a lot about the data warehouse announcements that we've made because we're really pleased and proud, and we think that's a big deal. I expect probably next time we'll talk more about what's going on in machine learning, but just to hit the specific issue you addressed, yes, machine learning is new and is moving fast and it's confusing to people. We offer an open platform. We were the first vendor in the market to adopt Apache Spark. We can take advantage of all the open source algorithms that are coming out of places like Facebook and the academic research community. Unlike, for example, IBM Watson or Palantir, Black Box systems that are opaque to customers that they can't understand, so simply that openness is helpful.
We also -- you may recall, introduced more than a year ago now, our data science workbench offering, and that's aimed at making data scientists and analysts much more productive. So, software tools to help them develop those applications. Finally, just under a year ago, we brought in a new capability at Cloudera, Fast-Forward Labs is our research and strategic consulting team that is able to deliver advice on state-of-the-art and where the market is going, and how these techniques can be applied to business problems, how organizations need to reorganize themselves in order to turn these AI tools into real actionable business insights. So again, we'll probably spend more time on that later, but I think we're well-positioned there. We've certainly been engaged in the market and thinking very hard about how we lead.
Your next question comes from the line of Sanjit Singh of Morgan Stanley. Please go ahead. Your line is open.
Hi, this is Josh Baer on for Sanjit. I had a few questions on gross margins, which outperformed this quarter. If you could review some of the dynamics that led to subscription gross margins improving two points year-over-year, and anything to keep in mind on the services margins line looking forward as you work on efficiency?
Yes. This is Jim. The subscription gross margin was driven by the factors I touched upon earlier. So, we use our own technology very comprehensively across Cloudera within our support organization. We will proactively look at customers' usage of the technology, and in essence, predict when failures may occur, contacting the customer ahead of time and preventing that failure. That leads to great customer satisfaction. It also leads to a much lower cost.
In addition, as we are implementing these new go-to-market changes, we are better identifying, I would say, customers and it makes sense to invest in that are early in their journey, and quite frankly, struggling, have low margins. And we have seen a significant increase in profitability of our lowest margin customers over the past few quarters. So those two areas have really driven the subscription improvement.
In terms of services gross margins, they're about 16%, that's roughly where they were last year. Movement you see in any given quarter on services gross margin is entirely due to we are still under the 605 accounting standard, we don't have VSOE, and therefore you will have ebbs and flows of revenue recognition. So there's no real news on the services gross margin piece.
Thanks. And as the new data warehouse products and Workload XM increasing adoption and mix, will that have any impact on gross margins?
Good question. I expect the data warehouse to be similar to our existing gross margin trends, which is increasing over time. To the extent, customers are adopting that in the cloud, we expect our gross margins in the cloud to be slightly higher than those on premises. We price essentially the same whether they are in cloud or in premises, but the cloud deployments are much more consistent, fewer failures easier to diagnose, so the cloud story will be more of a trend on gross margin improvement. Individual products, I don't expect to see much change in that one.
There are no further questions at this time. I will now turn the call back over to the presenters for closing remarks.
Well, thank you for all of you that had questions for us and spending the past hour with us. This is a very exciting time for Cloudera. We are making the right investments both on product and go-to-market to have sustained competitive advantage to build-out our modes, to lower our customer acquisition costs, to sustain our high net expansion rates. I appreciate you all understanding our journey, and thank you for the time today. We look forward to reporting back to you in a quarter from now with even better news. Thank you.
This concludes today's conference call. You may now disconnect.