Pure Storage Inc (PSTG) 2017 Q3 法說會逐字稿

Good afternoon. My name is Kelly, and I'll be your conference operator today. At this time, I would like to welcome everyone to the Q3 FY17 earnings conference call.

(Operator Instructions)

Thank you. Liz Lemon, VP of Finance and Development, you may begin your conference.

Good afternoon. Welcome to Pure Storage's Q3 FY17 earnings call. Joining me today are CEO Scott Dietzen and CFO Tim Ritters.

Before we begin, I'd like to remind you that during this call Management will make forward-looking statements which are subject to various risks and uncertainties. These include statements regarding competitive, industry, and technology trends; our strategy, positioning, and opportunity; and our products, business and operations, including our revenue and margin guidance, operating model, and growth prospects.

Our actual results may differ materially from the results predicted, and reported results should not be considered as an indication of future performance. A discussion of risks and uncertainties relating to our business is contained in our filings with the SEC, and we refer you to these public filings.

Also during this call, we will discuss non-GAAP measures in talking about the Company's performance. Reconciliations to the most directly comparable GAAP measures are provided in our earnings release.

This call is being broadcast on the web, and is being recorded for playback purposes. An archive of the webcast will be made available on Pure Storage's Investor Relations web site for approximately 45 days, and is the property of Pure Storage. With that, I'll turn the call over to our CEO, Scott Dietzen.

Thanks, Liz. Good afternoon, and thanks for joining the Pure Storage third-quarter earnings call. We have a special guest today, our VP of Products, Matt Kixmoeller, who will add color in our Q&A. As usual, there are new blog posts on our website, with more detail that we encourage interested investors to read.

Pure is winning in the market place, because our storage platform is enabling customers to derive unique value from their data. Modern analytics, including machine learning, fundamentally depend on faster access to larger and evolving data sets. Only Pure combines the scale, density, performance, and low cost of ownership necessary to develop next-generation, data-intensive applications, as well as the bandwidth to connect that data to all applications.

Legacy and commodity storage alternatives are too slow, too siloed, and too expensive to meet the demand of the data-driven cloud or enterprise. Pure has the right platform for our customers' most strategic data at the right time.

Our financial performance confirms our unique value proposition. Q3 revenues grew 50% year over year to $197 million, 3% above the mid-point of our guidance. We also saw improved leverage. Non-GAAP operating margin was 11 points better year over year, and five points ahead of the mid-point of guide.

In the quarter, Pure added more than 300 new customers, increasing our customer count to more than 2,600. Cloud deployments continue to account for more than 25% of our sales. Pure now serves over 400 software-as-a-service, infrastructure-as-a-service, and consumer cloud customers.

Meanwhile, Pure's international business remains robust, growing 70% in the latest quarter. Repeat purchase rates also remain strong, with our top 25 cohorts spending $12 more over the next 18 months for each dollar they spend initially.

We added many noteworthy new customers in Q3. I'll touch on just a few. Hyatt Hotels is relying on Pure to change IT processes around app development, deployment, and management. We also added the commercial real estate firm Cushman and Wakefield; wireless service provider Telecom Italia; CallidusCloud, a provider of SaaS, sales, and marketing software; Bill.com, a SaaS business payments company and Laika, an academy-award-winning animation studio.

Competition in the enterprise remains fierce, as our legacy competitors fight to protect their installed bases. Nevertheless, our win rates held strong, and are on par with where they were a year ago. As a result, we now count more than 100 of the Fortune 500 as customers, up roughly 10% from last quarter.

We win for multiple reasons. We have the most innovative products. Gartner placed us in the Leaders Magic Quadrant. We have the happiest customers. Statmetrics calculates our net promoter score at 83.5, among the top 1% of all business-to-business companies. We have the most reliable platform. FlashArray M delivered six nines of up time in its first year. That's less than 32 seconds of down time per year, and we did so uniquely without maintenance windows.

Pure was a pioneer in All-Flash arrays, and we are ahead of the competition again with MBME. MBME offers the potential for 100-volt increase in performance through processing many requests simultaneously, rather than one at a time. By extending MBME across the network, petabytes of data in Iraq can be accessed as if they were a local resource, enabling shared storage to offer better performance and better economics than the server local storage, or DAS model, popular in many clouds.

We believe All-MBME is going to prove as big a disruptor as All-Flash. Of the mainstream storage solutions, only Pure has been shipping MBME technology since 2015; and only Pure is ensuring customers' current investments are protected en route to that All-MBME future. Our competitors face expensive re-writes of their software, while their customers face having to buy all new storage, and migrate their data.

At the same time, we are ramping for the full launch of FlashBlade, our new product line for unstructured and big data. Customer feedback reaffirms that FlashBlade will revolutionize capacity-oriented storage the same way FlashArray re-wrote the rules for performance storage.

While revenue from FlashBlade won't be material this year, GA remains on track, and customer enthusiasm is sky high. Some recent examples include CUProdigy, a provider of private cloud solutions for credit unions, who is using FlashBlade to elastically scale operations transparently to their applications; and Paylocity, who is employing FlashBlade to support its SaaS payroll and human capital management software.

Pure's growth has been supported by strong channel partners. In the latest quarter, 76% of our net new logos came through the channel. That expanding partner leverage, which now includes key global systems integrators, is essential to extending our reach. We are also seeing strong traction for our FlashStack partnership with Cisco, delivering converged infrastructure that combines best-of-breed servers and networking from Cisco with Pure's best-in-class storage. FlashStack is growing triple digits year over year today.

To sum up, we are very well satisfied with our Q3, and even more excited about the road ahead. We are tracking toward our long-term goal of building a profitable, multi-billion-dollar revenue Company that will lead the global storage industry.

We are winning in cloud; we are winning in the Enterprise. Demand for flash array remains strong, with MBME set to tip the scales further in our favor. FlashBlade is ramping in a growth market that moves in our direction as flash costs drop. We are providing the data platform that uniquely empowers our customers to maximize the value of their data.

With that, I'll turn the call over to our CFO, Tim, to provide further detail on Q3, as well as our guidance for Q4. Tim?

Thanks, Scott. Q3 was indeed an excellent quarter. I am very pleased with strong execution around our operating model, which calls for consistent top-line growth, balanced with meaningful year-over-year improvements in operating leverage. We are making great progress towards becoming a profitable multi-billion-dollar revenue Company, as Scott mentioned.

Before I dive into Q3 specifics, please note that the gross margin, operating margin, OpEx, EPS, and free cash flow numbers I will use are non-GAAP, unless otherwise noted. A reconciliation of these non-GAAP metrics to the GAAP comparables is available in our press release, and in our earnings slide deck, which are available on our website at Investor.purestorage.com.

As Scott said, Q3 total revenue grew 50% year on year and 21% quarter on quarter, to a record $197 million, which is 3.1% above the mid-point of our guidance. Product revenue in Q3 grew 41% year on year and 23% quarter on quarter, to $160.5 million, driven in part by over 300 new customer additions, and excellent demand from our existing customer base.

Some of you have asked for more insight into our repeat business; so beyond the Top 25 that we have reported in the past, I am happy to say that across our entire customer base, our repeat business is thriving. For every $1 that our customers spend initially, they spend an average of more than $2 additionally within the next 24 months.

We continue to ramp our FlashBlade business, which had solid quarter-on-quarter revenue growth. We are excited about the long-term opportunity it represents, though as expected, the revenue contribution was immaterial to our Q3 results.

Support revenue in Q3 grew 105% year on year and 13% quarter on quarter to $36.5 million, driven by revenue recognition of ongoing support contracts. We continue to drive loyalty among our customers, demonstrated by a strong customer retention rate in the mid-90% range, and our industry-leading NPS score of 83.5.

Looking at Q3, from a geographic perspective, 77% of our revenue came from the US, and 23% from international, compared to an 80%-20% split in the prior fiscal year. We are seeing notable success in regions across the board.

Q3 total gross margins of 65.5% improved 3.8 percentage points year on year. Total gross margins declined 0.8 percentage points quarter on quarter. We continue to be in the range of our target long-term model of between 63% and 68% for total gross margin.

Product gross margins of 66.0% improved 2.9 percentage points year on year, and declined 1.4 percentage points sequentially. The year-on-year improvement reflects the shift to the higher-capacity, higher-gross-margin, FlashArray M Series products. Consistent with our Q3 guidance, the Q-on-Q decline was driven primarily by lower margin FlashBlade shipments, as we continue the manufacturing ramp of this new product.

Support gross margins of 63.2% improved 10.4 percentage points year on year, and 1.2 percentage points sequentially. This is driven by our expanding customer base, and the deferred support revenue associated with increasing product revenue. We continue to drive operational efficiencies within our support organization as we scale.

Moving on to expenses, as Scott mentioned, we continue to drive significant operating leverage year on year and sequentially, while making strategic investments in the business. In Q3, R&D expense was $46.3 million, or 24% of revenue, versus $34.9 million, or 27% of revenue in the year-ago quarter. R&D expense in absolute dollars increased 33% year on year, driven primarily by continued investments relating to our next generations of both FlashArray and FlashBlade hardware and software.

Sales and marketing expense of $82.5 million represented 42% of revenue in Q3, versus $59.2 million, or 45% a year ago. Sales and marketing expense in absolute dollars grew 39% year on year, as we invest thoughtfully in sales and marketing head count and programs. We remain focused on capturing the current market share opportunity, balanced by careful attention to sales productivity.

G&A expenses were $19.6 million, or 10% of revenue in Q3, versus $15 million, or 11% of revenue a year ago. G&A expenses spend grew 30% on a year-on-year basis, driven primarily by increased spend on legal, outside services, and finance, as we scale and continue to grow our business. Total head count at the end of Q3 was over 1,650, up from over 1,600 at the end of Q2, and up from over 1,200 a year ago.

Turning to operating margin, we continue to make excellent progress in our drive to profitability, and toward our long-term operating margin goal of between 15% and 20%. For Q3, non-GAAP operating loss was $19.4 million, or negative 9.8% of revenue, compared to non-GAAP operating losses of $28.1 million, or negative 21.4% of revenue, in the year-ago quarter. This represents an 11.6-percentage-point improvement in operating margin year on year, and 9.5-percentage-point improvement sequentially.

We are making significant progress in making this fiscal year our turning point in terms of absolute operating losses. Operating losses in absolute dollars have decreased by $27 million year-to-date over the same period last year, 23% ahead of our flat year-on-year operating loss guidance from the beginning of the year.

Our non-GAAP net loss for the quarter was $20 million, or negative $0.10 per share. This compares to the year-ago quarter non-GAAP net loss of $29.1 million, or negative $0.18 per share. The weighted average shares used for the per-share calculations were $195.8 million and $164.9 million, respectively. For those of you comparing to our Q3 FY16 period, please note that the share count numbers used to calculate net loss per share in that period assume the conversion of all of our preferred stock at the beginning of the quarter.

Please note that for Q3 FY17, our non-GAAP operating loss, non-GAAP operating margin, non-GAAP net loss, and non-GAAP net loss per share also exclude a $30-million, one-time cash charge related to a legal settlement.

Moving on to the balance sheet and cash flow, we finished the October quarter with cash and investments of $518 million. Our free cash flow was negative $35.8 million, or negative 18% of revenue, compared to negative $13 million, or negative 10% of revenue in the year-ago quarter. Please note that this includes $5.4 million of an impact related to our employee stock purchase plans. Excluding this amount, free cash flow would have been negative $30.4 million, or negative 15% of revenue. For Q3 FY17, our free cash flow also excludes the $30-million, one-time cash payment related to the legal settlement I mentioned above.

Let's now turn to our Q4 guidance. We are experiencing seasonality consistent with last year -- strong top-line growth in our second half, driven by the sales investments we made in our first half. As a result, our operating margin tends to be stronger in the second half relative to the first half.

With that in mind, for the fourth quarter of FY17 we expect revenues of between $219 million and $227 million, driven by several key dynamics. We are entering a seasonally strong period for our business. We have excellent repeat purchase rates across our entire customer base. We are now one of the top 10 vendors in storage, and we are seeing many more deals today than we were a year ago.

Our partner momentum remains strong, fueled by a geographically diverse combination of partners who are increasing their commitments with Pure. Perhaps most importantly, we have a superior storage portfolio, moving strongly into under-structured data, with a software advantage that customers value.

With regard to revenue contributions from FlashBlade for this fiscal year, I'll remind investors that while we are excited about the long-term opportunities for this product, revenue contribution is not expected to be significant in the near future.

Turning to gross margin, we expect Q4 non-GAAP gross margins in the range of between 64% and 67%. With the initial manufacturing ramp of FlashBlade behind us, our overall gross margin has now stabilized. This performance fits within our long-term model of between 63% and 68% gross margin, and we remain focused on execution in this area. We expect Q4 non-GAAP operating margins of between negative 5% and negative 9%, as we continue to drive absolute operating losses lower year on year.

In summary, our Q4 guidance implies for our current year that we will drive nearly 65% year-on-year revenue growth, while cutting our absolute operating losses by over 20%. We continue to manage the business towards sustained positive free cash flow in the second half of calendar year 2017, as we drive consistently toward our long-term operating margin target of between 15% and 20%.

With that, we will open the call for questions. Operator?

(Operator Instructions)

Your first question comes from Steve Milunovich from UBS.

All right, thank you very much. Last quarter I think you talked about pushing expenses forward into the October quarter. I don't really see that. R&D certainly didn't increase much. SG&A and G&A seemed up with what I expected. Did that not happen? Is that yet to come, or is that really embedded in those numbers?

Steve, this is Tim. What you're seeing really in Q3 is the leverage that we talked about here in the earlier remarks. That's dropping -- a lot of that is dropping to the bottom line, which we expected. I'm always on the lookout -- I think we're always on the lookout here for good quality talent in R&D, but we're happy where we are right now with the OpEx for Q3, and then the implied guide out into Q4, as well.

Okay. Then Scott, I wonder if you could talk a little bit about your position relative to the hyper-converged options, particularly Nutanix. Do you compete against them, and what do you view as your advantages?

Thanks, Steve. We do see hyper-converged in low single percent of our -- single-digit percent of our deals. Pure is competing for the most strategic data that our customers have. Where we see hyper-converged's sweet spot, if you look at the report we've actually cited on the blog that 67% of by their measure of the hyper-converged deployments in the world are for remote and branch offices in VDI, which is not the most business-critical data in the data center.

The larger data centers very much tend to specialize their infrastructure into multi-tier. There is an application and compute tier, and then there's a separate data and storage tier. That combination provides additional scale, additional performance, and actually better economics, because you can refresh and scale the compute and storage tiers independently of one another. Maybe I'll pass the baton here to Matt Kixmoeller, who can provide some more color.

The only thing I would add is that I think it's important to remind us that hyper-converge was born in the era of disk. The whole base architectural assumption was there would be lots of cheap disks you could throw in servers and get a storage layer for free.

I think the challenge with that thinking is it's ultimately quite a compromised storage layer. It doesn't have the performance, it doesn't have the scale and resiliency that enterprise customers expect. We really see this playing in the use cases that Dietz mentioned earlier -- remote office, VDI-type use cases, as opposed to the mission-critical, data-intensive analytics and business applications that Pure tends to serve.

I would add for cloud and enterprise, it's where we're less likely to see hyper-converged infrastructure.

Great, thank you.

Thank you.

Your next question is from Aaron Rakers from Stifel Nicolaus.

Yes, thanks. A couple real quick questions, if I can. First, I want to touch on the gross margin. You talked about getting through the initial ramp of FlashBlade. I think that was expected to be about a 100-basis-point negative impact this last quarter. As we look at the guidance, is that 100 basis points lifting out? In that, how do we think about the current trends that we're seeing on NAND flash pricing as it relates to your gross margin in the current quarter or even thereafter, as we've seen continued price increases out in the market? Then I do have a follow-up.

Yes, Aaron, this is Tim. I think on your observations on FlashBlade, you're absolutely correct. We're through that first phase now. FlashBlade margin's where we want to be. Now FlashBlade will become a bigger part of the overall revenue, so that also factors into gross margin, and that's why we guided the way we did.

As it relates to the overall cost of goods sold or the NAND pricing you talk about, we've seen some NAND increases, but I think as we've said in calls before, and I'll reiterate here, we've got strategic supplier relationships across many of the manufactures that have allowed us to go out and source the most cost-effective NAND for the job, which we've done.

We've modeled all of that into our pipeline. I'm really confident in terms of what we're seeing, in terms of maintaining and keeping that gross margin right where we want it, and right where we guided.

Aaron, maybe I'll add on that topic. From the outset, we designed our products to allow us to plug and play to flash from different vendors. We're using the same consumer-grade flash that goes into mobile devices and laptops, which gives us a much broader supply that we can access. We feel like we're in a very strong position. In fact, we think we're in a well better position than our competition. Maybe Matt has some thoughts on that?

To double-click there, if you look at the traditional Pure deal, because of our advantages in data-reduction software, our competitors will often have to configure between two and 4x more flash just to get the same usable amount of flash for the customer. That puts them at a much riskier position when you look at any kind of a supply shock in the system. If those prices rise, it's going to really we think clean out the market, and put the hurt on them, just doing these low-cost deals to try to compete with Pure.

These blips are very much part of normal operations. We move through them. The broad trend line is that flash prices will continue to drop in the quarters and years ahead, and that only brings more of the market to us.

Okay. Then following up a little bit on the competitive comments made there, we've seen a lot of talk about EMC and the positioning of their flash solutions, obviously now being a little tougher to see that under the Dell umbrella. There's been some discussion in the market about the repositioning, or lack thereof, of the extreme IO product relative to the all-flash VMAX product. I'd be curious what you're seeing competitively from EMC post the Dell transaction, and in particular as it relates to extreme IO versus let's say the VMAX All-Flash offering?

Hi, Aaron. Well, Dell EMC of course remains our most frequent competitor. Thrilled to say that in the quarter, our win rates overall held strong, as good as they've been in a year. Our win rate specifically against Dell EMC, we maintained the 2/3 POC win rate we've historically enjoyed. I think we are benefiting from the fact that the combined entity has nine different all-flash storage offerings, and that's creating some confusion in the field about which tech for which solution.

I will say broadly in our own experience, we are seeing a lot more all-flash VMAX than we are extreme IO today. Possible that moves over time, but it certainly seems that their cadence has shifted there. I would say the biggest opportunity up-tick we've seen is committed Cisco channel partners that really want to bring in products that they perceive to be long-term friendly, to Cisco networking and Cisco compute. Our FlashStack is enjoying triple-digit growth as a result of that.

Thank you.

Kix, anything you wanted to add?

No, think that's good.

Perfect. Thanks guys, good luck.

Thanks, Aaron.

Your next question comes from Katie Huberty from Morgan Stanley.

Thanks for the questions, and congrats on the quarter. The profitability up side to the back half, Tim, how are you thinking about that flowing through into the next couple of years? In particular, are you thinking that you hit profitability a quarter or two earlier than you would have thought a quarter ago?

I think, Katy, it's too early to tell. Again, as you might imagine, we're not going to guide going out; but what I will say is that Q3 and Q4 is really when we reap the rewards of those investments that we make. This has been a game plan that we followed for the last several years now. As I think into 2017, the first half becomes that growth or that investment portion of the business, as well, as we go back out to the last half where we will see those rewards come again. I think nothing fundamentally has changed in terms of how we think about the investment and then the return cycle in the business.

Okay, now that you have another quarter of FlashBlade under your belt, how are you thinking about the incremental TAM that adds to your business? What might that contribution to growth next year look like for that product, in particular?

Yes, Katy, this is Kix. We did a blog post about a month ago now updating our TAM with the result of a number of the expansions we've done recently, including FlashBlade, to re-state that to a $35 billion target market. I'd say as we've gotten into the first few quarters now of shipping in flash blade, we really have been enthused and encouraged by the wide variety of use cases.

When we first started, we began really positioning the product for some high-end use cases around analytics, software development, and chip design. As we had a quarter now of really coming on with the broader sales force, I think we've been excited to see the diversity. If you look at this last quarter, we've seen online gaming, a great example where the online experience plus the creative development experience are both accelerated by FlashBlade. Medical imaging and PAX, a whole new area; security log analysis; IAS and service provider environments. We're continuing to have confidence in the breadth of use cases of the platforms going on, and it still really early days.

We're proving out the time to value in the key verticals that we are targeting, and we're on track to hit general availability this quarter, which was always part of the plan. But we're going to avoid making specific guidance for next year's performance of FlashBlade at this time.

Generally speaking, would you expect a material revenue contribution in the first half of next year, or is it more back-half weighted?

Katy, I appreciate where you're headed with the question; but again, just too early to guide next year. We're happy with the momentum, but it still really is early days.

It's a good effort, though, Katy. I admire that.

Okay, thank you very much.

Your next question comes from Simona Jankowski from Goldman Sachs.

Yes. Hi, this is Matt Cabral on behalf of Simona. I wanted to circle back to one of the earlier questions about the competitive landscape. I heard the commentary about EMC, but it also seems like now NetApp is running over $1 billion of all-flash revenue. HP is up to about $750 million or so of a run rate. I'm curious if you've seen any change in how often you're going up against those two, and what you feel your real competitive advantages are against both of those?

Again, no material change to the competitive environment. We've maintained our win rates. We've been going up against all-flash for multiple years now. All of the core competitors took the effort to swap out the mechanical disk in their legacy systems and put in flash.

The key thing is that doesn't change the challenges. Those platforms are still too slow, too siloed, and too expensive. Pure is competing for the most strategic data in the enterprise because we're able to do things that customers can't do with the legacy retro-fitted products. You're seeing that in our success in the market, in our overall growth.

I would say the best way to look at this market is not just growth in the flash business, but compare that relative to the overall company's performance. If somebody's flash business is growing, but their overall business is shrinking in double digits, then really all they're doing is share-shifting inside of a shrinking installed base, which is not compelling going forward.

I would say that we are very happy with where our win rates are now; but if you look forward, I think we have the wind at our backs, in the sense that FlashArray demand is holding strong, and we've got the MBME differentiator just now coming to market, and FlashBlade ramping. There's just nothing like FlashBlade out there in the market. That is a huge and growing TAM that's going to continue to shift our way as flash prices fall.

One extra comment from Kix here. I would say if you look at the discussions we've been having technically with customers over the last couple quarters, two net new things. I would say the FlashBlade discussion has materially even helped the flash array business. At the end of the day, every customer has both structured and unstructured data. The product is so strategic.

Also, if you look at the competition, there just has been anemic response. NetApp really hasn't responded at all. EMC recently came out with a version of Isilon where they just simply retrofitted big, slow SSDs in the existing Isilon product. We really feel like the combination of FlashArray and FlashBlade is changing that longer-term strategy discussion around flash.

Then the second thing I'd say is this path to NVME is becoming real for customers. One of the things -- I was recently talking with a Gartner analyst, and he said look, I think the customer base is finally starting to internalize the rapid pace of change of flash. This isn't the three, five-year cycle any more. You have to invest in an architecture that's always going to evolve.

We're showing that very clearly with some announcements we recently made with NVME. NVME transition's going to happen faster than people think, and we're out there really differentiating our platform being built for NVME today, so the customers understand that they're investing in a platform that absolutely can be upgraded to the latest and greatest flash next year.

Got it. Then just a quick follow-up. I just wanted to clarify that the $30 million legal settlement in the quarter, that was all cash, and that was the big drag on free cash flow? I just wanted to clarify that. Then I'm sorry if I missed this earlier, but I just wanted to make sure you're still on track to hit free cash flow break-even in the second half of the calendar year?

Matt, Tim here. On free cash flow, yes, we did make the $30-million payment to the legal settlement in the quarter. It has been stripped out of the FCF numbers that you saw in the presentation, so that excludes that calculus. But if you look at our overall cash balance change, $30 million actually did have to do with the legal settlement. Then on terms of cash-flow positive, no change in that still signaling a sustained cash flow positive in end of calendar -- towards the end of calendar 2017.

Thank you.

Your next question is from Alex Kurtz with Pacific Crest.

Yes, thanks for taking the questions, guys. Tim, back on the questions that Aaron was asking around product margin. Would you say that 100-basis-point estimate on the FlashBlade impact was played out in the quarter?

Yes, it was in that ballpark. Certainly, the majority of the change, Alex, is indeed due to FlashBlade.

Should we be thinking about a similar kind of impact as you scale the FlashBlade business in the January quarter here, and think about product margins maybe in the same levels because of that?

Yes, we've guided the implied mid-point of our guidance is about flat quarter on quarter, and I think that's exactly it. You're going to see FlashBlade become a bit of a bigger portion of the overall revenue. Its gross margin is going to probably improve, as well. I think when you swirl those together, you get kind of that stabilization that we've seen.

The other thing I look at is I look at overall momentum in the business on the FlashArray business on a gross margin business. I think one of the important take-aways here is that the gross margin business that we were writing at the end of the quarter was stronger than the gross margin we were writing at the beginning of the quarter. Really -- go ahead.

Just to tie this into the service margin discussion with Scott for a second around competitive opportunities, it looks like you're improving your efficiency on the services side, which is going to give you some flexibility if services margins continues to trend up here into the mid to low 60%s. Are you going to use that opportunity to take some margin from services and apply it to product to go after some incremental accounts? Is that how you're thinking about it?

This is Tim again, Alex. That's very deliberately why we've guided the overall gross margin number the way we have. You'll note that we don't guide between product or support, because we've got the leverage and the capability to do a variety of things, as you allude to. The thing also I'd point out is we are best-in-class in product margins and best-in-class in gross margins. A lot of our competitors have numbers that are fives or fours in front of them. We're really proud of that. It speaks to the differentiate of the technology, and gives us that flexibility to go out and capture share and grow the business.

All right, thank you.

Thanks.

Your next question comes from Rod Hall with JPMorgan.

Hi, this is RK on behalf of Rod. Thanks for taking my question. I have one question and one follow-up, if I may. Could you let us know what provided the upside in the quarter in terms of both revenue, as well as the EBIT margins?

Yes, on the revenue side, it really comes down to demand. It comes down to demand on both the product -- on the FlashArray product and the FlashBlade product. We had good momentum on both of them; really good solid sales productivity and pipeline generation capabilities in the quarter. Again, it's really a testament to the technology and the quality of the product and service offering that we have in the business.

Then on the bottom line, a handful of things I'd point out to. Number one, naturally up side on the revenue side will drive to the bottom line, which certainly helped the bottom line. Then we're starting to get some benefits in leverage in terms of productivity that I talked to alluded to. Then the other thing I would say is we're now starting to see reductions in some of our legal spend, as well, as a function of settling the trough.

Okay, great. Also, I wanted to make a clarification about FlashBlade. I think last quarter you alluded to FlashBlade revenues being material next year. I wanted to confirm if that's true or not, and if so if you're reiterating it now.

No, what we said was that FlashBlade was not material for the current fiscal year, and that we were not offering guidance on FlashBlade out into the next year, so calendar 2017. If you want to make it clear on that, we're not providing any guidance on FlashBlade any farther -- or not at all into 2017.

Okay, and could you give us any indication on how much lower FlashBlade gross margins are compared to your FlashArray?

We don't provide that, and the reason being is it's early days. We talked a little bit about this on the last call. Very similar to what happens with our 400 Series product in the old days and the M Series product, there's a ramp, a manufacturing ramp. It's really until we hit the stability that we would start talking about FlashBlade gross margins.

The other thing I would say in FlashBlade gross margins, they will run right in our long-term margin targets over time. It's just a question of getting it ramped up over the next couple of quarters.

Okay. Lastly, whether any other factors like -- other than FlashBlade -- like international mix or pricing which impacted your gross margins this quarter?

No, the predominance was again FlashBlade gross margins. That's what we had talked about and signaled. It came in largely as we expected.

Okay, great. Thank you.

Your next question comes from Simon Leopold from Raymond James.

Great, thank you for taking my question. You made a comment in the prepared remarks about the fourth-quarter seasonal -- seasonality. I'm wondering if you could talk a little bit about what you anticipate for seasonality longer term? Really, the basis for the question is not simply to get guidance for FY18, but to get a better understanding of how to think about the cadence through the year?

When I look back at the April 16 quarter it was a negative 6.8% sequential versus the April 15 quarter. That was a 12.5% sequential rise, moving in opposite directions with current consensus indicating that the expectation is the first fiscal quarter's a down quarter. I just want to get a better understanding of what are the drivers? Is it related to things like sales incentives or customer behavior? Help us understand what seasonality to expect. Thank you.

Yes, Simon, great question. In terms of seasonality, enterprise storage has always been a very seasonable business. Q1 tends to be the weakest quarter, Q4 tends to be the strongest quarter. That's why you typically see most companies have this sequential decline. What that really is driven on is it's largely to do in part to customer buying behavior -- the classic budget flush at the end of the quarter, and then the starting of the new quarter. I think you're thinking about it absolutely correct.

The other thing I would say is that as we think forward -- again, we're not offering guidance for next year -- but as we think forward, we've got two dynamics in the business going on. We've got what I call the secular growth, which has been growing very fast, hyper-growth. Then the seasonal trend, which has always been there. How I think about it here internally is that seasonal growth trend will always be there, and that secular growth trend will be slowing down. It's natural in terms of the law of large numbers, so just factoring that into your over, and thinking how you're thinking about the modeling.

Great, I appreciate that. Sorry.

Sorry, one quick add. I agree with Tim. It's predominantly customer behavior that the end of the year is the time that storage is purchased. We do have the extra impact that in Q1 we're ramping new hires into our sales force. That's when we add the most people. We're re-jiggering territories and segmentation and so on. That's another reason that Q1 is likely not to be as strong as Q4. We would expect that same seasonality that we experienced last year to occur again next year.

In terms of your October quarter, do you have material federal exposure? Is that a factor in terms of your October quarter performance and trending into the January quarter?

Yes, your federal quarter, the October quarter is always the federal quarter, their buying cycle, so you're spot on there. We have a nice federal business, state and local government business as well. That's some of the dynamics you're seeing, and why Q3 was a bit stronger than Q2.

Great. Thank you for the help on that.

Absolutely.

Your next question comes from Maynard Um from Wells Fargo.

Hi, thank you. Regarding the $30 million settlement in the quarter, I just wanted to see if you can clarify. Are you making any ongoing quarterly royalty payments as part of the settlement? I'm trying to figure out whether or not that had any impact on the gross margins, either in the quarter or going forward? If you can, help us quantify the OpEx benefit. If I'm not mistaken, I think the last time a legal issue went to court, I think there was an associated increase in OpEx related to the legal fees that I think might have been $1 million or $2 million. If you can help us with the quantification, that would be helpful? Thanks.

Yes Maynard, Tim. On your first question, the $30 million, all rear-view mirror -- nothing in gross margin, no ongoing royalty payments. We are completely closed off on that topic.

Then on a go-forward basis in terms of legal spend, I'm not going to offer a specific number, but as you saw in some of the leverage in our guide in Q4 relative to where consensus was, it was born about less expenses from a legal perspective. It does make a difference on the bottom line, and one of the reasons why we're happy to put this behind us.

Great. Then lastly, you talked about flash pricing dropping and helping you as more of the market starts to come to you. But do you need to talk about your expectations for when you expect to start seeing more material declines in flash pricing, because we've seen the NAND pricing actually going flat to up more recently? I'm wondering when you expect that will start to help you more significantly? Thanks.

Yes, our earlier comments were actually in reference to some anticipated questions about a bit of a supply bubble that's happening right now. We're anticipating slight increases over the next few quarters. Those tend to be seasonal, and nothing we're particularly worried about riding out.

As we said earlier, we have multiple suppliers. We've architected our product for that from day one. Because of our superior data-reduction technologies, we believe we're much less impacted by any shifts in NAND pricing than some of our competitors who rely on low-cost pricing and raw flash to win.

Just to be clear, the trend we were outlining is the very long-term trend that we believe that costs will continue to come down for flash as we see more capacity ramp, just because it's an explosively growing business.

Great, thank you.

Thank you.

Your next question comes from Jayson Noland from Baird.

Okay, great, thank you. Scott, I wanted to ask about your blog post where you say NVME will be as big as all-flash array, and you mentioned an all-NVME array. Could you talk a little bit about the timing, and any required investment on the part of Pure?

We're going to avoid making new product announcements at this particular juncture. I would say we have been making R&D investments to prepare for that all-NVME future for more than 18 months now. We are tracking to our long-term plan

We absolutely believe this is going to be a profound differentiator in terms of how much value customers can extract from their data, how many different applications can use that data, how much more real-time analytics we can accommodate. We are shipping NVME technology in the current flash array M product, and we are providing our customers with an investment-protecting path that will be non-disruptive to their business, to get the full benefits of all-NVME going forward. I think we're the only major storage vendor doing that.

Just to give you a little more clarity on why we think this is going to be so disruptive. If you look at the way most storage arrays are built today, all-flash arrays, they use SaaS technology, serial-attached SCSI. The first S in there is serial. That means a single point-to-point length from CPUs down to the flash.

Moving to NVME, it literally gives you 64,000 parallel cues. It really helps us take advantage of multi-core CPUs connected to massive amounts of flash. As we move to these denser and denser all-flash arrays, that becomes critical. You first of all need a hardware connection that can support that, and we of course own our entire hardware architecture and design that from scratch. More importantly, you really need the software technology to be able to take advantage of it. You can't just plug it in and exploit that parallelism easily.

We believe this is a pretty major transformation. It's one that we saw coming anticipated three years ago when we started building FlashArray M, and built NVME from the start. It's one that we intend to lead the transformation industry around.

As Scott says, we're not ready to announce any new products yet; but like we said from the early days of FlashArray M, we've built the ability for that upgrade to come within the product already, and we anticipate taking customers there in a totally non-disruptive fashion.

We think this is going to be a substantial challenge and hurdle for our competitors that are starting with legacy technology. It is likely a multi-year software re-write, as well as -- that's for the vendor. Then for the customer, it is a re-purchase of the storage, as well as a data migration. They're going to sacrifice investment, whereas our customers get to preserve investment.

It's better across the boards. It's going to allow us to deliver a lot more technology in a smaller form factor, so we're not just going to be able to add more performance. We're going to be able to cut cost of our customer infrastructure with this technology.

One final thing I'll add. We really are excited about this, not only because of the customer benefit, but we view it as TAM expansionary. If you look over the last few -- maybe call it five years, many of the newer-skill web generation applications were built in a world of local data storage. NVME allows us to go in and attack that market now with a shared storage footprint.

We can take those local SSDs, those local flash cards out of those servers, and replace them with a shared array that offers all the benefit. From a customer's point of view, it's best of both worlds. They can potentially get the performance of local flash in a server, but all the benefits of manageability, reliability, and economics of a shared environment.

Okay, appreciate the color on that. Then a follow-up on international. It looks like you saw strong growth year on year. Which markets is that coming out of, and is it a trend that we should expect to continue?

Yes, on international we had momentum in a number of markets. I guess I'd highlight two of them. We talked about Telecom Italia on the call as a named customer reference in Europe, very strong momentum there. Our sales leader James Petter in EMEA has done a phenomenal job in terms of growing that region in general. Another one I'd call out is Germany. We installed a new leader there recently, and had a very strong big win in a large German financial services firm, as well. A lot of things kicking on all cylinders, both there as well as the APJ Markets, as well.

There's so much addressable market globally, and we're so excited with the prospects that we have to continue to expand. I would say we have to continue to get better in our consistency. We have countries that are exploding; we have others that are still ramping. We've just got to keep doing our work, and you'll see that international growth continue in the years ahead.

Thanks, guys.

Your next question comes from Eric Martinuzzi from Lake Street Capital.

Thanks. You commented a little bit on the gross margin linearity in Q3. What about the revenue linearity?

Revenue linearity in terms of -- in quarter, Eric, or trying to get to understanding of the question?

Yes, within the quarter. Typically, I would think an October quarter would be very September-October loaded with the doldrums in the summer months. What did you see?

That's exactly it. That's the dynamic we've seen for several years, is that particularly some of our international markets, August is a very slow time, just because people are away. I think you've characterized it perfectly -- consistent behavior that we've seen in years past, as well.

Okay, and that also for the services side of the house, I know the primary support revenue is around maintenance contracts; but as far as professional services to the extent that you were performing pro services, was that similar as well?

Yes, and pro services is, as you allude to, a very small portion of the business. The vast majority of the number you see on that line is just amortization of the support contracts.

A key part of our practice is we try to encourage the services business to go to our partners, because it's a key place for them to add value around our product offerings.

I understand, thank you.

Your next question comes from Tim Long with BMO Capital Markets.

Thank you. Two questions, if I could. Just following up on the services line, you talked earlier about the gross margin having flexibility between the businesses, but you had a nice ramp in the services line. Is there more runway there for you to play with?

Then separately, just curious on the -- you said the win rate's been pretty stable, despite some of the incumbents maybe catching up with all-flash array products. Anything stand out as far as win rates by different verticals or geographies? Is there any area that you think your win rates are a little bit better than the competitors, and more defensible? Thank you.

Tim, on the first question, is there leverage in the support number, I'll give you just a very short and simple answer, yes. You've seen that grow very nicely over time, and we still see line of sight to greater things there.

On win rates, it is consistent, right? We're measuring across all of our competitive engagements and all geos. The fact that those win rates are holding strong is reflective of our performance broadly.

I just want to hit again, the differentiation gap is not closing. You take these legacy and commodity storage platforms and you put flash into them, they are still too slow, they're too siloed, and they're too expensive and too complicated. They do not do the things that Pure does. That's why we've been able to sustain these win rates for the past couple years, despite the competition doing the best. That's why we're growing, and our competitors are shrinking in their overall market footprint.

Okay, thank you.

Your next question comes from James Kisner from Jefferies.

Thank you. I first want to tell Scott that you must have misread the script because you didn't mention Jefferies as an important new customer. (laughter)

But can you talk about linearity in the quarter? DSOs were spiking up a bit; last October DSOs went down. Separately also on gross margin, if FlashBlade were excluded from your guidance, would you expect margins to be up or flat for product? I would expect it to be up, given that flash would be -- I should say I think it would be flat, if we're not seeing any pricing pressure? Thanks.

Yes James, on your first question, from a DSO perspective, DSO in terms of timing of billing, as you allude to, can go up and down a little bit. The internal metric that we use is weighted average days to pay, and that's been nice and stable. It gives me confidence that the overall business is performing well. As it relates to gross margin, what I'll say again is that yes, we've hit that period of stability. We're happy where the FlashArray margins are right now, and FlashBlade will blend in very nicely, as reflected in our guide.

James, thank you so much for working in the product plug. We're extremely grateful to have you as a customer, as well.

Thank you very much.

Your next question comes from Rich Kugele from Needham & Company.

Thank you, good afternoon. One last technology question on NVME. As you said, you have been shipping it in systems that are capable; but what is the inflection point this year that makes -- or this coming year -- that makes NVME suddenly get adopted? Is there any work that has to happen on the customer end in order to make that so? Then Tim, if you could just talk about the hiring plans for next year relative to what happened this year? Thanks.

Yes, on NVME -- first of all, to clarify how we've been shipping it -- when we first shipped FlashArray M, we actually built our own dual-port NVME-NVRAM devices. We also laid the groundwork in the product -- the plumbing, if you will, to support NVME flash drives for the bulk storage. We've been anticipating this transition.

What we think is going to push the transition is the adoption of NVME across the broad industry, especially in the consumer space. Consumer demand always pushes flash. If you've bought a modern laptop recently, if you bought a brand-new phone, it's got NVME flash inside. That transition is starting to happen, where the entire industry is gearing over towards NVME as the volume product.

The third reason I think is it's just a better gateway to open up and create the next level of growth in industry for data. If you look at our highest-end customers that are trying to do things that haven't been done before in the data path, this opens up the next level of performance. It's just good for the industry. As a leader, we intend to push it, because we believe the customer impact is going to be hugely positive.

It's one of the very pragmatic considerations. If you just look at SSDs in the data center, when Pure got started, they were just 256 gigabytes. Now, the SSDs that are used can range from 8 terabytes up to 32 terabytes, and they're road-mapped to go well higher than that.

The problem with the interconnects that we have for those SSDs is they're way too slow. The capacities have gone up 100-fold, and we're still stuck with the same access model. That has to be replaced; otherwise we can't continue to scale and drive density. Density is what allows you to mine a lot more value from data. NVME is coming along at the perfect time to allow us to continue to advance the data platform.

Two other things I'll just say about the way we do things at Pure. One is that we believe in making flash mass market technologies. In the early days of Pure, flash was this high and exotic thing. People thought it was only for the 1% of storage use cases. We really broke the market open by making it mainstream, and let others follow in that strategy.

We intend to take the same path with NVME, and be very aggressive about allowing it to really hit a broad swath of the market.

Then the second thing is we really believed in a non-disruptive upgrade path. Customers shouldn't have to ditch their first-generation storage all-flash array and move to a brand-new all-flash array to support NVME. We built that transformation right in.

When we look at our competitors -- let's pick on, let's say MC for a minute -- they are following the opposite strategy, right? They've got a product called DSSD. It's at the uber high end of the market from a cost and performance perspective, and it's a complete island in terms of upgradability from any of their other eight-odd all-flash arrays.

Rich, I want to make sure we cover off on your second question around investments and how we think about next year. Obviously too early to talk about guidance, but I'll give you a frame work about how we think about it.

Really, it's the play book -- and we talked a little bit earlier on the call about it -- it's that play book of invest early and then reap the rewards in the latter half of the year. When the best sales people are coming available to us, they're coming off a good year somewhere else, and we'll pull them up.

I think in terms of just said sequencing and seasonality of investments, you're going to see this coming year be very similar to how you've seen years past. It's also when we have our big demand-generation events. Our user conference, which was a significant success for us last year, we're going to do that again this year. That's how to think about investments as we get into the next year.

Great, thank you.

Your next question comes from John Lucia from JMP Securities.

Hi, guys. Thanks for taking my question. You announced an expansion of your Cisco partnership in mid-October, and you've talked about triple-digit growth with that partnership. Can you just touch on the Cisco partnership and what's driving the momentum there? Then longer term, what kind of opportunity does Cisco represent as a partner for you?

Yes, I'd start with there's an absolute segment of the market that wants to buy conversion infrastructure solutions. It's a full-stack mentality. Customers want an end-to-end tested solution that's designed for the highest-end environments.

When we first looked at this market space a few years ago, we really saw the first generation of CI solutions built in the era of disk. They weren't particularly efficient, and they were pretty complex. There was a huge services wrapper around getting them to work. We saw an opportunity to do a lot of the same innovation we did in the all-flash space for CI in general, to go in, to bring flash, to make it more efficient, and to modernize for a next-generation set of applications. We've seen a lot of positivity around there.

I'd say the other thing that's important to realize is there's a lot of pull for CI from the channel; because what do channel partners want to do? They want to sell complete solutions for their customers. CI makes that easy. I'd say now with this transition, particularly in the channel around the Dell EMC merger, we're seeing a lot of excitement around FlashStack, and wanting to look towards the next-generation of CI solutions.

It's why we've invested to make Pure as channel-friendly or partner-friendly a storage Company as there is in the industry, because it's such a great opportunity for us to align with these committed Cisco partners that are looking for storage that fits very well into the overall solution.

Okay. In the press release, it talked about thousands of deployments with Cisco. That would suggest it's a pretty large partnership. Can you give us any sense for the relative size of the partnership with Cisco?

When you talk about thousands of deployments, we do indeed have thousands of configurations where Pure Storage is interconnecting with Cisco systems. Then we also sell this converged infrastructure where the technology is combined before it goes into the customer site.

Both of those are key parts of our business, right? We sell independently and our infrastructure gets used. Our platform gets used along with Cisco systems; but then through our channel in particular, a lot of our customers are buying FlashStack converged infrastructure. It's that FlashStack converged infrastructure that's growing at triple-digit rates right now.

Okay, thank you.

Your next question comes from the line of Mehdi Hosseini from Susquehanna.

(no audio) national. Two follow-up questions for Scott. Can you help me understand how the cost per gigabyte for your customers are changing, so we can better understand the value you're creating? I have a follow-up.

Sorry, so how the cost is changing per gigabyte?

Yes, effective cost per gigabyte, including de-duplication and all the proprietary savings that you provide for the customer?

Yes, I think the best proxy for that, Mehdi, is that our product margins are stable. What that means is as we slowly drive down cost in our supply chain, we are handing some of those savings along to our customers, and that we continue to drive the industry's most efficient technology, in terms of having the best data reduction and the most efficient erasure coatings, that will allow us to get the most mileage out of common hardware. We don't specifically publish dollars per gigabyte, but I would say our customers purchase the storage based on how much data they can store on it, not how much flash it comes with the device.

Okay. Then moving on to my second question, which is more of the bigger picture. I understand your commentary about the hyper-converged. It's entirely different than what you're doing, but hyper-converged could be like a 100-plus-billion TAM. You're focused on 10, 20, 30 billion TAM.

In that context, are you just being very diligent with your R&D dollar, and focused on your next two, three-year road map? I see both converged and hyper-converged co-existing, but hyper-converged could be a much bigger TAM. I've got a feeling that you just want to stay focused; but why not think about the bigger picture, especially in the longer term? I want to better understand your thought process. Is it just being diligent with the R&D dollar and not wanting to encroach into other TAM, or is it something else?

Mehdi, we're extremely excited about our target market. Being the preferred platform for our customers' most strategic data, the data that they are seeking to mine the most value from, we think is a supremely large, terrific market opportunity, maybe the biggest available market opportunity in tech.

We see growth in cloud. We're substantially exposed to cloud software-as-a-service, infrastructure-as-a-service, as well as consumer cloud companies. We're seeing great growth in the Fortune 500. I would say we're right where we want to be, and playing for a market that we think is growing and tipping in our direction.

Great, thank you.

Your next question comes from Steven Fox of Cross Research.

Yes, just one quick one for me, please. Looking at your SG&A expenses for the quarter, it seemed like it was a decent amount less than I was thinking. I could be off, but I was wondering if you maybe realized an abnormal amount of sales productivity during the quarter, or if there was some spending that was deferred, or maybe that's about where you expected it to be? Thanks.

Yes, Steve, this is Tim. I think largely where we expected to be. I think the R&D line could have been up a little bit. Then we did start seeing some savings in what we call outside services or professional services, as our legal spend starts winding down a little bit.

So productivity is pretty much tracking as you thought overall from a sales force standpoint?

We're happy with productivity, yes.

Okay, thank you.

There are no further questions. I will now turn the -- now turn it over to Scott Dietzen for the closing remarks.

Thank you, operator. Thanks for your time today. We're tremendously excited about the position we're in. Pure's data platform is allowing our customers to extract value from their data they can't get from any other solution. We're big, we're fast, we're easy, we're elastic, and we have the lowest total cost of ownership of any of the solutions in the market place.

More importantly, we're uniquely capable of building the most data-intensive applications and supporting many different apps all using that data simultaneously. The legacy and commodity storage we compete against is too slow, it's too siloed, it's too complicated, and it's too expensive.

FlashArray demand is strong. With NVME, it's poised to tip the scales further our way. FlashBlade is just ramping now into a huge and growing TAM. As I mentioned earlier, as flash prices decline over time, ever more of that big-data market opportunity is going to tip in our direction. With our platforms well differentiated, operating at scale, one of the biggest -- if not the biggest available markets in tech, the future is incredibly bright in front of Pure.

We want to wish you a happy holidays. We're very grateful to have your time. We'll see you in three months. Cheers.

That is the end of today's conference call. You may now disconnect.