Splunk Inc (SPLK) 2020 Q4 法說會逐字稿

Ladies and gentlemen, thank you for standing by, and welcome to the Splunk Inc. Fourth Quarter 2020 Conference Call. (Operator Instructions) Please be advised that today's conference is being recorded. (Operator Instructions)

I would now like to hand the conference over to your speaker today, Ken Tinsley, Corporate Treasurer and Vice President of Investor Relations. Thank you. Please go ahead, sir.

Great. Thank you, Dilem, and good afternoon, everyone. With me on the call today are Doug Merritt and Jason Child.

After market closed today, we issued a press release, which is also posted on our website. Also note that we have posted supplemental material on the Investor Relations web page as well. This conference call is being broadcast live via webcast and following the call, an audio replay will be available on our website.

On today's call, we will be making forward-looking statements, including financial guidance and expectations, including our forecast for our first quarter and full year of fiscal '21, '22 and '23, duration, revenue mix and long-term cash flow and our expectations regarding our products, technology, strategy, customers, market, acquisitions and investments.

These statements reflect our best judgment based on factors currently known to us and actual events or results may differ materially. Please refer to documents we file with the SEC, including the Form 8-K filed with today's press release. Those documents contain risks and other factors that may cause our actual results to differ from those contained in our forward-looking statements.

These forward-looking statements are being made as of today, and we disclaim any obligation to update or revise these statements. If this call is reviewed after today, the information presented during this call may not contain current or accurate information.

We will also discuss non-GAAP financial measures, which are not presented in accordance with generally accepted accounting principles. A reconciliation of GAAP and non-GAAP results is provided in the press release and on our website.

With that, let me turn it over to Doug.

Thank you, Ken, and thanks, everyone, for joining us. It was a great quarter and year-end for Splunk. I'm very proud of our results and our team, and I'm grateful for the trust and commitment of our customers and partners.

We ended the year with annual recurring revenue of $1.68 billion, up 54% over last year. I'm happy to report that this is ahead of what we forecast last quarter when we introduced our transformation metrics of ARR and operating cash flow. Our shift to renewable model is 99% complete, as customers are now predominantly opting for term and cloud contracts.

The flexibility and predictability of our new Data-to-Everything pricing options have made it easier to do business with us and for customers to bring data to every question, decision and action. And now we're focusing squarely on cloud as the driver of our next phase of growth. Cloud-first has become our mindset from both a business model and a product strategy perspective. 35% of our software business was cloud this past year, and we expect cloud to ramp at a strong pace over these next few years, reaching over 60% of our total software bookings in FY '23.

I'm extremely excited about the opportunity we see ahead of us, and I'm confident in our ability to navigate another successful business model transformation, as Jason will elaborate shortly. Stepping back, what we know for sure is that digitization, or digital transformation, is becoming a real, tangible and critical global initiative. Digital security has become a board level discussion for global enterprises to start-ups planning to go public, and there is increased focus on digital technologies to help solve some of humanity's biggest challenges. This makes it clear that over the next 5 years, every organization must become a digital organization. And data is the fuel powering a digital organization.

At Splunk, we're empowering customers with capabilities they simply can't get anywhere else at the time that they need it the most. We see tremendous opportunities across our core markets of IT operations, security operations and application development. And we're investing in products and technologies that extend Splunk's value and relevance to all of our customers' decisions and actions, regardless of whether the data they need sits in the Splunk index or somewhere else.

Capabilities such as stream processing ensure customers can access massive amounts of data in real-time then learn, refine and adjust that data as it moves across the stream. Our federated search capability searches across Splunk deployments and non-Splunk data sources to provide users a holistic overview of their entire data landscape, at much higher speed and scale than before.

Event management, orchestration and automation, help our customers proactively remediate issues and take action on the insights and analysis they're capturing. And by focusing on new user interface frameworks like consumer-grade design, collaboration, mobile and natural language processing, we're working to ensure that everyone in the organization can get value from their data, regardless of how skilled or technical they are.

Our vision to bring data to everything is resonating with our customers. As an example, I'm proud to report that we recently completed our largest order in company history. Thanks in large part to the value that data stream processing, SignalFx and our mission brings to our portfolio, plus the combination of streaming, traces, metrics and logs, combined with unparalleled scale and enterprise capabilities. That portfolio enabled us to beat 16 competitors, while our new pricing gives the customer the confidence and clarity needed to commit to a massive enterprise scale order. This customer win demonstrates the power of our Data-to-Everything portfolio. And most importantly, our reputation for delivering huge value to our customers.

A few more highlights of customer wins in the quarter include McLaren Racing, a prominent technology innovator and racing team, who purchased Splunk Enterprise to capture data across their organization, including racing analytics as well as their infrastructure, network and server environments.

Splunk's Data-to-Everything platform will help McLaren Racing improve customer engagement on and off the racetrack. Discovery, Inc., provider of television brands, including Discovery Channel, TLC, HGTV, Food Network and more, expanded their use of Splunk Cloud to gain real-time visibility into their consumer streaming experience. Splunk is the central investigative and monitoring solution to their video delivery platform. And the streaming analytics powered by Splunk improved the global viewing experience and provide faster mean tender resolution when IT or application issues occur with direct-to-consumer offerings.

Macquarie Government, the cloud security platform trusted by over 40% of Australian governmental agencies, purchased Splunk Enterprise, Enterprise Security, Phantom and UBA, to gain real time visibility, monitoring and response into the agency's future security state. Splunk is a key player in Macquarie Government's multiyear transformational data journey and will be used to support the agile nature of the Australian Taxation Office as well as help protect their IT environment from security threats, harmful code, fraud and spam.

Earlier, I mentioned that in Q4, we saw momentum build for our new pricing programs as more customers turn to Splunk for an increasing set of use cases. We've answered their call for more flexible and predictable pricing options, including those not based on data ingestion. This is making it easier for our customers to bring even more data into Splunk and find faster value from that data. This quarter, I've had so many customers tell me that our new pricing options have unlocked adoption within the business, which has given them the overall confidence, predictability and the right total cost of ownership to standardize on Splunk as their data platform across their organizations.

Although it's early days and awareness is still building, let me give you a few examples. Long time customer, Washington Post, expanded their use of Splunk Enterprise for deeper insights and analysis for all Post properties and for their fast-growing digital content management system platform, Arc Publishing. Leveraging our flexible pricing programs, the Washington Post uses Splunk for real-time monitoring and data insights to help improve their overall customer experience. Splunk is also used to improve information security and threat detection.

Another example, NHS Digital, who significantly expanded their use of Splunk this quarter through dynamic pricing, adding Splunk Data Stream Processor and Splunk Enterprise Security to help strengthen and secure their networks. NHS Digital already uses Splunk for security and operational monitoring of England's national health systems. This includes e-referral services, Health and Social Care Network and Spine, which indexes over 1 billion messages a month across 23,000 health care IT systems in over 20,000 organizations across NHS. With Splunk, NHS Digital can enhance security defenses, helping them to securely deliver important health services to patients across England.

In summary, it was a milestone quarter with our cloud bookings more than doubling Q4 to Q4, and it was a strong finish to FY '20. I'm so proud of the entire Splunk team. We're uniquely positioned in Data-to-Everything and are 100% focused on capturing the tremendous opportunity ahead. We're investing for scale, growth and innovation. We adopted a cloud-first mindset. And most of all, we're committed to helping our customers exploit their massive Data-to-Everything opportunities.

I'll now hand over to Jason to walk you in more detail through the quarter and fiscal year results. Jason, over to you.

Thanks, Doug, and good afternoon, everyone. Thanks for joining us today. Q4 capped a very strong year for us, highlighted by a faster shift to cloud than we expected.

We ended the period with total ARR of $1.68 billion, up 54% over last year. ARR was comprised of $442 million from cloud and $1.238 billion from term license and maintenance contracts. As we outlined last quarter, because customers have been opting for term and cloud contracts over perpetual faster than we anticipated, we discontinued new perpetual license offerings. As a result, in Q4, 99% of software bookings were either term or cloud and our transition into a renewable model is complete. Our focus now shifts to the move to predominantly a cloud model.

The bookings mix between term and cloud becomes an important balance to understand since it can have a meaningful impact to the financial statements. Going into FY '20, the year we just closed, we expected cloud would contribute about 25% to software bookings. Turns out that cloud demand was significantly higher and ultimately, 35% of total software TCV was from cloud, most of which came in the second half of the year.

And because revenue recognized from a cloud contract is fully ratable, even a large cloud contract on a TCV basis will yield significantly less revenue than a comparably sized term contract. As we've said, because ARR normalizes bookings activity irrespective of cloud or term, ARR is the most appropriate metric to evaluate overall momentum of our business.

We ended Q4 with total RPO of $1.8 billion, up 43% over Q4 last year. The portion of RPO, which we expect to recognize as revenue over the next 12 months, was $1 billion at period end, up 23% year-over-year. RPO bookings were $1.145 billion, up 23% from Q4 last year and full year RPO bookings were $2.9 billion, up 28%.

The accelerating cloud contribution in Q4 led to a slower pace of growth in current RPO bookings compared to total RPO bookings. This is due to the out years of a multiyear cloud contract being classified as noncurrent, whereas the majority of a term contract is captured in current RPO bookings immediately. Again, ARR normalizes for this mix shift.

In Q4, we recorded 221 orders greater than $1 million in total contract value, up from 179 in Q4 last year. Over the course of the year, we booked 35 orders greater than $10 million compared to 24 in the prior year. Recall that several large federal government contracts with 5-year term skewed over all duration in Q3. As expected, weighted average duration reverted to prior period levels of 34 months in Q4.

Now on to the P&L. Fourth quarter revenues were $791 million, up 27% year-over-year, and full year revenues were $2.359 billion, up 31%. Cloud revenue was $99 million, up 86% over last year. Full year cloud revenue was $312 million, up 82% from fiscal '19.

Q4 software revenues, which is the total of license and cloud, were $617 million, up 33% year-over-year. On a full year basis, software revenues were $1.686 billion, up 40% over last year.

On margins, which are all non GAAP, Q4 gross margin was 86.7%, down slightly on a year-over-year basis due to a greater proportion of cloud revenue contribution. Full year gross margin was relatively flat at 85%. Operating margin was 24% in Q4 and 14.2% for the full year, both in line with our plan.

Finally, operating cash flow was negative $288 million for the year, also in line with our expectations following a midyear shift to annual invoicing.

Looking forward, we are confident in the continued acceleration of customer adoption and high-growth trajectory, and we maintain our guidance for mid-40% ARR growth this year, fiscal '21.

Looking further out, we expect to sustain a 40% ARR CAGR through fiscal '23, and we reaffirm our $1 billion operating cash flow target for FY '23. Revenue guidance for this year is a little trickier given the high variability that comes from cloud mix and related ratable revenue. With current expectations that cloud will contribute to a substantially higher proportion to overall bookings in fiscal '21, it follows that revenue growth will flatten this year even as software bookings and ARR continue at high-growth rates.

As we begin to benefit from the renewal of previously booked contracts in FY '22 and '23, revenue growth should rebound sharply. So we expect total revenues in Q1 of about $450 million and full year revenue of $2.6 billion and will follow our normal seasonal pattern of 40% front half and 60% in the back half.

With anticipated revenue snap back starting in fiscal '22, we expect annual revenue growth in the high 20% range in both FY '22 and '23. With this revenue trajectory, it follows that operating margin will trough in fiscal '21 before rebounding in '22 and '23. We expect negative 25% operating margin in Q1, and roughly breakeven for the full year and a long-term target of 20% by fiscal '23.

At our upcoming analyst and investor meeting, we intend to provide a framework on how varying degrees of cloud mix will translate to revenue and profitability metrics over time. And we look forward to walking you through more details of this transition and our next phase of growth.

In closing, it was a strong finish to the year, and we're entering fiscal '21 with excellent momentum as the cloud business is poised to ramp rapidly.

With that, let's open it up for questions.

(Operator Instructions) Our first question comes from Kash Rangan from Bank of America.

Congratulations on the forward-looking guidance. It looks absolutely terrific and significantly higher than what we expected.

So I want to get this right. So what you're seeing is in fiscal '23, you are poised to doing about $4.6 billion. So just take your numbers -- guidance of CAGR 40%. So about $4.6 billion in ARR we're talking about in calendar '22 time frame. And revenue growth, it looks like based on the guidance for high 20s. If I were to CAGR the fiscal '21 guidance into fiscal '23, that's about 35% compounded revenue growth rate. And you're reiterating your operating cash flow guide for about $1 billion. It looks absolutely stunning. It looks like the -- some similarity to the Autodesk model transition, which I think was well received.

The question for you is with respect to the total RPO bookings, is there some seasonality? Because it was up about 42% in Q3. We got lulled into thinking that it could be up massively again. I know, Jason, you've talked through some of the mechanics of this shift towards the cloud. But total RPO bookings, that number was up 22%. Can you shed a little bit light there? And also with respect to the revenue guidance for next year, what are your assumptions in the 2.6 -- clearly, everything else looks even better than expected.

Congratulations again. These are terrific numbers -- forward-looking numbers.

Thanks, Kash. This is Jason. So I guess just to first tackle the RPO booking question. So yes, it was 42 in Q3, but also recall that we added -- we added SignalFx into the quarter. So when you normalize for that, we grew 35%.

That said, it is a deceleration. The reason for that is because on -- RPO bookings is, obviously, it's revenue plus change in RPO. And so there's a number of things that move. The reality is that RPO bookings is a bit volatile. And you see this, if you look over the last 4 quarters, you've seen a couple -- you've kind of seen it go up and down, and up and down, and the reason why is because it's really the combination of the timing of when deals close. And as we get larger, we have a larger number of larger deals that can actually cause that volatility.

And then, also, of course, it depends on what we're comping against in prior period. And because of the lumpy mix kind of -- of when deals hit, it moves around a bit. I actually think looking at RPO bookings on a TTM basis is probably the best way to look at it because it's kind of -- what do we add into the quarter? Now -- in terms of new deals.

Now the best metric, as you've heard me say, is ARR. And the reason ARR is better than RPO bookings is because it's going to cut through timing of when a deal closes. It's going to cut through whether it's revenue or whether it's on-prem or whether it's cloud. And it's just going to show what is it that we're actually delivering in the next 12 months. And so that's the number.

And then you see the ARR growth has actually accelerated for, I guess, now 6 quarters in a row from when it was 46% from the first quarter reported in Q3 of a year ago, and then now it's kind of been stepping up every quarter. So we're really excited about the ARR growth. And that's the number I'd focus on.

It looks fantastic. Doug, what made Splunk win that megadeal against 15 companies? And what makes Splunk win in this market that seems like a little crowded? Why should Splunk win in the long term? That's it for me. Congrats.

Thanks, Kash, and thank you for that softball question. So this was a...

I can ask a hardball, too...

I know. I know.

But these numbers look good. They look good.

Yes. Super -- no, I'm super appreciative because it is -- we're really, really proud of that deal. We've obviously been making a ton of investments across the portfolio in all the buying centers, security, AIOps and the app-dev/observability area. Even adding things like data stream processing and federated search. And this was, I think, a testament to what we've been driving across the portfolio.

What the prime use case for this large customer was app-dev and the next-generation dev ops model. But like all customers, they also want to leverage the offerings for security overall, data center management in the infrastructure side. And what distinguished it is when you paired the incredible throughput and scale capability of DSP to even further enhance the automatic metrics extraction and dash-boarding of SignalFx, and then the -- I think, the strong synergies of tracing and what we've done with distributed tracing with the SignalFx on mission with the underlying log management platform from Splunk.

There was nobody that could even get within, literally, miles of the scale and fidelity and enterprise features that this customer rightfully demanded. They're a massive, massive organization that powers huge chunks of dollar transactions around the globe.

So it really was -- the investments we've been making in high scale, high-fidelity, high-capability, enterprise-class functionality, from everything from logging to metrics, to streaming, to tracing, that came together for this customer.

Our next question comes from Brad Zelnick from Credit Suisse.

I echo Kash's congrats on a strong finish, especially the large deal performance and a lot of the wins that you talked about. I don't know if I can do that as nice as Kash does it, but I'm as enthusiastic.

So I guess, maybe my first question, Doug, I know the budget for what you offer is very high priority across just about all verticals. But I have to ask the question, is there anything you're seeing real-time in terms of the demand environment just given global health concerns out there that factors into your outlook for the business right now?

So the forecast we gave for Q1 and then for the subsequent 3-year outlook, FY '23 outlook was based on everything that we have available to us today. Right as of this moment, we are not seeing any immediate impacts from all the reactions to COVID and the other volatility we see in the market. Like every other company, we're monitoring it literally on a multiple time per day basis.

What I can -- what I'm super excited about and happy that we're able to pull off as we did have our sales kickoff last week in Las Vegas. I know a lot of companies are reevaluating whether that is something they're going to do or not. I think that's an absolutely pivotal event for every company. It's so difficult to ensure your entire go-to-market organization is aligned with your product organization. Especially in a year where we are leaning in hard on annual contract value versus total contract value, and the cloud has got a lot of dynamics to it, and we've got a brand-new observability suite and the pricing that we're communicating is so important.

So the -- part of that forecast was built on talking to roughly 2,500 people over the course of 5 days and a lot of the forecast calls we've been driving on a daily basis as well.

And Jason, maybe just for you, as we look out past fiscal '21 to get to your '22 and '23 guidance. How should we think about what you're embedding from a maintenance conversion perspective? Are you -- and are you expecting any uplift upon renewal of term licenses as customers transition to cloud?

Yes. It's a great question. I would say, last quarter, I gave the metric that I think it was 5% of the perpetual install base had converted to either term or cloud. That metric hasn't moved much, but I think we're going to provide a more helpful metric for you within the next quarter. And that is to look more at net expansion rate because we're doing very, very well against that metric.

So in terms of the expectations, there is a -- there's definitely -- we know exactly kind of what the renewal cycles look like, which is why we are confident in the ARR growth. That's why I was able to say that we have a 40% ARR CAGR from now through FY '23.

So that will certainly be a big part of growth. We also -- certainly, there's going to be continued new logo growth as well. But overall, I think the best way to look at growth today is, of course, ARR. I think shortly, we'll give you some expansion rate metrics as well. The combination of those 2 will help support the guidance that we gave.

And you can really see with that 5% number, that there was nothing that we were doing special incentivized. It's either within the customer base or the sales force to try and drive that to a different number. And there's nothing we're doing this year, and there's nothing I can conceive of doing in any other year to pay for that conversion. We are more than happy to have customers layer term or cloud on top of perp, it's something they paid for and they rightfully own. We're happy for those maintenance dollars.

But as Jason said, what we're really focused on is the overall expansion and success rate of those term and cloud contracts. And that's where all the core growth is going to come from.

Our next question comes from Matt Hedberg from RBC Capital Markets.

I'll echo my congrats on the ARR outlook. I think that's a lot better than pretty much any of us were thinking.

Two questions. One, last quarter, you guys talked about 50 customers that have converted to sort of more predictive pricing, and you'd really seen a dramatic sort of increase in their footprints. I know it's still early, and there's another quarter, I'm wondering if you could give us an update on sort of how that's progressed sort of the conversions.

Yes. We -- actually, we've got a very focused pricing group that -- and paired with the marketing team, is driving high visibility on making sure that our customers have the new pricing, whether it's predictive pricing bands on data volume or whether it's infrastructure-based pricing exposed to them at SKO. My keynote heavily focused on hey this pricing is really important, and I expect you to view the -- bring it to your customers. And Susan reechoed that, and then Sendur, our Head of Product, reechoed that. And then Christian, the Head of Global Field reechoed it over the course of 3 days.

So we are seeing a lot of positive signs. There's not a customer that we've encountered yet that doesn't feel this is a big advantage to them. Virtually every customer that has looked at the new pricing and digested it has -- one, they've made the decision that's right for them. Some people looked at infrastructure and then went back to a predictive pricing data volume-based piece and others went the other way. They really went back to the use cases they had and what was the highest value return equation for them.

But a consistent feedback I've gotten is, all right, got it, thank you. Now I can focus what I wanted to focus on, which was value and number of use cases, and making sure that I'm driving the change I want across the company. So as I said last call, we are extremely focused on making sure that from an awareness perspective, a tooling perspective, so customers and reps know how to evaluate the pricing and make the best decision and a customer satisfaction perspective, that we are extending this to every customer out there.

That's great. And then maybe just quick follow-up. I know that there's always been a debate with you guys on the land versus expand. I know we want both. And there's obviously a large expansion opportunity in the base. But with the new pricing, Doug, I'm wondering, as we move forward, say, the next couple of years, would we expect new customer adds to perhaps accelerate from kind of the 400 to 500 customers you guys have historically been adding per quarter?

I am still deeply expectant on that movement. When I look at what have we done over the past 3 years since we laid a bunch of metrics on what we -- how are we going to try and convert the overall company? We really impressively overachieved on what we -- on what we thought was a relatively decent lean forward. We predicted $2 billion in revenue. We just closed $2.359 billion. We exceeded our upper end of our op margin target. We gave a 12% to 14% range. We were at 14.2%. We say we moved renewables from less than 50% to 75%. We just ended at 95% for the year, 99% for the quarter. We said deals over $1 million would hit 300. We just delivered 494 for this year. The overall cloud percentage of software bookings we thought we were being super aggressive at 26%. We just finished at 35%. Cloud revenue, we predicted $250 million. We just ended a $312 million in cloud revenue.

So the -- what all that tells me is there has been so much momentum within the customers that understand the value, that no matter what we do to try and entice people to look at net news, that has -- one metric of all the metrics we gave that's kind of stuck. And I think our cloud offerings, they have to help because it's so much easier to get up and going in cloud. But we just need the reps and the partners to find ways to pay attention to new customers. And as long as current customers are expanding at the rate they are now, I know it's going to be a constant tension push for us to achieve.

Our next question comes from Phil Winslow from Wells Fargo.

Congrats on a great quarter and outlook. Two questions. First one for Jason. On last quarter's deck, you gave an operating cash flow chart that showed the trajectory from '20 to '23.

In fiscal '21, you essentially guided to a negative operating cash flow, effectively in line, maybe a little bit lower than fiscal '20. Has that changed at all? Or maybe just directionally give us a sense. And then one quick one for Doug, after that.

Thanks, Phil. So first, on cash. What we said last time, and I'll stick with it, was -- or I'll reaffirm it, and that is that Q -- or that last year, FY '21's cash should look the same as FY '20. So we were minus $288 million. So that's a good forecast for this year.

That snapback starts to occur in the middle of '22. So we'll flip to positive then, and then we will be approaching $1 billion in '23. So no change to the cash guidance.

Okay. So no change in the past. Great. And then, Doug, a question on the data stream processor. Obviously, there's a lot of interested in that at comp and as well as just our follow-up conversations with customers. I wonder if you could give us your perspective now that DSP has been out there for a bit, what you're hearing from customers.

Yes. I think a lot like the observability suite. Really positive indicators with a quarter and a month of observed buying patterns. It was included in a number of deals in Q4. The biggest transaction in company history that we talked about and couldn't name the customer, I can say with assurance that would not have happened without DSP. This customer processes more transactions per second than I think just about any customer in the world. And DSP was key to be able to have a high-fidelity transport and metric extraction across the millions of metrics per second that they drive.

So I think we made the right investment. Streaming is, I think, a very important and critical complement to data at rest. And that's -- it's absolutely something we'll track over the course of the year. I think the real traction will be what we just saw with this big deal, which is attaching DSP to specific use cases. And app-dev is a key one because the volume of data is happening across that next-generation environment. But we see the same thing with AIOps and the ITOps landscape and certainly trying to get real-time bead on events and incidents in the current system. Security is an important component to that as well.

Our next question comes from Raimo Lenschow from Barclays.

Let me echo Kash's excitement about the numbers. Quick question on the cloud transition that we see this year. That's a huge change and a huge momentum. Can you talk a little bit about what's going on in terms of the salespeople pushing it? But I'm also surprised to see that have much of a shift because you still have data gravity as well on the locking side. So can you talk a little bit about the drivers there? That's pretty amazing.

Yes. It was -- while we're reluctant to give guidance for FY '21 was, you really need -- we need to see a full year pattern with FY '20 to even have the beginnings of a model for FY '21 and the one thing that I think Jason and I will stand behind is, we're probably wrong with the FY '21 model because it's really hard to predict.

But it did. It was the largest spike that we've ever seen in Q4. And the interesting -- even further interesting thing that surprised me is we had a little bit of a decelerant on cloud this past year in the sales force. The value of a cloud contract tends to be higher than term because we are providing a service and consuming infrastructure.

And despite the fact that there was a difference in commission credit between cloud and term, we saw it push through at the highest rate we've ever seen in Q4. The Q1 forecast on a deal-by-deal basis, had an equivalent level to Q4. So that just tells me that, one, the health and quality of that offering has continued to progress. And we've done a lot of really interesting things I think unique in the industry from this program called Autobahn, where you actually get to preview exactly what your production instance looks like as part of the POC. And then when you buy, you just instantaneously flip on to some of the interesting pricing and option characteristics that we're providing on rich features within cloud.

But I think the underlying piece is customers want to go to cloud. They see a really rapid return to value. Despite that, that means that 65% is on-prem. And I've said for 5.5, 6 years, that I can't foresee us ever being 100% cloud. There are workloads and use cases where a public cloud is probably not going to be the right thing, at least the next 5 to 10 years. So if we -- our target now is we think we'll get to 60-plus percent cloud by FY '23, which is driving some of the immediate impacts on revenue by some of the benefits as we begin to lap in FY '20 to FY '23. But I don't think that we'll -- that, that number will get to 95% to 99%.

Okay. Perfect. And then one follow-up for Jason. If I look at the operating margin situation for FY '21 is like what's the, obviously, we had a change in revenue, and that's kind of obviously the impact. But in terms of your investment focus and cadence, like, was there any change? Or is this just what we see in terms of operating margin guidance as the reflection of the revenue situation?

It's a reflection of revenue. Revenue is coming down because of the cloud shift. The reason why we guided to a target of 20% by FY '23 is if you're kind of going to layer in the 200-or-so basis points improvement that we've been targeting, you get to 20%. So it's purely kind of a GAAP effect of the cloud transition -- GAAP income statement effect.

Our next question comes from Brent Thill from Jefferies.

Doug, now that you teased us with the largest deal ever, I think everyone wants to hear if you could help just quantify maybe not the exact size, but help us ballpark it.

And for Jason, there's a lot of questions around how the existing install base will migrate. Also the pricing change in how you factor that into the guidance for this upcoming year.

Thanks, Brent. So my hidden internal goal has been, we've obviously been excited to share 7-figure deals. We now have been sharing 8-figure deals with you guys, which have gone from, one $8 million to $10 million plus in '17 to 35 in FY '20, which was a nice move.

My internal guidepost, I want to see a 9-figure deal. And this biggest deal ever was getting us close. It was trying to chin up to be 9-figure deal. So very proud of that. It's not so much the raw number. It's the implication of what that means as far as the importance to the business. And for organizations that are fully digital and their entire business is online, then having a unified architecture that allows you to manage everything from security, to your data centers, to your app-dev portfolio, and then ideally, the line of business requirements as well. I mean that -- when you're a billions and billions of dollar online company, $50 million, $100 million, and $150 million contract deals like that is a reasonable ask.

In terms of the other -- the second question on the -- kind of the assumptions on the perp installed base transition as well as pricing impacts that we baked into the ARR guidance, I would say, as I said in one of the earlier questions, the perp installed base, that's a -- it's a relatively small percentage that have moved thus far. We've been -- we've had that motion for about 3 years. So it's something that moves relatively slowly. There is, of course, folks who are expanding as they're adding new contracts, who are expanding on top of a perp contract. And so that's why I think the better metric to look at going forward is probably going to be an expansion rate. So I'll be providing that, say, within this quarter.

And then in terms of the pricing impact, we actually started testing that back last summer. And so we've been able to kind of see how that impacted, I think, the field's ability to actually kind of digest that and then help customers understand it. And then see exactly kind of some of the elasticity impacts. That's something that we've actually been kind of watching closely for a while.

So we kind of baked in what the experience we've seen thus far has been and feel pretty confident about the forecast that we provided.

Our next question comes from Michael Turits from Raymond James.

I think at the risk of beating a dead horse a little, can I ask for more clarification of Brad and Brent's question. Can you just tell us for the 50% plus growth in ARR this quarter and the 45% next quarter, how much of that is from migration? So we really know what is the sustainable rate in ARR.

And by migration, you mean customers who are moving from perp to either term or cloud?

Correct. And both the let's call it the inorganic impact of that and then the expansion impact of the increasing ASPs.

Yes. So why do we provide that 5% number? If we look at all perpetual licenses, 5% transitioned -- turned in their perp license and turned it for a term or a cloud license. We're still at 5%. It's below the 1 percentage point metric between Q3 and Q4. Going back to -- we're not incentivizing. There's no benefit for a sales rep to do it. There's no benefit for a customer to do it.

So when a customer -- those 5% that moved looked at the offering and say, I want a single contracting vehicle. I get where you're trying to go, give me all cloud, I'll give you my perp license, send it back. So it's a very customer-driven decision.

For our Q1 and for FY '21, we factored 0 perp customers trading in their licenses for term or cloud. They're all organic purchases of term or cloud. And some of those have come and will come from existing perp customers, but there's no program to have them -- to pay for them to get there.

Okay. So 0 perp is in the guide. In other words, just to be clear, there's de minimis or 0 impact from migration from perp in your ARR forecast?

Yes.

Okay. And just one clarification on RPO because we know what SignalFx was last quarter, what was the inorganic contribution from it this quarter?

We -- yes, so last quarter, we disclosed it because it was the first quarter, and we actually absorbed it partially through the quarter. At this point, we've integrated the offering. So we're not breaking it out.

Our next question comes from Fatima Boolani from UBS.

I have one for you, Doug, and one for you, Jason. Doug, just coming off the heels of your sales kickoff and how you're sort of putting pedal to the metal on the cloud software transition. I'm wondering if you can talk to us about what sales incentive changes and the magnitude thereof you're planning to roll out within the sales organization and both to partners as well?

And how are the new pricing structures, which I imagine have now officially rolled into the price books in the partner channels, how the sales forces are responding to that? And I'll follow-up with Jason.

Great questions. So the excitement at SKO was palpable. You're not always -- I guess, you're going to have [happy years], especially if you're the CEO of the company, but I have -- that's the most people I've ever had come up to me, expressing both clarity of vision and then, ultimately, excited about the comp plan and how we've simplified the business

I think the core of excitement comes from, we've finally been able to move to an ACV compensation structure. TCV is so tricky because you're trying to guess, are you going to do a perp license or a term or a cloud? Are you going to do a 1-year, 2-year or 3-year?

And it's a very inexact science, and it is frustrating for everybody in the -- from the sales rep all the way up to me. And ACV is simple and clear. So there's a straight line of sight for them for the year.

Last year, we had a decelerant on cloud. They didn't get full commission credit for cloud deals because the size of cloud tends to be larger than term. We're trying to neutralize term and cloud. This year, we took that decelerant off. So it's 100% credit, whether it's term or cloud, which I think will have them leaning into cloud because the size of the deals tend to be larger.

And when I also look at the infrastructure-based pricing, the Splunk Virtual Core within a cloud environment, I think that is a really advantageous architecture to cloud discussions. It makes more native sense to the buyer than a data volume-based metrics since they're already used AWS, Azure or GCP infrastructure usage charges. So I think when you're talking about a cloud deal, the SBC is a very tight fit for them. Although we've seen cloud customers choose predictive data volume pricing over SBC depending on use case. And all of that is extended to the partners as well.

Fair enough. And then just using that as a jumping-off point. Jason, for you, appreciate that the operating margin compression on a year-over-year basis is just a mechanical output of the revenue moving lower. But as we think about this transition to cloud and cloud software bookings mix shift, how should we think about the trajectory on the gross margin front? If you can help us with that.

Good question. So what we said last quarter was that the gross margin in cloud, which is where the compression will come. The -- obviously, the on-prem business has been in the kind of 90-plus-percent gross margin range and don't expect that to change. The cloud gross margin, obviously, has come up substantially from being in the 30s last year. Last quarter, we gave the milestone update that it had achieved 50%.

We are -- I think we said we were going to get to 65% to 70% over the next few years. Our expectation is we would expect to see cloud get into the 60% gross margin range by the end of this year. And then probably be approaching the 70% range by the end of next year. So then the question is on overall gross margin, what's your mix assumptions? So you can kind of flow-through. And if you assume 60%, you can take 60% of that 70% gross margin or somewhere around there. And then the rest at the 90-ish-percent gross margin and get to a number that I think you can see where it comes.

So it's still going to -- to be able to hit that 20% gross margin, then obviously, we have to see some further kind of efficiency gain across go to market, R&D and G&A, and we do expect that to happen as well.

Our next question comes from Keith Bachman from Bank of Montreal.

I also had two. On the total RPO bookings, Jason, could you talk about how the migration, the cloud over the next 12 months is going to impact the growth rate? I heard Doug say the deals tend to be larger than, say, the term deals?

And I'm just trying to understand really how does the growth rate of the ARR, which I know you're pointing us to, but just trying to see if there's any call-outs to how you want us to think about the RPO bookings, the total RPO bookings, over the next 12 months? And then I have a follow-up.

It's a good question. I mean RPO bookings, honestly, is a little harder to forecast because duration plays such an important piece of that. And as we are -- we've now got our sales team completely aligned on the kind of more of the ACV versus TCV model, so focused on really making sure we're driving the right ARR growth, that again we're confident in our forecast.

And so because of that, duration is going to move around. And so -- and then, of course, timing, as I said, is already volatile. So as a result, I would expect that you will see the numbers probably -- they probably will not look unlike last year, where you might see some volatility across quarters and the overall growth rate. I have to decide if we're going to start duration adjusting things. It makes it more complicated. But if there's going to be a lot of focus on it, you kind of need to look at it that way. So that's something that I will probably talk about at Analyst Day.

And the duration of cloud is shorter than term, right? So it will...

Well, so today, and I think we'll determine the view -- we'll see if we can create a little more clarity. The one thing I would just make absolutely certain of is the reason we gave the ARR CAGR kind of guidance for the next 3 years is because that's effectively what's going to translate to whatever the duration is.

And so whatever you see in RPO bookings, ultimately, what we're going to deliver is what's in ARR. And so the reason I wanted to -- the reason I wanted to give that ARR guidance is to make sure that regardless of whatever volatility you see in RPO bookings, you're focused on comparing that to ARR because ARR is the one that's more durable, more stable and what actually relates to what we're delivering. And regardless of the timing of when multiyear duration deals hit or all of that complexity. It really -- it can create volatility that honestly does not translate to -- ultimately to revenue and ARR.

And so that's why you're going to hear us continue to push on focusing towards ARR. Once we're all ratable, this won't be -- it will be less complicated, but we're obviously in the period of transition.

Okay. Well, let me ask my second question before Ken cuts me off rudely. You talked about -- you gave a '21 margin, and you talked about a snap back in '22. I just wondered if you wanted this on any business operating margin I'm referring to.

Just give any kind of dimensions on how you're thinking about that word snapback against your longer-term targets?

It totally relates -- it completely relates to the speed of the cloud transition. So if this year goes incredibly well, we'll have an even higher cloud mix than our current plan, which will be outstanding and fantastic for the business. It will make the GAAP metrics more confusing and then snap back shorter.

So the reason I'm focused on '23 is I don't know how -- it's just hard to say how fast it's going to go because we're not going to tell our customers what they should consume. They're going to tell us. And so as a result, we're very confident about where we're going to be in '23. But where we're at on the halfway point at the -- in '22 is a little harder to see. So that's why we gave the guidance the way we did.

Our next question comes from Mark Murphy from JPMorgan.

Doug, following up on an earlier question, you had said that you recently closed your largest order in company history. So I just wanted to clarify, I'm presuming that, that was in fiscal Q4, but is there any chance that, that or part of that was in Q1?

And then I think you were -- you sized it up in terms of TCV. Can you help us at all to try to roughly understand the ARR size of that? And then I'll have a follow-up.

Yes. So the transaction itself closed in Q4, in January, which, again, looking at the volatility of this environment, and it was a nice testament that despite already some craziness in the overall global markets that this customer is willing to lean in at the size of the transaction they did. It's -- being a very, very large TCV deal, you can imagine that it may be an 8-figure ARR deal, annual deal, but we're not going to -- we won't disclose the dynamics of that specific deal. Other than -- the purpose why I brought it up was, it was a portfolio buy and it reemphasizes that the 3 years of focus we've had on trying to bring DSP to market and DFS to market and doing the SignalFx-Omnition acquisitions. Now I didn't fathom the security fold in that -- that, that drove the outcomes that we are looking for, which is enterprise-grade features from security to ITOps and infrastructure management to app-dev and dev ops.

Okay. Understood. As a follow-up, on the topic of the coronavirus. What actions are you taking with respect to your own employees to try to safeguard them?

Just, for example, are you pulling out of conferences? Are you restricting travel? And if you are doing the latter, is there any reason to think that some of the on-site implementation work could be impacted at some point, if the problem continues to grow?

Yes. It's -- like, again, assuming many other companies, our employee and customer health is our #1 priority. And so we have issued guidance internally to all employees and all managers that we value health and safety, number one, that please restrict business at -- any travel to business-critical travel. If someone is feeling uncomfortable, they get to choose, even if it's business critical, we'll find -- we'll honor the employee and/or the customer and make sure that we're keeping health and safety top of mind.

For noncritical internal meetings, we've begun to move some of those to be virtual. For external conferences, so far, we have not deemed that we need to retract. And again, I guess, we're probably lucky enough on our SKO timing that it happened before some of the latest swells within the United States. Where we've got a really good safety and security group, and they're issuing updates multiple times per day. We're staying as on top of this as any other medium to larger-sized company can be.

In Asia Pacific, we've rotated a lot more of our professional services work to virtual work wherever we possibly can, wherever it makes sense for the customers. There is the possibility there in other locations around the world that if everything move virtual, we all know there are some things, rack and stack hardware that customers usually do, but that might gate our ability to deploy Splunk.

And again, we -- I've got daily calls with our Head of Customer Success, and it's something that we're going to keep as on top of as we possibly can, given how dynamic the environment is.

Our next question comes from Steve Koenig from Wedbush Securities.

One financial question and one quick product question.

On the financials, the ARR metric is key, and that's really helping us understand the business. We do keep getting surprised with some of the other metrics. I'm just wondering, so that we can calibrate going forward, can you share with us or give us some understanding of what you're planning for the cloud mix this year? Should it be like a linear progression towards that 60%?

And what's the impact on revenue, I don't know, per point of mix shift? Or how do we think about modeling that?

So we're going to have our Analyst Day, I think, in 20 days. So we're going to -- we'll probably shed a little more light on exactly what -- how to kind of think about those things. I mean, I'll just say, high level, we went from 26%, 27% in '19 to 35% this year. I expect it to be, certainly, into the 40%-plus range next year.

And so in terms of how linear that's going to be, well, I'll say, stay tuned because it is -- again, it's customers that are choosing. And so it's hard to -- it's a little hard to predict. And that's why the ARR metric is really the one that's most important to focus on.

Got it. Okay. And Doug, for you. On the cloud adoption, can you maybe just give us a little color on what kinds of use cases? How much of the data is being forwarded from prem and it's being used as a deployment option versus de novo cloud apps? And kind of how does the mix do between ITOps, app security? Does it kind of mirror your premise mix?

So just some understanding of what it's being used for.

Yes. Three years ago, I would say that security was definitely a smaller component and IT and other use cases were larger. The big shift that I've seen in the past 24 months is the comfort, even with some relatively conservative -- excuse me, regulated industries and companies with moving to cloud. And with that, the growth -- we've seen a growth in security data going into cloud, and then the deployment of enterprise security and Phantom to complement that.

The -- what I would say, 3 years ago, probably a higher majority of the data was cloud-based data. But again, as people become more comfortable with cloud, and they're moving more workloads, and it's a classic hybrid environment, we're seeing that it depends on the use case, but a healthy mix of on-prem data streaming to cloud as well as cloud data being gathered either needed within the cloud, if you're very centered on AWS or having your sales force in Okta and Workday data go in to commingle with the other data.

Our last question comes from Karl Keirstead from Deutsche Bank.

Maybe 2 for Doug on this cloud mix. Doug, as you see a larger mix of Splunk deployed on cloud. One would assume that your competitive set might change such that you might start seeing customers evaluate cloud-native alternatives, such as Microsoft Azure Sentinel. I'm just wondering if, in fact, you're seeing that. And as a result, how does your pitch in the field change?

And then I'll ask my second one. Of all the Splunk on cloud, Doug, can you comment, is the bulk of it occurring on AWS? Or are you starting to see some Splunk on Azure or Splunk on some other clouds?

Absolutely. So I'll take your last one first, just as context, and we'll go back into what does our cloud growth mean overall?

So our SaaS offering, our cloud offering, which is now 35% of revenue, that is -- that is wholly on AWS. We have architected that solution so we could begin to spread to other clouds besides AWS. But I would foresee AWS being our Tier 1 strategic vendor for as long as I can see out given the relationship we have between the 2 organizations.

But as we all know, customers have a growing bias on whether that offering is housed in AWS or in Azure and GCP, and we want to be able to play across all the clouds. From a -- how does that change the dynamics? Even when companies are going hard at cloud, what I've certainly seen for the Global 5000 is 90% of them have a multi-cloud strategy. I think so many of us have been on this road before of trying to be an IBM-only shop or HP-only shop or Microsoft-only or Cisco, and that doesn't tend to work out super well for the IT department and for the company overall.

So I think this time, especially given the -- given more complete control and value that the cloud providers are delivering, I think people are really, really attuned to, I better find a way to make sure them spreading the workloads around across the cloud. And I'd better be smart about where I'm storing my data because the egress costs can be ridiculous, which I think is a good value prop for Splunk, if you store it within Splunk. And we've got some interesting stuff that we have on the road map, we're going to help you ensure that the data is yours, that it's masked from anybody, and that, that data will be portable for you.

When I look at higher order offers. If I go above the data as a service landscape to something like Sentinel, last year at RSA, there's a lot of discussion about, hey, what's going on with Microsoft and Google and others as far as their security offerings?

I think after a year of evaluation, what we heard over and over at this year's RSA was very little discussion about Microsoft or anybody else on the security front. Almost no discussion about Open Source. I think the maturity -- where, obviously, this was last week. So we've already got some of the effects of coronavirus and people being concerned about the volatility of the economy and that would -- what last week told me is people are kind of circling the wagons and making sure that they have vendors they can trust, that the value is super clear on what they're actually working with and if this goes with any patterns we've seen in past economic disturbances. The companies are very, very customer centric. And have a track record of value and delivery wind up doing pretty well.

And that, in my view is, in a heterogeneous cloud world, you better have a vendor that's going to be able to span all those clouds and non-clouds. And ideally provides you and insulates you from the lock-in that customers are afraid of. And I think we've got a very strong value prop there.

Thank you. This concludes the Q&A session. At this time, I'd like to turn the call over to Ken Tinsley, Corporate Treasurer and Vice President of Investor Relations for closing remarks. Please go ahead, sir.

Great. Thanks, Dilem. Now we appreciate your help today, and thanks, everybody, for joining us. Hope you have a good night.

Thank you, ladies and gentlemen. This concludes today's conference call. Thank you for participating. You may now disconnect.