數據狗 (DDOG) 2020 Q3 法說會逐字稿

Ladies and gentlemen, thank you for standing by, and welcome to the Datadog Third Quarter 2020 Earnings Call. (Operator Instructions) Please be advised that today's conference may be recorded. I'd now like to hand the conference over to your host today, Mr. A.J. Ljubich, Director of Investor Relations. Please go ahead, sir.

Thank you, Liz. Good afternoon, and thank you for joining us today to review Datadog's third quarter 2020 financial results, which we announced in our press release issued after the close of market today. Joining me on the call today are Olivier Pomel, Datadog's Co-Founder and CEO; and David Obstler, Datadog's CFO.

During this call, we will make statements related to our business that are forward-looking under federal securities laws and made pursuant to the safe harbor provisions of the Private Securities Litigation Reform Act of 1995, including statements related to our future financial performance, including our outlook for the fourth quarter and for the full year 2020, our strategy, potential benefits of our products, partnerships and investments, anticipated hiring, our ability to capitalize on our market opportunity and the impact of the COVID-19 pandemic to our customers using our platform, and industry trends as well as our ability to benefit from these trends. The words anticipate, believe, continue, estimate, expect, intend, will and similar expressions are intended to identify forward-looking statements or similar indications of future expectations. These statements reflect our views only as of today and not as of any subsequent date. These statements are subject to a variety of risks and uncertainties that could cause actual results to differ materially from expectations. For a discussion of material risks and other important factors that could cause actual results to differ, please refer to the quarterly report on Form 10-Q for the quarterly period ending June 30, 2020, filed with the SEC on August 10, 2020. Additional information will be made available on our quarterly report on Form 10-Q for the quarterly period ended September 30, 2020, and other filings and reports that we may file from time to time with the SEC.

Our filings with the SEC are available on the Investor Relations section of our website. A replay of this call will be available there for a limited time.

Non-GAAP financial measures will be discussed on this conference call. Please refer to the tables in our earnings release, which you can find on the Investor Relations portion of our website, for a reconciliation of these measures to the most directly comparable GAAP financial measure.

With that, I'd like to turn the call over to Olivier.

Thank you, A.J., and thank you all for joining us today. We are very pleased with our performance in Q3, which showed continued high growth at scale and demonstrated efficiencies. It was an exciting quarter in which we maintain our high velocity of product delivery, starting with the new products and features announced at Dash, our annual user conference. We are building on our strong track record of innovation and expanding our lead as the most complete and cloud-native end-to-end observability platform.

With the majority of our employees continuing to work from home, I am extremely impressed by our productivity. Our engineers continue to build and ship innovative solutions. Our go-to-market teams continue to efficiently deliver value to our customers, and we continue to hire rapidly across R&D and sales and marketing to best position ourselves for the future.

We are strategic partners to our customers like never before as the importance of being digital first and agile is more pronounced than ever. Companies everywhere are continuing to migrate to the cloud and invest in their digital operations, especially [dot coms]. And as a market opportunity, driven by massive replatforming and cloud migration, has become clearer during this time, so has our ability to execute against it.

Now on to a review of Q3 results. To summarize Q3 at a high level, revenue was $155 million, an increase of 61% year-over-year and above the high end of our guidance range. We ended the quarter with 1,107 customers with ARR of $100,000 or more, up from 727 last year. These customers generate about 75% of our ARR. We have about 13,100 customers, up from about 9,500 last year, which means we added about 1,000 customers in the quarter, meaningfully more than the 600 that are in Q2.

We also continue to be capital efficient with free cash flow of $29 million. And as in past quarters, our dollar-based net retention rate was over 130% as customers increased their usage and adopted our newer products.

Now to review Q3 in more detail. Throughout the quarter, usage growth of existing customers was robust, which was a return to more normalized levels after slower usage expansion in Q2. To be more specific, the pace of user growth in Q3 was broadly in line with pre-COVID historical levels. As a result, we feel comfortable that some of the rationalized cloud usage from our larger customers that we've seen in Q2 was transitory as many of those customers have now returned to steady growth for multiple consecutive months. Strength was also broad-based across customers of different sizes and within different industries.

In addition to that, new logo generation continued to be robust with customers additions in line with pre-COVID levels and churn remain consistent with historical rates. Taking all of this into account, total ARR at the end of the quarter was a new record for the company, making this a very successful quarter.

Next, our platform strategy continues to resonate and win in the market. As of the end of Q3, 71% of customers are using 2 or more products, which is up from 50% last year. Approximately 20% of customers are using 4 or more products, which is up from only 7% a year ago. We had another quarter in which approximately 75% of new logos landed with 2 or more products. And we continue to be pleased with the uptake of our newest products, including Synthetics from NPM and security.

I will point out that synthetics has now been commercially available for about a year; and today, it is used by thousands of customers, has reached 8 figures of ARR and continues to be in hyper growth. Adoption of Synthetics has exceeded our expectations, which I will attribute to the combination of the strength of the product itself and the power of our platform. As a reminder, frictionless adoption is a key value proposition of our platform, which we expect would benefit all of our products.

To conclude my review of the quarter, our ability to both land and expand during what has been a time of uncertainty demonstrates that is of importance as companies of all sizes and across all industries, even in the most challenged sectors, are turning to their digital operations as the most strategically important segment of their business.

Now on to R&D. We have a proven and long track record of innovation, and our team lived up to that standard in Q3 with the introduction of 8 new products and major features at Dash. Those announcements include: the introduction of the Datadog Marketplace to enable technology partners to build applications on top of our platform; the general availability of Continuous Profiler to measure code-level performance to an always-on and low-overhead solution; extending Synthetics to continuous integration deployment pipelines, which ended testing earlier in the development process; the introduction of Mobile Real User Monitoring for both Android and iOS; the general availability of Error Tracking to aggregate triage and prioritize front-end application errors; the beta launch of Incident Management for dev ops and security teams; the beta launch of Compliance Monitoring, which extends our security solutions to proactively notify on misconfigurations and compliance drift; and lastly, the beta launch of Recommended Monitors, a suite of preconfigured, curated and customizable alerts.

Additionally, we recently announced a strategic partnership with Microsoft currently in public preview, which will make Datadog available to purchase, implement and use directly from the Azure console and Azure and Datadog sales teams will increase collaboration for co-selling to enterprise clients.

Today, we also announced an expansion of our strategic partnership with Google Cloud Platform, which extends our GCP presence into new regions and enhance our go-to-market collaboration and sales alignment between Datadog and GCP. We believe these partnerships, along with our existing alliance with AWS, demonstrate our leadership in cloud environments as well as the collaborative nature of our relationships with the public cloud vendors.

These are just a few of the new features and enhancements we shipped this quarter. Rather than listing them all, I'll conclude on R&D with 2 main takeaways. First, I am very proud of the continued productivity of our teams. Together, we have not allowed the pandemic and work from home to slow down our road map; and we have also been able to successfully hire and onboard at scale throughout the challenging times. Second, we continue to deliver the most complete and cloud-native end-to-end observability platform; and yet, we are only getting started.

Now on to sales and marketing. As you know, this quarter, we hosted Dash, our annual user conference. While this was our first time hosting it in an all virtual format, this enabled us to reach a broader audience of over 7,000 attendees, which is more than 5x last year's count. That was a great success, bringing together our customers, prospects and partners to show the power of Datadog and many of the new products discussed earlier. And I want to give a special thank you to our events and community teams for excellent execution in a quickly shifting environment.

Now let's discuss some of our wins this quarter. First, I'll highlight 3 notable upsells that demonstrate the broad move to digital channels that has been catalyzed by the pandemic and our ability to rapidly scale with our customers. First, the 7-figure upsell with a large Latin America e-commerce company that has been handling record level of orders. This company brought down silos and now has hundreds of users in Datadog collaborating around the shared view of the IT stack; next, a European on-demand delivery company that has seen its business more than double from last year and has grown its Datadog usage more than 4x. The company has standardized monitoring on Datadog with a 7-figure deal featuring all 7 of our platform products, including security to enhance real-time threat detection. And last, a U.S. gaming company that has seen material growth in their platform and now has a 7-figure commitment to Datadog. In addition to using all 3 of our pillars, they are using Datadog to support their serverless architecture they harness on machine learning to detect anomalies before they occur, and they also report on key business metrics to finance and advertising teams.

We also had good success with customers from traditional industries that need to transform. We had a 6-figure new logo win from a 150 years old postal service in Europe. This organization aims to deliver more digital services to its customers while undergoing a transition to a multi-cloud and containerized environment. They are using 4 products, including Synthetics, to monitor back to front end across both on-cloud and on-premise environments.

Next, we had a sizable upsell to a European financial service institution. After joining as a customer just over a year ago, this company increased spend more than 4x for their consolidated monitoring on Datadog to now exceed $1 million in ARR.

Finally, we had an upsell with a 100-year-old global shipping company, which we mentioned on last year's call for Q3. Back then, this customer was spending mid-6 figures on Datadog infrastructure monitoring. Today, this customer's ARR has crossed into 7 figures using infrastructure monitoring, APM, logs, Synthetics, RUM, NPM and security. This is a great example of not only how companies of all stages are undergoing digital transformation but also of our powerful cross-selling motion as we introduce new products to market.

Now moving on to our outlook. As we look ahead to the final quarter of 2020, we continue to be excited about the market opportunity ahead of us, and we are confident in our ability to execute given continued strong performance through challenging times. After some of the rationalized cloud usage we saw in Q2, we've seen a clear return to normalized usage growth. It is apparent that cloud migration is not only resilient in the current environment but may even grow stronger longer term. Companies globally and across all industries are prioritizing digital operations like never before, and the cloud is a clear strategic winner to enable greater agility and innovation.

We continue to believe Datadog is a primary beneficiary of this trend and that we remain very well positioned to win in the market. In other words, while the near-term macro environment remains uncertain and could cause bumps along the way, we are very confident in the long-term opportunity and our positioning. And we believe that we can continue to sustain strong growth both in the near term and over time.

With that, I would like to turn the call over to our Chief Financial Officer, David Obstler. David?

Thanks, Olivier. As mentioned, we delivered strong third quarter top and bottom line results amid a difficult macro backdrop. Revenue was $154.7 million, up 61% year-over-year. Usage trends were robust and returned to more normalized growth after the pressure that we saw in Q2. Meanwhile, platform traction continued to be strong. New logo generation was robust, and churn was in line with historical norms.

To provide some more context, first on usage. Usage from existing customers was robust, and our third quarter dollar-based net retention rate remained above 130% for the 13th consecutive quarter. After some pressure seen in Q2 driven by optimization efforts from larger customers at scale in the cloud, Q3 was characterized by a decisive return to more normalized growth from our existing customers.

Throughout the quarter, we saw usage growth that was more in line with pre-pandemic historical levels. The trend was broad-based and sustained throughout the quarter. This provides us with confidence that what we experienced in Q2 was a transitory optimization effort that were related to the challenging macro environment. While further optimization may happen periodically as we've talked about previously, we feel confident that cloud migration is very much intact and perhaps even strengthening longer term.

Recall that we have a ratable SaaS model. Therefore, while Q3 usage growth was back to pre-COVID levels, the pressure experienced in Q2 can still be seen in our year-over-year comparisons for a number of quarters. Our powerful land-and-expand model continues to be driven by both usage growth of existing products as well as the cross-selling to our newer solutions.

Next, in the third quarter, we saw continued strength in our platform strategy, with over 70% of our customers using 2 or more products and about 20% of our customers now using 4 or more products, up from only 7% a year ago. Continued product traction is being driven by adoption in the initial land as well as strong cross-selling. The newest products continue to perform well, growing run rate quarter-over-quarter. I note that many of them, such as RUM, NPM and security, are still early and, therefore, small contributions to results in the quarter.

Next, new logo results were solid, with customer additions in line with pre-COVID levels and strength across sales channels and regions. And lastly, churn was stable, in line with historical levels and improved from the slight elevation seen in Q2. Our dollar-based gross retention rate has remained largely unchanged in the low to mid-90s.

Now turning to billings, which were $155.9 million, up 39% year-over-year against an exceptionally difficult compare. We have said that we will provide a pro forma to more clearly align billings growth with the economics and growth of the business. And in Q3, there were a number of timing and duration differences that affected the growth of billings in the quarter.

First, on last year's Q3 call, we had disclosed that there was an invoice timing difference, which increased billings by $6 million. Pro forma, excluding that bill, billings would have been $106.4 million. Next, in the quarter, we had an $11 million impact to billings and deferred revenue from a few long-term large customers, which moved from annual billings to semiannual or shorter durations upon their renewals. Notably, all of these customers continue to commit on an annual basis. They're simply breaking up their bills, as we talked about on the last call, into smaller increments. If we pro forma for these timing effects, Q3 '20 billings would have been $166.9 million. So now comparing the pro forma billings of $166.9 million to the year ago apples to apples of $106.4 million, the growth is 57%, relatively in line with revenue growth.

Remaining performance obligations, or RPO, was $316 million, up [50] per year. Current RPO growth was similar to pro forma billings growth. As a reminder, billings and RPO can fluctuate, as we've just discussed, versus revenue based on the timing of invoicing and signing of customer contracts, while revenue incorporates customer usage.

Now let's review the income statement in more detail. As a reminder, unless otherwise noted, all metrics are non-GAAP. We have provided a reconciliation of GAAP to non-GAAP financials [in the] release. Gross profit in the quarter was $121.5 million, representing a gross margin of 79%. This compares to a gross margin of 80% last quarter and 76% in the year ago period. Year-over-year improvement in gross margin was driven by more efficient use of cloud infrastructure. The slight decrease in gross margin sequentially was due to minor inefficiencies created from our investments in products and platform innovation, as we have discussed. As a reminder, our gross margins may fluctuate quarter-to-quarter within an acceptable range as we prioritize product development and innovation as well as the build-out of cloud data centers in newer geographies.

R&D expense in the quarter was $45.8 million or 30% of revenue compared to 28% a year ago. We've continued to invest significantly in R&D, including high growth of our engineering headcount. We've been able to attract talent and execute on hiring and onboarding, enabling us to deliver multiple record quarters of engineering headcount additions. We continue to see a meaningful opportunity to innovate and expand our platform and therefore, plan to continue to make meaningful investments in R&D.

Sales and marketing expense for the quarter was $49.7 million or 32% of revenue compared to 39% in the year ago period. Similar to R&D, we continue to make substantial investments in sales and marketing, but the pace of revenue growth continues to outpace that investment. This was another quarter of no in-person trade shows or marketing events. While we have successfully redeployed much of the events budget to advertising and other lead-generating activities, it was not on a 1-for-1 ratio.

G&A expense was $12.1 million or 8% of revenue, slightly lower than the 9% in the year ago quarter. And operating income was $13.8 million or 9% operating margin compared to operating income of $726,000 or 1% in the year ago period. In addition to the improvement in gross margin, the continued reduction in marketing events, travel and entertainment and facilities overhead due to COVID contributed to operating margin. Headcount growth was approximately in line with revenue growth in the quarter.

Non-GAAP net income for the quarter was $16 million or $0.05 a share on 333 million weighted average diluted shares. We have a very efficient business model and have experienced a high return on our investments in sales and marketing and R&D. While we have delivered 5 consecutive quarters of breakeven or positive operating income, we note that our priority remains top line growth, and we intend to continue aggressive investments in R&D and go to market.

Finally, turning to the balance sheet and cash flow. We ended the quarter with $1.5 billion in cash, cash equivalents, restricted cash and marketable securities. Cash flow from operations was $36.3 million in the quarter. After taking into consideration CapEx and capitalized software, free cash flow was $28.6 million in the quarter or a margin of 19%. Cash collections have been very strong amid COVID, a testament to the importance of our solution to our customers.

Now I would like to turn to our outlook for the fourth quarter and the full year 2020. While we saw usage growth in Q3 that was consistent with pre-pandemic historical levels, the pandemic is still ongoing and uncertainty remains. Therefore, we are being prudent by factoring into our guidance usage growth trends below what we have seen in Q3 and conservative new business assumptions as well as continued strong investment in R&D and sales and marketing.

Beginning with the fourth quarter, we expect revenue to be in the range of $162 million to $164 million, which represents year-over-year growth of 43% at the midpoint. Non-GAAP operating income is expected to be in the range of $3 million to $5 million, and non-GAAP net income per share is expected to be in the $0.01 per share positive to $0.02 per share based on an approximately 335 million weighted average diluted shares outstanding.

And for the full year 2020, revenue is expected to be in the range of $588 million to $590 million, which represents 62% year-over-year growth at the midpoint. Non-GAAP operating income is expected to be in the range of $48.5 million to $50.5 million, and non-GAAP net income per share is expected to be in the range of $0.17 to $0.18 per share based on approximately 332 million weighted average diluted shares.

Then some notes below operating income. We expect approximately $1.9 million of quarterly non-GAAP other income, which is net including the interest income on our cash and marketable securities less the interest expense on a convertible debt. And we do not expect to be a federal taxpayer again but have a tax provision related to our international entity, which we estimate to be approximately $450,000 in Q4.

To summarize, we are pleased with the results for the quarter. Usage growth was strong as companies are prioritizing cloud migration and digital transformation more than ever, and we continue to execute at a high level. We delivered an impressive velocity of product innovation in the quarter and are investing to continue that track record. While uncertainty related to the macro environment remains, we feel very well positioned to capture what is a large and growing long-term market opportunity.

And now with that, we will open the call for questions. Operator, let's begin the Q&A.

(Operator Instructions) Our first question comes from Sanjit Singh with Morgan Stanley.

Congrats on the Q3 results. David, maybe just to start with you. I think the message I heard off the script was a pretty emphatic view from your guys' perspective that the cloud rationing that you saw certainly improved. They became less of an issue. Expansion trends look like they've gotten back to pre-COVID levels.

If you could just bridge for us the slower revenue growth sort of in the low 60s versus the 80s, but at the same time, I think you mentioned -- or I think Oli mentioned a record new ARR quarter. And if I look at kind of the RPO based bookings, it seems like there's an acceleration there. If you could sort of just help us understand how those 3 metrics sort of tied together and give us a sense of whether the business is truly rebounding versus what seems like slower revenue growth.

Yes. We had -- as we mentioned, organic growth is a very strong contributor, and that rebounded, particularly in the larger customers, to more historical trends. And we continued to have new sales in line. It's the combination of the 2 that contributed to the record ARR growth. And so those are the main factors. The organic is always the majority of the growth in a quarter complemented by the new business.

And just to complement on that -- this is Olivier -- if you compare it to last year, so one thing to remember is we have a ratable SaaS model. So the growth we did forego in Q2 is going to be with us in the year-to-year comparison a bit. Last year at the same time, we had an acceleration also, which makes it for a tougher compare. And it's going to -- the increases in ARR only show up in revenue when they're actually incurred in the usage. So it depends on when we added those in the quarter and also what the exit velocity of ARR was in Q2. So basically, this is how you bridge the record ARR with the revenue as it were not record.

Okay. Understood. That makes sense. And then, Olivier, on the partnership front, you signed a follow-on partnership with Google Cloud that you announced today and then you signed another strategic partnership with Azure sort of intra-quarter. On the Azure front, can you sort of frame out what the opportunity is here? Can you give us a sense of what Azure represents as part of the business today? And with the ability of customers to sort of draw down on their Azure credits using Datadog, how much of an opportunity is this in terms of sustaining revenue growth or potentially accelerating revenue growth with this Azure partnership? If you can just to sort of frame out the opportunity for us, that would be great.

Yes. So first all, I should say, we already have customers lock up providers. We've been working with Azure before. So today, our customer distribution follows a little bit the arrival of the cloud providers. So we still have many more customers and more revenue volume on AWS, for example, than we do in Azure.

We -- what's interesting with this new partnership is that we get more tightly integrated into the Azure console itself, and we're the first vendor to do that. What's interesting there is the user base for Azure tends to be a bit more homogeneous or used to be a bit more homogeneous in the technology they use, mostly for Microsoft. And so the ability to be more integrated with Microsoft really can reduce friction and help onboard new customers onto our platform. So we see that as a great opportunity to get more customers, more usage from Azure as Azure itself is growing and as their customers are diversifying their tech stack.

I should say that this is only one of the cloud providers we're working with. We have -- we're active on all of the other platforms. We announced another partnership with GCP today. We have another one with AWS. So this is an evolution, not a revolution.

To your point also, our customers can use their Azure credits, and there's going to be some cross-selling involved with the Microsoft team. So we're excited about the lift in go-to-market there.

One last thing to remember is that this is still in preview, so the -- we don't expect any significant volume for that in the near term. That's something that's going to kick in over the next few quarters.

Our next question comes from Sterling Auty with JPMorgan.

Oli, you kind of answered it with your last comment there, but I just want to make sure that we level set in terms of the timing of the ramp of the partnerships not only for Azure but also what you announced with Google as well. How should we think of the timing of that flowing into revenue?

It's not going to be immediate, all right? So the Azure one is just in preview right now. It's not completely live yet. The GCP partnership involves a number of new technical things that need to happen but also some new go-to-market motions we're putting in place. So there's not going to be an immediate impact. But we see that as being potentially meaningful contributors in the mid to long term.

That makes perfect sense. And just one follow-on to that. Are we looking at across the 3 pillars or just specifically for infrastructure on both of these partnerships?

Well, the entry point for most of our customers is usually infrastructure. But the -- as we mentioned earlier in the call, in 75% of the cases, it also comes in with another product. So we see that really as a gateway into the platform and then a way for us to have customers use all 3 or 4 products in the end.

Our next question comes from Chris Merwin with Goldman Sachs.

Okay. I was hoping you could talk a bit about what you saw across the customer segments in the quarter. I think you mentioned that the strength was pretty broad-based. But I guess within the enterprise segment specifically, are you starting to see more standardization around the Datadog platform not just with infrastructure obviously, but the log, APM and, of course, there's much, much broader suite of products that you now offer?

Yes, that's definitely what we see. I mean we see -- I think it's very obvious for customers that they need to integrate various parts of observability together, and then having all that on top of the same platform is desired outcome. So we see more and more of that, but it's not specific to the enterprise. We see that happening over all segments at this point.

Great. And maybe just a follow-up. In terms of usage stepping back up, it's very encouraging to hear. As you talk to customers, is there any pushback on pricing? I mean I would think not given that you're seeing usage step back up here, but just curious what those conversations have been like and how you're thinking about the pricing model from here. Or is there something that really works well for your customer base as it is and there's unlikely to be any evolution of that?

Yes. Look, the one thing I should say is anytime you charge customers for millions of dollars, they're going to ask questions about the price. It's a whole department in enterprises that are built just for that. That being said, at the end of the day, what matters is how much value you deliver for that price; and I think for that, we're doing a pretty good job.

I think the one thing everybody is grappling with is as more and more applications move into the cloud and those applications generate more and more data, what's the best way to align [who is that] being used for observability reasons, for security reasons, other reasons, to align that with the -- as well the value with the price being paid. And when you look at our product offering and what we've been adding, that's why we've added all of these new SKUs to really unbundle what we're doing with the data so customers can really buy what's aligned with the needs they have.

For example, we announced more recently Tracing Without Limits, which is a way to send extremely large amounts of data to our APM but only retain the part of it that actually makes sense to customers and they want to retain long term. We've done the same thing with logs before. We're doing the same thing with our infrastructure product. These are the way we basically put our customers in control so they can align what they pay with the value they get.

Our next question comes from Brad Zelnick with Crédit Suisse.

Great. And congrats on another strong quarter and really nice to see usage growth returning to more normal levels. David, as investors try to think about the growth algorithm for Datadog in the years to come, clearly, you're levered to digital transformations and cloud migration. And we can look to many proxies for this opportunity, perhaps the growth rates of the large public clouds themselves, which, by the way, great to see the deeper partnerships with Microsoft and Google. But I'd be curious to know how internally, from a planning perspective, how you all think about this. And although I get you're not going to give us guidance right now for next year, how should we think about how you're framing it and the rate of investment against that opportunity? And along with that, how would you characterize the ability to hire in this environment?

Yes. Yes, good question. So on revenue, we said that over -- about 60% of our revenue growth comes from existing customers. So we start with the land and expand and look at those cohorts inorganic. And I think we referred to looking at pre-pandemic and historical trends. So that's at the sort of linchpin, making the business relatively predictable even in uncertain times.

And then we look at the market size, the opportunity, which tends not to be a limit. And so what is, is that the execution, how many salespeople we get in, how we can ramp them, et cetera. And we then essentially have some experience in understanding ramp and understanding productivity, and those are the algorithms we use in looking at growth.

I think we feel, and we said this over and over again, that there's a very big market, and we're very early on. So both in terms of product investment but also in go to market, there's a lot of areas that we are still building out. There are a lot of opportunities. There are a lot of successful territories where we have to put more feet on the street. So we tend to build that from a bottom-up with sales headcount, and that's resulted in sales and capacity as we've talked about in the 60% to 70% growth.

And we see that and we see that opportunity and based on our success in bringing people in and getting them productive. As Oli mentioned, we have been successful in COVID in hiring, and it hasn't really held us back. We've seen a lot of good opportunity and continue to do that throughout the pandemic.

Yes. And I'll complement that. Sorry, I'll just complement that on the growth side. I mean, right now, we're so early in the opportunity that the way we think about the way we grow our team isn't directly related to the way we think about the growth we're going to get next year. We really think of it in terms of how fast we can grow them while optimizing for both short-term and long-term growth.

So we are growing the team as fast as we can, basically, and we think there's enough opportunity to justify it. And that's true both on the R&D side and on the sales and marketing side, which is why you didn't see us slow down at all during the -- I would say the heart of COVID in Q2, we maintained our hiring and we are very convinced that the opportunity is there. And just to put that into -- to frame that a little bit, we're growing a lot faster than the cloud providers as a whole, which would means there's plenty, plenty of market for us to get.

Maybe just a follow-up and perhaps put a finer point on this correlation of sales headcount growth with top line growth just as the portfolio continues to expand, especially recently with Dash, all the new product announcements. How should we think about the evolution of the sales force to drive efficient cross-selling and specialization that might be required in order to hit all the different buyers and cover all of the product capabilities within the customer base? Are there -- do you envision having to make any significant changes in the go to market?

It's very possible. Every company that -- which is large scale and grows at some point, starts specializing their sales force. We haven't had to do that at this point. And I would say we're optimistic that, for the observability part of the stack at this point, we've been very successful at having one sales team sell that.

I think the issue might come first for the new categories we're entering such as security, where the barriers might be a little bit different. There, again, these products are new enough. And the customer base is for these products and the, I would say, the customers for which these products are mature enough are a small enough group targeting enough group that we don't need to specialize the sales team yet. But that's definitely something that's on our minds for the future.

Our next question comes from Raimo Lenschow with Barclays.

Two quick questions and congrats from me as well. Olivier, like what are you seeing in terms of competitive dynamics? So we had like someone there in your -- in the broader space is kind of bring pricing down like crazy and is doing a lot of free stuff. Other guys are just trying to broaden the portfolio coming up with new cloud product. Just sort of just a word on what are you seeing in that space.

So broadly, there's no change. The space is very much the same as it's been for the past few years. We don't see anything different day to day with customers. We don't see anything different in the adoption cycle. We hear a lot about it still around earnings time, but that's about it. No big changes.

Okay. Perfect. And then as you think about -- like as you kind of look at the landing motion and expand -- more the landing motion with all your new customers, have you seen like a change in terms of what people are taking up? You mentioned at the beginning of the call, a couple of extra stats around how many modules, et cetera. But do you see a change in nature in terms of people understanding observability better and kind of going from more than just infrastructure and just kind of thinking more about this whole thing holistically. What are your observations there?

Yes. Definitely, customers expect definitely to get -- to do more with observability than just infrastructure. I would say the field is understanding what it needs better there. At the same time, when we land with customers, we -- the balancing act is between having them use more products on day 1 but also slowing down the landing, which is why the number has been pretty stable around 75% of the lands that include 2 or more products. I think that corresponds to the right balance right now between landing fast and landing with more than one product. Again, landing fast is critical. Like it's a very important part of business, and we've been very successful at it with our infrastructure products in particular.

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

Great. At Dash, it was really good to hear the application marketplace going GA, really seems to open up a nice halo effect for developers to build application on your platform. Can you talk a little bit more about what sort of interest you're having from developers?

And ultimately, you're not alone in software vendors doing this. How do you think about monetizing this? Or is it more of just trying to build up more awareness for additional features that you don't deliver as first-party features yourself?

Well, it's a collection of all of these, right? I think we're still super early in this. At this point, we're working with our very first few partners to make sure that we define -- we develop the platform in a way that's helpful to them. And we think if we make them successful, they'll make us successful in the long run.

So as far as monetization comes in, we worry about them more than we worry about us at this point, just to make sure they're successful and they get what they want out of it. So super early. We think it's important. We think it's going to play a big role in the future, but we have many, many more quarters of innovation to come on these to fully deliver on the business.

That's great. And then maybe just one for David. Your explanation of billings was super helpful. Trying to kind of think about that on a normalized basis given all the nuances from Q3 of last year. I guess I'm curious, though, when we look at deferred revenue this quarter, and maybe there's an obvious answer to this, but it looked kind of sequentially flat. Was there anything that impacted deferred this quarter, not necessarily looking on a year-over-year basis, but just sequentially from Q2 to Q3?

Yes. It's the same factor that resulted in the pro forma analysis. There were $11 million of billings from some large customers that had -- would have been 100% in billing, which were chunked up anywhere from semi-annual to monthly billing. So the way to look at that and do the apples to apples would be at that same $11 million, which reflects the economics, and you'll see the deferred revenue in the mid-50s, like the other metrics that we talked about.

I got it. So really just a relic of what happened last year then repeating itself this year in terms of the renewals.

Exactly. It's deals that renewed annually and grew, and we said last time that we're trying to work with our clients keep the commitment. These are some very long-term large customers who are staying with us, but to accommodate them in this environment in their requests to chunk up bills, that's what it is.

And just to point on that, we'll see more of that in the future, right? I mean the beauty of our model is because of the efficiency of our go to market, we don't actually care if customers pay us upfront. We want to align with them on that. And so we're -- I'm pretty sure we'll see more of that as we grow.

Our next question comes from Brent Thill with Jefferies.

David, I'm curious if you could just expand on the normalization comment you mentioned from Q2 to Q3, add a little more detail in terms of just overall customer behavior. And I'm curious if you could also just drill down on the security initiative, what you're seeing there and what metrics are standing out to you from that business unit.

Yes. Let me just go over some of the metrics around organic growth and then Oli to security. So what we said in Q2 was that we had some of the larger customers rationalize. So therefore, the slowdown of organic was concentrated first in the larger customers and that about 10% of our customer base and ARR was in impacted industries where there was some pressure. And what we saw in Q3 was a sort of return to normalcy in those, meaning the larger customers continue to now, after that optimization, grow in pre-pandemic rates. And we also saw that the COVID impacted -- the impact was taken -- the medicine was taken, and they also continue to be stable to slightly up. So essentially, it was across enterprise, mid-market and SMB, and it was across also the large customer, the small customer that exhibited similar types of organic growth, which is what we've seen over the long period in the company but saw a different effect in Q2.

Oli, on security?

Yes. On security, I think it's too early to have a lot to share there, but the product is growing very nicely. We're getting great adoption stories from customers, but it's growing very quickly from a very small number. So -- and it's still very early in its life cycle.

So we fleshed out the offering a little bit at Dash with the compliance product. There's a lot more we're working on both of the existing bids and on new bids for security, so let's say, it's still early.

And just to restate some of what David said on the growth, we're very happy with the growth we've seen in Q3. It really showed a reversion to normal for the month-to-month growth. I mean if you look at the monthly growth in ARR, any of the months in Q3, like it could have slowed this month in Q3 or Q4 -- sorry, in Q4 or Q1, they would have fit right there. So I think it's not -- we're very happy with what we've seen. At the same time, we're still very careful of that because we don't -- given the macro backdrop, we're still not quite sure what can happen towards the end of Q4.

Our next question comes from Robert Majek with Raymond James.

As part of the continual shift to a more multi-silo approach, curious if you could elaborate on the experience or results you've had as of late penetrating further into the log monitoring pillar.

Into log monitoring, you said?

That's correct.

Yes. I mean, look, the log product is in hyper growth, right? And it's a very exciting and also a challenging one for customers because it's one of the products for which it's the easiest to generate a lot of fairly useless data and have to pay for it, which is why a lot of the innovations we've made initially on that product were around giving customers more control to align the -- what they pay with the value they get. So we've done quite a bit of that.

As we keep growing that product, we're getting into situations where customers are not standardizing on us for all of their observability, including logs. And we've been basically pulled by our enterprise customers to add a lot of the enterprise features that they would expect from a platform that they send all of their extremely confidential data into. So that's a lot to do what we've been doing more recently.

And then I know we're just 3 months out of Dash, but how has customer feedback been so far on some of the new announced products and features? What's resonated the most with customers?

It's interesting because it's -- we've announced many new products, and they all got quite a bit of attention. I would say, we've got quite a few eyes on the new Incident Management product, which is only in beta. We've got a lot of excitement for the profiler product. So there's a number of different things that customers are excited about from Dash. I think we still have -- in all of those cases, it's still early for those products, so we still have to fully bring them to market. For most of them, we still have to charge for them, and we have to basically work with our customers to make sure that these products fully deliver on their promise and then our vision.

But that's the way we build, right? The way we build is we build with our customers. We put the products in their hands. We make sure they see the value. We make sure they have the controls, so they can align what they pay with the value they get. And then we grow those customers on those products.

Our next question comes from Brad Reback with Stifel.

Great. Oli, as your product set continues to diversify and the revenue stream with it, do you think that has any impact on gross margin longer term?

Today, we don't think so. There's no reason to think we're going to deviate from the model we've had so far. We might still see some fluctuation mostly because of the back and forth between building more products and optimizing our use of the infrastructure because they are the same engineers that do both and also because of our expansion to new geographies and new data centers and things like that, which might push the numbers up and down a little bit. But there's no reason to believe that we're heading to completely a different ZIP code there, at least not today.

Our next question comes from Jack Andrews with Needham.

Congrats on the results. I was wondering if you could just perhaps frame for us what percentage of your deals today are partner sourced and just how we should be thinking about new logos that could be generated from your partnership ecosystem given the launch of your partner network, I think, in January of this year.

Yes. Today, it's a very small part of our business that is partner sourced. And so all of the partnerships we've discussed are basically all upside for us, and that's why we're investing in all of those. We think there's a number of things we can get thanks to these partners, and we're investing in the partner organization in general. We launched our partner program earlier this year. It's starting to see a bit amount of success, but it's still a small part of the business.

Okay. And then if I could ask as a quick follow-up, could you describe some of the -- if you need to make some go-to-market changes, in particular, to scale your federal business and how we should be thinking about how big this business could become relative to some of your other vertical markets.

Yes. So we started building a team for that, right? And we are -- as you know, we have a number of things in process for FedRAMP and et cetera. We think it's going to be similar in many ways to the way we sell to all our customers and different in a few other ways.

In terms of the importance of the business, we think it can be a big part of the business. And if you look at other companies in comparable spaces, like it is a big part of their business, but -- whether it's cloud providers or other vendors in security or observability. So we believe that there's really a real opportunity there. So again, this is all upside for us.

Our next question comes from Ittai Kidron with Oppenheimer.

Oli, I'll start with you on the cloud partnerships that you've talked about and mentioned before. Can you highlight what's unique? Or -- first of all, is there anything exclusive in those relationships, number one?

And number two, in what way will it be either easy or difficult for some of your competitors to replicate either the quality of the relationship or the depth of the relationship with those cloud guys?

Well, I mean there's nothing exclusive in most of these relationships, right? But in many situations, like the way they are implemented and what they actually entail is a little bit different. And they're also different depending on the cloud provider we work with. So for example, the way we integrate -- we're going to integrate with Azure plugs directly to their console, which is not the case with some of the other cloud providers. So that's something that's interesting there.

Now in terms of what others could or could not do, it's hard for me to comment. I mean I believe that it's software, so if we can do it, others might be able to do it. I think what we've proven time and time again is that, because of the overall structure of our products, because of the structure of our customer base and our go to market, we end up having a product that is a lot easier and has a lot less friction to adopt ends up being more widely adopted by our customers and more successful in the end. And I think that's what we -- that would guide basically the way we run all those partnerships. So what attracted us to the Azure partnership was the ability to reduce friction there, and that's why we decided to invest in that first.

That's great. And then, David, a question for you, again, sorry, going back to the duration on that large customers, the $11 million. Can you at least confirm that those customers have expanded at your -- roughly your net dollar expansion rate? Just trying to make sure that this isn't just taking a bill and splitting it by half, but he was taking a bigger bill and splitting it by half.

Yes. No, they've been customers that have grown substantially. One of them is a customer that's more than doubled over the last year or so. So this is merely that we have changed the billing terms, but they are all customers that have been growing with us over the last few years.

Got it. And then lastly, on your underlying assumption for some softness in usage in your 4Q guide. Half pretty much into the quarter, it doesn't sound like there has been any unusual usage softness in that 1.5 months. Correct me if I'm wrong. So aren't you just being a little bit too conservative here? I mean what is it, the scenario that truly worries you with so little time left in the quarter?

Yes. I think, just overall, we've tended to be conservative in our guidance to incorporate usage growth rates that are lower than what we have seen and new logo accumulation that's lower. And I think we said last time that given that velocity and the fact that we're in the pandemic and we can't predict what might happen around the world, we wanted to continue to roll that conservatism forward. So it's really that at the core of the guidance rather than anything that in particular, that we've seen that's different than what we said on the call today.

Our next question comes from Bhavan Suri with William Blair.

Olivier, I guess I'll touch on first a little bit about the technology architecture. This comes up a lot even we're talking to investors. Maybe a little help in understanding. So today, when you look at sort of even the infrastructure monitoring space and the way Datadog works and the sampling model with logs, which data makes sense, which doesn't. But this is idea that with sampling, you're not getting all the data, and some of the vendors in other markets are saying we can absorb all the data. How do you think about that? And would it be difficult for you to absorb all the data? Is that too much data? How do you explain that?

We totally absorb all the data. We totally absorb the data, absolutely. There's no sampling. Sampling is something customers can choose to do if they don't want to retain everything and store everything. But absolutely, we take all the data.

And actually, one thing we announced recently with our Tracing Without Limit product, it actually allows us to be -- we're the only ones to actually take absolutely 100% of the data sent to our infrastructure and available in real time, even in extremely high-volume environments and then [these are the] environments where they're keeping track of millions and millions of requests per second. So we do that, and it's not -- something that the other vendors don't do.

So we -- there's no sampling required. Sampling is just a way for customers to decide, hey, the millions and millions of debug logs my developers are sending, I only need to keep [1%] of them to see what's going on as I retain that in storage for a longer time and [for many things] without having to pay for that.

No. Great. I appreciate the clarity. And my second one is kind of a more longer-term question, but it's about the machine learning. So some of the other vendors are marketing heavily the focal point of machine learning, and I'd love to understand sort of how you think about that vis-à-vis competition. And so to what extent are you seeing APM customers, especially leverage the Watchdog automatic anomaly detection service you introduced, maybe new was last year, 12 months ago? So I'd love to understand a little bit of the thought process there.

Yes. So a couple of ways to [make]. So the first one is we actually think it's a customer model and we're fully SaaS. Machine learning is a strength of ours because we actually see all of the data. We can train our algorithms and all of the data. We can seek to choose the problems and the use cases that we want to solve with machine learning versus the ones that we don't. So it actually strengthens the long-term structural strength for what we do.

That being said, we don't like to lead with the promise of an AI that fixes everything for you because we think, in general, those commitments underdeliver. They look great in the demo, but then a number of specific use cases, you're going to find them to fall short, and that's not what we want. And that's true of pretty much all the products you see there today.

So our approach is to basically underpromise, overdeliver, and we think we have long-term structural strength in there because of the way we run our products and the data volumes that we see, which is not something that most of us [competitors] can do or have.

Now in terms of the adoption we see, we've been purposefully selective in the situation in use cases we saw with Watchdog today. We see Watchdog being used by our customers in real situations, and they rely on it, and that's something that we keep building upon.

Our last question comes from the line of Andrew Nowinski with D.A. Davidson.

Just a quick clarification on the billings. I understand the larger customers are growing at pre-pandemic levels and your COVID-impacted companies are slightly up. So it sounds like the environment and the usage -- the cloud usage significantly improved from Q2 to Q3. But I'm wondering, if you look at your billings on a sequential basis after -- even after normalizing, that is only up $6 million sequentially despite that significant improvement in the environment. So I'm wondering, is there any other factors there that we should consider given that the billings only went up $6 million?

No, we said all along that billings has to do with when bills went out. The advice is to take the revenues for the quarter and then in -- multiply that by 34%, 35% or something, and that's the -- that's -- and then times 12, and that's sort of the linearity. And that's what drives the business.

What we're doing with billings, despite the fact we don't run, is we're trying to clean away some of the noise of when a bill went out this quarter versus that quarter, which isn't germane to the top line growth of the business. So I would say urging everyone back towards the ARR approximation and the top line, and we're just basically giving pro formas here, which all are sort of in the 50s percent and RPO, et cetera, which is more in line with revenue growth.

Okay. And then I'm just wondering if just -- could you touch on the competitive landscape both for the APM space as well as log management?

Yes, what would you like to know on the competitive space? I mean it's very much a same as it used to be, fortunately or unfortunately, a little bit boring. So any particular question on the competitive space?

Well, have you seen any sort of win rate improvement now that your cloud usage is certainly getting better? I'm just wondering if your win rates have improved versus competitors like Splunk and Elastic in the log management space or Dynatrace and New Relic because they (inaudible) in terms of different pricing changes.

Most of our business is not replacements, right? Most of our business is net new. So the -- we do see some replacements from time to time, but that's a small minority of the accounts will end. The world is moving to the cloud. Most companies are new to the cloud. Most companies are new to this environment. And they need observability, and they don't have anything yet in their cloud environment even if they have on-prem, and they start using us for that.

I don't have a win rate improvement to report, but I'm also not unhappy about win rates. Actually, I don't even look that often at win rates because that's not what drives most of the acquisition for us.

That concludes today's question-and-answer session. I'd like to turn the call back to Olivier Pomel for closing remarks.

Thank you. 1 second. My apologies for that. I have a 1 year old who was just barging in. That's what you do -- what you get for working from home.

In any case, in closing, I'd like to reiterate that we are very pleased with our performance in the third quarter. I'm very proud of our execution, and I want to thank our employees for the strong productivity that they've shown during the quarter.

We recently celebrated our 10 years anniversary this summer, and I want to say I'm incredibly proud of the culture we created, incredibly proud of the work we have completion and the value we deliver to our customers. But I'm even more excited about our future, about the strength of our team and about the magnitude of opportunity. So in other words, the message you should get from this call is that we're just getting started, and we're also very excited about it. Thank you all for attending.

Thanks.

Ladies and gentlemen, this concludes today's (technical difficulty)