Good afternoon and welcome to the MongoDB Third Quarter Fiscal 2021 Earnings Call. All participants will be in listen-only mode. [Operator Instructions] After today's presentation, there will be an opportunity to ask questions. [Operator Instructions] Please note, this event is being recorded. I would now like to turn the conference over to Brian Denyeau. Please go ahead.

Thank you, Eely [ph]. Good afternoon and thank you for joining us today to review MongoDB's third quarter fiscal year 2021 financial results which we announced in our press release issued after the close of market today. Joining me on the call today are Dev Ittycheria, President and CEO of MongoDB; and Michael Gordon, MongoDB's COO and CFO. During the call, we will make forward-looking statements including statements related to our market opportunity and future growth, the benefits of our product platform, our competitive landscape, our financial guidance, and the anticipated impact of the COVID-19 pandemic on our business and results of operations as well as on our clients in the macroeconomic environment. These statements are subject to a variety of risks and uncertainties that could cause actual results to differ materially from our expectations. For a discussion of material risks and uncertainties that could affect our actual results, please refer to the risks described in our SEC filings including our most recent annual report on Form 10-K and quarterly report on Form 10-Q. Any forward-looking statements made on this call reflect our views only as of today and we undertake no obligation to update them. Additionally, non-GAAP financial measures will be discussed on this conference call. Please refer to the tables in our earnings release 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 Dev.

Thank you, Brian and thank you to everyone for joining us today. I will start by reviewing our third quarter results before giving you a company update. Looking quickly at our third quarter financial results, we generated revenue of $150.8 million, a 38% year-over-year increase and above the high-end of our guidance. We grew subscription revenue 39% year-over-year. Atlas revenue grew 61% year-over-year and now represents 47% of revenue. And we had another strong quarter of customer growth ending the quarter with over 22,600 customers. We're very pleased by our third quarter performance against a difficult and uncertain macroeconomic backdrop. We saw another record quarter of customer additions both in our direct sales and self-serve channels. As we discussed in the past, given the strong product market fit of Atlas, we decided to make a number of changes that make it easier for new customers to get on to our platform. We continued to reap the benefits of these adjustments in Q3 with record customer growth and strong net ARR expansion. On the self-serve side, our continued efforts to broaden our digital marketing funnel have resulted in over 2,000 net Atlas customer additions in Q3. As a reminder, our self-serve business is increasing not just an important revenue generator in its own right, but also a source of leads for our sales force. Strong self-serve net additions in Q3 indicates that this flywheel effect will continue in the future. I'd like to share some themes we have heard from senior level customers across a large swathe of industries and geographies. Customers are feeling more pressure than ever to innovate quickly to seize new opportunities and to respond to new threats, and 2020 has only exacerbated the need for speed meaningfully increasing the urgency to move to the cloud. When evaluating…

Thanks, Dev. As mentioned, we delivered another strong performance in the third quarter both financially and operationally. I'll begin with a detailed review of our third quarter results and then finish with our outlook for the fourth quarter and full fiscal year 2021. First, I'll start with our third quarter results. Total revenue in the quarter was $150.8 million, up 38% year-over-year. Subscription revenue was $144.1 million, up 39% year-over-year and professional services was $6.7 million, up 19% year-over-year. To put our performance in the quarter into perspective, we thought it would be helpful to provide an update on how COVID-19 has impacted the growth of our business. First, let's talk about new business. Our ability to execute on new business opportunities with both new and existing customers continued to surpass our expectations in Q3 despite a difficult and uncertain macro environment. As Dev mentioned, customers are approaching digital transformation and cloud adoption with a heightened sense of urgency, which our go-to-market teams have capitalized on. I want to be clear, even though we performed better than our expectations in terms of new business, COVID-19 did have a negative impact on our quarterly performance. Second, the trend in existing Atlas customer spend has been steadily improving since the modest, but broad based slowdown we experienced in Q1 due to the impact of COVID-19. In Q3, the growth from existing Atlas customers has returned to our pre-COVID trends. Overall, Atlas' strong performance continues to be the largest contributor to our growth. Atlas grew over 61% in the quarter compared to the previous year and now represents 47% of total revenue compared to 40% in the third quarter of fiscal 2020 and 44% last quarter. During the third quarter, we grew our customer base by over 2,400 customers sequentially, bringing our total…

[Operator Instructions] Our first question today comes from Raimo Lenschow with Barclays.

Hey, congratulations from me, a great quarter. One quick question, I see the really nice increase in new customers and congratulations, amazing what you're doing there, but I'm also then seeing a deceleration on growth on Atlas self-serves. Can you just help me understand like how this kind of links together, I assume it's a lower ASP that you are signing there, but just try to help me understand that a little bit. And then one for Michael as a follow-up, like if you think about next year and I think about kind of improvement around profitability, like how do you think about it because you had a lot of like kind of call it like one-off benefits this year with less travel, etcetera, like how should we think about modeling next year? Thank you.

Thanks, Raimo, it's nice to talk you. With regards to your question on self-serve, actually, we feel really good about our self-serve business. What's happening there is that the self-serve is a great source of new customers that makes it just very easy for customers to engage with us in a very frictionless way and then quite a number of those customers end up being leads to our -- both our corporate or inside sales team as well as our field team. So that's what's happening there. It's not because we're seeing any slowdown in our self-serve business.

Okay, perfect. Thank you.

Yes, I think on the second piece of the question, obviously, we'll get to our guide in fiscal '22 in the March call, but I think what we tried to be clear about is, we are definitely looking at the business from a long-term perspective and the guidance in Q4 includes pulling forward some of the hires and other investments that we would have made in fiscal '22 given that we are seeing some COVID-related savings and we talked in an earlier call about our intention and desire to reinvest those now and that continues to be the plan.

Perfect, thank you. Congrats.

Thanks.

Our next question comes from Sanjit Singh with Morgan Stanley.

Thank you for taking the questions and my congrats as well on another strong quarter. Dev, I guess my question sort of relates to the topic of machine learning and sort of the data management and data workflows supporting machine learning. There is definitely a growing ecosystem around that and in terms of the like -- in terms of the application development component where maybe there's a new generation of applications that are going to be more machine learning infused, MongoDB hasn't necessarily been a huge player in that core analytics market, but if you think about supporting this next generation of applications, what's the positioning for the document model in MongoDB specifically for this kind of next generation of applications that are emerging.

So one, I would tell you that we already have a number of customers who have built-in machine learning capability on top of MongoDB. So this has been happening for a number of years. So, two, the power and flexibly of the document model is even more profound as people want to do more and more sophisticated things with applications and keep in mind that MongoDB is a very scalable platform and so being able to have a platform that allows people to leverage the massive amounts of data to train the algorithms is incredibly helpful and just today, Amazon made an announcement around their SageMaker product where we were part of that announcement and our CTO, Mark Porter was quoted where we are, you know, allowing customers of SageMaker to basically leverage data sitting in MongoDB. So from an application development point of view, we're very well positioned. This is a natural path of people building more sophisticated applications and the document model is set up very nicely to help people build more and more sophisticated and complex applications just given the power of the model.

Understood -- well, understood on the distinction there. So let's go back to the acceleration in new customers -- customer adds, it's been going on for a couple of quarters now. I think what also impresses me is the acceleration happening on the direct sales side as well. So you can sort of talk through whether it's the same sort of playbook that you're using to accelerate the self-service side and any sort of indications from these new cohorts that you're bringing on in terms of their profile. Is there any reason not to believe that these customers will expand in sort of the similar sort of trajectory as your prior cohorts before you really accelerate the initiative to increase your customer velocity?

Yes, sure. On the direct side, we actually made a decision prior to COVID becoming front and center for all of us that we just saw last year a lot of friction where our sales people were incentivizing or trying to incentivize customers to make some sort of commitment typically at least an annual commitment because that's the way that they were paid and what we realized with a lot of these customers, sometimes these workloads were new and just didn't know what -- how much you know the kind of resources they would need on Atlas. They just -- didn't -- the friction of getting the approvals to sign a longer-term commitment was that much longer and they liked the flexibility of being able to basically just start using Atlas directly. And so we encouraged -- we saw the strong product market fit of Atlas and once people started using Atlas, the growth was quite profound and so we said, why don't we take away that friction. So, we changed the comp plans for this fiscal year to encourage our sales people to sign up customers more quickly and that has paid huge dividends. It's allowed us to engage customers more quickly and allowed us to acquire a lot more new customers.

Understood. Appreciate it. Congrats, guys.

Our next question comes from Brent Bracelin with Piper Sandler.

Good afternoon. One quick one for you, Dev and one for Michael. I mean, Dev, what type of customer cohort does multi-cloud functionality appeal the most to. Is it the start-ups you think or are there some large enterprises that have appeal there. And then for Michael, EA looked like it was a little stronger than we had expected. Was there a federal tailwind this quarter or were there other factors that kind of drove upside in EA? Thanks.

Yes. So Brent, on the multi-cloud clusters, it actually can -- it just really depends on the sophistication of the organization and how broad a user base and customer base they are going after. For example, we already have customers for example in Canada who are trying to serve the Canadian market and are running their application across different cloud providers all in Canada because they want to leverage the reach -- the broad reach of being able to leverage all three cloud providers and they also wanted to have the benefits of resiliency in case one cloud provider goes down and there were some high profile outages right before Thanksgiving. They don't have to worry about their application going down, and that is, I wouldn't say is one of the -- is not a very large customer. Then, on the other hand, you could also have large enterprises who increasingly want to leverage the different capabilities of the different cloud providers. Each cloud provider is differentiating themselves across different dimensions of capabilities and so being able to do that reach. Again, some cloud providers have more presence in certain markets than others do and also just keeping the vendors honest is something that plays well especially with senior level decision makers who've lived the movie before about being held hostage by particular vendor. So it really ranges depending on the organization, but we think it's very applicable for very sophisticated early-stage companies as well as large enterprises.

Very interesting. And then, Michael on EA, any color there?

Yes, so just quickly. So, not a Fed -- push or Fed impact. EA continues to be an important leg of the growth stool. I certainly understand that Atlas obviously steals a lot of limelight occasionally but EA continues to be quite important and relevant. It's very much a customer choice situation and we're just trying to make it easy for them to consume some MongoDB wherever they are vis-à-vis their cloud posture. Obviously, there'll be variability as we've talked about on EA given 606. I would say there's also variability kind of quarter-to-quarter as we've talked about in terms of relational migrations, but as you heard from Dev in the prepared remarks and some of the case studies, it was a good quarter from that perspective as well. So generally just good broad-based participation.

Good to hear. That's all I had. Thanks.

Thanks, Brent.

Our next question comes from Brad Reback with Stifel.

Great, thanks very much. Dev, as you look at the Atlas self-service business, are there any constraints around how many customers you could add in a quarter?

There's clearly -- there's obviously some limits to how many we can actually add, but there is no natural constraint. It's not like we're constrained by having enough sales people in place. The sign-up process is automated. We can handle tons of volume. So, technically, we can add a lot of customers. What we're obviously continuing to hone is develop the marketing program and the experiments and use content marketing to go acquire new customers and we're getting more and more sophisticated in that. So our self-service pipeline continues to grow nicely.

Great. And then, just one quick follow-up. In the press release you talk about a DoD DevSecOps opportunity. As you look at that longer-term, do you think that gets fulfilled mainly via EA or Atlas? Thanks.

Yes. I think like many customers, I think every organization is in some way on some path to the cloud. We've talked about this in the past. Not only do we see this in the federal sector or even the state and local government sector, but clearly we see this even in the commercial sector where there are certain customers who are much more cautious, maybe because they are in the regulated industry or just culturally they are more cautious, but they are all moving in some way, shape or form towards the cloud. The benefit of using MongoDB is that we give them a very seamless migration path to the cloud and what I mean by that is they don't have to rewrite one line of code to move from an on-prem deployment to a cloud deployment. So that takes away a lot of risk. In some ways, we future-proof their applications. I think with this particular fed situation, I think much like in enterprises and other organizations, I think there's some workloads that are more naturally predisposed to stay on-prem and then obviously there is a big initiative obviously with FedRAMP where you have to certify yourselves against a bunch of criteria to be able to run in the fed -- in the clouds that the Fed's authorize and so that is work that we're already doing because we do see a huge amount of demand there and I should mention that the federal government has been long time users and customers of MongoDB even before I joined the company. So that is a segment that's been early in its adoption of MongoDB.

Great, thanks very much.

Our next question comes from Jason Ader with William Blair.

Yes, thank you. Hey, guys, on the Postgres migration comments that you made, Dev, can you talk about some of the underlying limitations that customers are facing as they do that and how -- and what you guys bring to the table to help them in that modernization effort?

Sure, so there is a number of limitations. One with using a relational database, you basically have to stuff data into a tabular format into rows and columns and that's not the way developers think nor is it the way they code. So what MongoDB allows you to do is one, be able to think about your data much like -- in a much more natural way. So, for example, a customer record does not have to be disaggregated into a bunch of different tables. You can treat a customer holistically and keep everything about that one customer in a document. Second, making changes is that much easier because you can make changes on the fly without having to change all the documents for all the other assets you're tracking nor having to update a centralized catalog etc. So making changes in MongoDB is that much easier versus a relational database. Third is scalability. Relational databases were designed to be single node systems. MongoDB was a distributed -- was designed from the ground up to be a distributed database. So not only do we allow you to move quickly, but we'd like to scale, very, very fast, which is why lots of customers pick MongoDB because they have such high performance and scale requirements. The fourth issue is resiliency. By definition because it's a distributed architecture, we have multiple copies of the data. So if one node goes down, whether it's for a system or a network failure or some other failure, your application is still up and running. You have to do a lot of work on the relational database side to build in that level of fault tolerance and it can be quite expensive and still not really deliver on your requirements. So for all those reasons, people are recognizing that the document model is a very powerful way to model your data, it's the most natural and the most productive way for developers to work which is why we've become so popular.

And just a quick follow-up on that. So understanding all of the advantages, is the issue just from an adoption standpoint that moving to kind of a similar model, relational database in the cloud is just easier and modernization is more of an effort on the part of the customer. Is that the main issue?

Yes, the perception on many customers part, moving from a legacy relational database to an open source database in the cloud will require less work. So when someone is say under pressure to move off particular legacy contract because the contract is up for renewal, they may in their mind believe that moving to an open source relational database in the cloud is a better option. What we are spending time on -- what we are -- the analogy I use with our sales people is that we're selling a very powerful language and for some people, because it's a new language, they've been using the old language for 50 years, it can be a bit intimidating. So we have to educate them on the benefits of our language and when they get educated, they suddenly see the power of the document model, the power of our distributed architecture, the scalability of our platform. They basically just recognize the benefits of our platform and that's why they start moving to MongoDB. And so depending on where they are in their journey, I encourage our sales people to really educate our customers about the benefits of MongoDB and they don't have to sell very hard. As soon as they explain the benefits, customers get it. And as I said in my prepared remarks, we had this banking customer who remarked to us that a number of their people who were long time relational -- I could almost say bigots, just when they started learning about MongoDB became huge enthusiastic converts to MongoDB and we see that happening all over the place and we're seeing that happen now in all these accounts that we've been working in where more and more of the workloads are now coming towards MongoDB.

Thanks very much.

Our next question comes from David Hynes with Canaccord.

Hey, thanks guys, congrats on the results. Dev. So, sales and marketing spend ticked back up as a percent of revs. You told us that would likely happen and you're clearly seeing some nice yield from that spend, what's the plan from here? Is it more of the same or do you think, does it feel like more spend could even yield faster customer growth?

I'm trying to grow -- I think we collected as a leadership team we're trying to grow our sales organization as fast as possible, but there's limits to how fast you can grow because you don't want to cut corners on quality, you don't want to cut corners on investing in ramping these sales people, giving them the right management support and the right technical support for them to set up for success, but I think we've talked about this in a number of earlier earnings calls. We still feel like we're vastly underpenetrated. There are a few slots of the market that we just don't have enough salespeople and so we are trying to grow our reach and expand our reach as fast as we can and that's not just in North America, but in Europe as well as in Asia and we're seeing great demand in all regions even Latin America as well and so for us it's really finding the right leadership team to invest in building a team around them and then scaling that over and over again is really the constraint in terms of growing our sales force.

Okay. Yes, makes sense. And then Michael, a follow-up for you. So a more significant jump in the $100,000 plus cohort this quarter. Is that just timing that more happened across that threshold in Q3 or is there something more insightful happening there?

Yes, I think it's really hard to get super precise in sort of the sequential changes or anything like that in that bucket. Sometimes, we've had quarters where it's been very large. Other times we've had quarters where it's been a little bit lower and all the growth was in people who are already above the threshold. So it's somewhat arbitrary threshold designed just to help people understand that people are making meaningful levels of investments, but I think it is when you get that many number of customers in the absolute who are spending above that level, I do think it's an indication that we're a meaningful part of their technology infrastructure.

Yes, makes sense. Okay, congrats guys.

Thanks, David.

Our next question comes from Rishi Jaluria with D.A. Davidson.

Hi guys, this is Hannah [ph] on for Rishi. Thank you for taking my questions. First, could you just talk about how traction with Realm is going and given it's a newer product, how do you think about your competitive positioning with that product specifically relative to mobile database offerings from other modern database providers?

Yes. So one, I should just remind you that we [indiscernible], Realm product earlier this summer. So we still believe we're in the very early days of this product and I would say the early traction has been really good and we're seeing really nice month-over-month increases in usage, but again, it's early and the revenue for Realm won't show up as a different SKU, it will show up as Atlas revenue because that will drive more consumption of Atlas. And in terms of adoption of Realm as the mobile database, we do recognize that when we acquired Realm, they had not really invested in kind of invigorating the developer community and so we're making investments in doing that as well. Excuse me. And that will also take some time, but we're very pleased in terms of our positioning. One of the killer features that and we will be announcing or we've announced, but will be available -- will be early next year, which is called Realm Sync, which will provide very sophisticated data synchronization capabilities between the client and the back-end. That is one of the hardest problems for developers to solve and automating that will make the platform that much more attractive for developers to build mobile apps on. And we're really bullish on the mobile space because with the advent of 5G and other related technologies, we see that mobile apps will become even mission critical, have much more rich features, streaming types of data and so forth that will require a very sophisticated platform to support.

Great, that's helpful. And then, how do you think about your ability to serve the world's most complex use cases. And what do you feel like technologically differentiates your platform relative to the other no-sequel more modern database providers.

Yes. So I would say, one, when I joined the company about six years ago, I would say that we were considered to be just in a basket of no-sequel vendors, it wasn't clear who the breakout company was going to be. I would say, six years later, I think we've proven that one, we had a significant technology advantage because -- even the size of our community at that time was far bigger than any of the other no-sequel providers. We were able to marry that technology advantage with really strong execution and we've grown our business faster, we delivered better financial performance, we've got more customers than any of the other no-sequel vendors and along the way, we've also done a lot of product innovation, more recently with multi-cloud clusters, earlier with multi-document ACID transactions and a bunch of other capabilities, that's really pushed the envelope. So MongoDB today is not what I'd call a no-sequel database, I would call it a modern general purpose database and we have some of the largest, most sophisticated, and savvy customers across almost every industry and geography using MongoDB to transform their business and we're very proud of that and we feel we're very well positioned to go after the most mission critical workloads no matter the use case.

Great to hear. Thank you.

Our next question comes from Tyler Radke with Citi.

Hey, thanks very much. I wanted to ask you about the sequential improvement that you saw in Atlas. You added about $10 million of incremental revenue, which was up from the $7 million that you added a year ago and up from the $6 million that you added last quarter. I know you obviously saw record net adds and you're seeing some incremental improvement in Atlas expansion rate. I guess what do you kind of attribute the improvement and then the incremental revenue to the most and was there any kind of one-time items that you would caution us just kind of from a extrapolating the sequential growth in Atlas to future quarters?

Sure, yes. Thanks, Tyler. I think the overall -- it was a very strong quarter really across the board, but including within Atlas, we saw obviously a lot of new customer additions, but those tend to be when a customer comes on tends to be smaller and they grow from there. So it's really, it's most of that growth in the expanding -- of the existing customer base. And as I mentioned, we saw both Q1 and then for Q2, we talked about in the various calls, that's slightly slower growth that really reverted more to normalized levels, pre-COVID, which is great to see and I think you see that showing up in the numbers. So, I think it's really kind of all the parts of the chain that you would expect strong additions, good underlying behavior, stickiness, kind of expansion within the customer base and generally the sort of very strong cohort dynamics that we have really just plying outside; so I wouldn't point out [ph]. It wasn't really any one thing in particular or any particular meaningful outliers worth calling out.

Okay, helpful. And then, just wanted to clarify, as you think about the factors impacting the Q4 guidance. Obviously, you called out a tough comp on the EA business, part of that's related to 606 accounting. Are you adding any kind of incremental caution just on the macro environment in Q4 relative to where you were a quarter ago. Obviously, the world is kind of -- it's really evolved to -- the world business environment is at a different place than people thought at the beginning of the year, but just kind of curious how your expectations are for Q4 versus [ph].

Yes, I'd say a few things when looking at the Q4 guide. First of all, obviously, we significantly raised our outlook for Q4. So that clearly indicates our confidence despite the macroeconomic environment. That said and our Q3 strong results notwithstanding, we do expect to see a continued impact on new business. I think it's hard not to. We've seen great customer engagement, but there is plenty of macro uncertainty and I think you can see deals receiving more scrutiny given the current environment and so I think that's sort of captured and reflected. The second thing that I would add from a big picture perspective is we've talked about Atlas. It's great to see that the cohort dynamics and the behavior are back to pre-COVID levels, but we're entering Q4 with just a smaller recurring revenue base than we would have had we not had those COVID headwinds in the earlier half of the year. And then, the third thing I'd point out is just a tough compare which you referenced, which we had called out last year was about $3.5 million incremental of EA and again this has sort of impacted not just the denominator, but also the numerator given how 606 works.

Thank you.

Our next question comes from Patrick Walravens with JMP Securities.

Okay, great. Thank you. And let me add my congratulations. So, Dev. I'm wondering what your R&D priorities are right now and maybe an easy way for us to think about it is, what do you want your software to be able to do in the relatively near future that it can't do today?

Well, I think what I want MongoDB to be known for and I think we're on a path to get there is to be the best place to build modern applications and I believe that the ability to innovate using software and data will determine the company's competitive advantage long-term and that's going to be really, really important. I think the power of the document model, I think our really powerful but easy to use query language, our ability to handle any workload, any size of data all play to our strengths and I think we talked a little bit about, you know, the fact that we have a huge developer community that's only growing by the day and that we provide platform independence. So, I think we have all the ingredients to be a very viable if not compelling place for people to build all their modern applications and that's what we're striving to. Clearly, you're going to see us and we've talked about this in the past, you're seeing us expand from being a database to a data platform. We announced a whole bunch of new products. Those products will continue to grow and mature overtime, we'll probably announce new capabilities over time. So, we're just going to make it easier for developers to build very sophisticated applications, applications that span system of record, systems engagement and systems of insight and we think application of the future will embed all that functionality and we're really well positioned to be the premier place for people to build those applications.

Great. Is there one thing you can say you can't do now but that you're excited that's coming in the next year or two?

Obviously, we're a 13-year old company. So we have a lot of capability and I wouldn't say there's anything that prevents us. I mean, frankly, the one thing I think that really prevents us is awareness or perception of us from a company that we were six, seven years ago. So the more people learn about us, the more people are updated about our capabilities, the more people spend time learning about the technology, the quicker they become enthusiastic converts and that's a big part also of our go-to-market strategy.

All right, great, thank you.

Our next question comes from Ittai Kidron with Oppenheimer.

Thanks. Hey guys, great quarter. Good stuff. Dev, want to start with you for some on Atlas. Clearly, very good progress over there. Can you talk about conceptually how a new Atlas user today is different from perhaps an Atlas user a year or two ago in the sense of who he is and what he or she are doing and how quickly they're growing their usage of the platform?

Yes. I don't know, maybe a couple of years, maybe two, three years. I remember, Atlas was launched in 2016. So I would say our first cohort of customers were probably people who were kicking the tires, trying to really understand how to use the platform, not really put mission-critical workloads on Atlas and so they were probably early adopters. What we've seen now and it's clear is that enterprises are going very aggressively into the cloud. So we have very mature, very conservative organizations who are now using Atlas. What we're also seeing is actually small development teams sign up through self-serve who are parts of big organizations, but they just want to get going on Atlas, play with it, launch some applications on it, get a feel for it as a precursor to maybe a big deployment of Atlas. And then, the other thing that we've done is we just become much more sophisticated in terms of expanding the coverage of access to customers through our self-serve channels. So we're seeing a lot of people sign up for Atlas in other markets, markets in Asia, markets in Latin America and so and so forth that we don't necessarily have a lot of sales people in and that gives us a good way to go after those markets as we continue to build out our sales force. So that's where I would see how the market has or the customer base has changed overtime.

That's great. Very helpful and then Michael, on, I want to kind of dig into a little bit into your return to pre-COVID levels. Could you be a little bit more specific on what you mean by that. Does that mean that expansion rate has bounced back up? ACV -- first deal ACV. Help me amend and what does that mean, if you could break that down for us?

Sure. You know very specifically in Q1, we had talked about how we saw a slowdown in the expansion rate of existing customers within Atlas and then that got a little bit better in Q2. So that's really what we're talking about. So think about the Atlas customers -- Atlas revenues is all consumption based. This is not about deal activity or commitments or things like that. We walked everyone through back then was clearly you can see the adds continue to be strong when we saw slightly slower expansion. One of the first questions was, was there any churn or increase in churn. We haven't seen any changes and that would sort of underscore the mission critical position that we occupy within a customer's environment and what we saw was, if it's not gross adds, if it's not churn, really it's about sort of the expansion dynamics within a customer, what we saw was a slight but broad based slowdown really starting kind of the second half of March in Q1 and continuing into Q2. The first hypothesis was that it was sort of industry driven right maybe industries that were particularly affected by COVID, hospitality, travel, things like that, but instead, it was really much more broad based and modest impact as opposed to a narrow number of customers are having a major impact. We talked about that in the September call that how we've seen some improvement, but not all the way back to historic levels. And so what we're saying here is that based on the data we've observed in Q3, those dynamics are now back to pre-COVID levels. So continued expansion without sort of slight decline that we'd seen. Obviously, continued strong adds and very sticky, which you can see in the overall numbers, but we wanted to try and disaggregate that a little bit for folks and sort of give an update based on the fact that had been a factor and the way it plays out in terms of the Q4 numbers is really, you wind up entering Q4, even though the expansion dynamics are sort of back to normal, you are beginning Q4 with a smaller recurring revenue base than you would normally have if you hadn't had slightly slower growth in Q1 and Q2, etcetera. So hopefully that helps people understand.

Got it. Very good, thanks. Good luck, guys.

Thanks.

Our next question comes from Jack Andrews with Needham & Company.

Good afternoon and congratulations on the results. I want to see if we could dig in a little bit more into the legacy migration opportunity. It sounds like from your prepared remarks that there is a good appetite for these types of projects. I know you've announced some recent partnerships in this area. So who is really helping you go after this market opportunity today and who do you think can help you tackle this moving forward.

Yes, sure. So just to make sure people understand why we see there's a big opportunity. The premium innovation has never been higher, because obviously what COVID has done has created both opportunities and risks for companies and people need to move very, very quickly to either seize opportunities or respond to new threats of the business. And so that means you need to innovate very fast. What that means is that you need to be able to and obviously innovation today means being able to use software and data to build new applications, to add new features, to add new capabilities etcetera and that means you need a database platform that enables you to do that. One of the challenges of relational databases is that as you add more and more capability, the data model becomes increasingly more and more brittle, so it becomes harder and harder to add those capabilities. So when you're in an environment where you need to move fast, you start saying do I want to stay on this on this brittle platform or do I want to move to a much more modern architecture where I can move very, very fast as well as change directions quickly and that's what's driving this demand to modernize our legacy application portfolio and so we feel that we're really well positioned for this opportunity. We're a proven technology, we have reference customers in almost every industry and so we think the legacy modernization or migration opportunity is really big and what we're seeing now is in terms of partnerships we've historically had a lots of SIs who worked with us because their business model is very complementary to ours. In some ways, they take on the development onus for the customer and so we just announced obviously people who are joining our program to help customers migrate off relational databases and these are some of the largest SIs in the world. We also work with some really interesting boutique SIs who have some real deep MongoDB expertise. The other players, actually the cloud providers themselves they work with us, they provide marketing space. We work with our sales people to help customers migrate off legacy platforms to the cloud and obviously running Atlas on their particular cloud. And then we also work with ISVs. A lot of these ISVs, some of them are legacy ISVs who are now realizing that rather deploying or offering an on-premise legacy solution, they have to be much more modern and offer a SaaS-based solution. So they are re-platforming their own product with MongoDB as the underlying data store. So, those are three examples of partnerships that we work with partners that we work with to help people migrate off legacy applications.

Got it. Thanks a lot for your perspective.

My pleasure.

This concludes our question-and-answer session. And I would like to turn the call back over to Dev Ittycheria for any closing remarks.

Well, I would like to thank everyone for joining us today. I'm really proud of the strong execution this quarter in a difficult macro environment. As we discussed, we're making continued investments in product innovation that further establishes MongoDB as the preeminent independent modern data platform, and our differentiated value prop is really best on best-in-class technology, developer mind share, and platform independence; and this is resonating with both, our existing as well as new customers, and we're looking forward to finishing the year. So, thank you for joining us, and we'll talk to you soon.

The conference has now concluded. Thank you for attending today's presentation. You may now disconnect.