All right, thank you to everyone for joining us. I'm Mike Cikos, the lead analyst here covering GitLab. With me, I'm pleased to say that we have the CFO, Brian Robins. Thank you for joining us.
Thank you for having us.
In the front, we also have Kelsey as well. We'll be going through some prepared questions on my side. Again, this tends to be the best when we are actually getting inbounds from you guys. Just make it interactive. You guys might have questions that are way smarter than mine, so please feel free to just launch them. I think just to level set, because I know we're a little bit removed, numbers are a little foggy, but coming off the October quarter, which feels like a lifetime ago, can you just give us a quick refresher on the numbers and how it played out?
Yeah, absolutely. Before I jump in, I just want to thank everyone for being here today and your interest in GitLab. I know everyone has busy schedules, and so it's great to see a lot of familiar faces, and so super happy with how we did in Q3. We had a great quarter, and we executed really well across every vector within the company. We reported 31% year-over-year revenue growth. We did that while maintaining best-in-class non-GAAP gross margins of 91%. We also continue to get great operating leverage in the business, and so one of the things that Sid and I have been really vocal about is number one objective within the company is to grow, but we'll do that responsibly, and so we added 1,000 basis points of operating margin improvement. Ultimate continues to do well. It's 48% of ARR, greater than 50% of total bookings.
Overall, it was just a really, really strong quarter. We talked about a number of new logos and AI. Happy how we're coming along with AI as well. It was the best quarter in Pub Sec in the company history, which was great to see that our products are really resonating in the Pub Sec market. Happy with the execution and happy with how the team did.
Excellent, excellent. And then one of the things that we've gotten off that call specifically, I think at some point Sid had mentioned something about consumption pricing. And I just wanted to see if we could flesh that out now while we have that.
Thank you for the question. This is probably the one question we got most after the call was, are you considering a change in pricing? And if so, when? And so I appreciate you bringing that up. And to clarify, there's no changing in the pricing model. That was specifically related to Duo Workflow, which is in conception right now. We said we get a private beta out by wintertime. And so when we have Duo Workflow and that comes out, there may be a consumption component to it, but there's no contemplated pricing changes right now with the model.
Excellent. OK. If I just think broader market at this time, we're in the new year. We'd love to get a sense, how are customer conversations shaking out from a budget or demand perspective? What are you guys hearing currently?
Yeah, absolutely. So because we do a number of different conferences, some are broadcasted, some aren't, we've made it a policy not to do sort of mid-quarter updates on customer demand. But I will do a refresher for those who did not hear the conference call. We said that the environment remains cautious. I think that's what a lot of companies are reporting, software companies specifically reporting. However, with that said, it's very stable. And so the last three quarters has been what we anticipated. And so it's cautious. Sales cycles aren't getting elongated, but it's been stable out there.
This last quarter, again, strongest quarter for the company in its history as it relates to Pub Sec. Has there been historically any change in demand from federal when we go through these changes in administration now that we know that we have a Trump administration? Or was Pub Sec maybe even too small to really be talking about four or even eight years back at this point?
When we talk about Pub Sec, it's roughly about 12% of the business, so it makes up a meaningful size of the business. There's no changes in the upcoming. There was no sales related to the upcoming administration that actually drove Pub Sec. It was just great execution by the team. There are some exciting things. There's a new DOGE Committee, I think it is, DOGE. And hopefully, with sort of the savings within the government, that's really what the GitLab platform does. It takes out all the point solutions, puts it in a single source of system or record, if you will, and the payback period is really quick. And so we hope that there will be many more opportunities within the Pub Sec market.
Got it. And I mean, the platform is obviously meaningfully expanded. You guys now have Dedicated. You have Duo Pro. You have Duo Enterprise. When we come back to the strength of Pub Sec in this most recent quarter again, is it fair to think that part of that is based on having something like Dedicated, I'm sorry, yeah, the Dedicated offering as well as the Duo offerings? Or is this just more consistent execution and maybe that's to come still?
Yeah, I would say it's more consistent execution than, as it relates to Dedicated. Some of our, if you look at our top 10 customers, some of our largest customers are three-letter agencies, and so it's really the benefit they're getting by having one platform and a single pane of glass that can do the entire software development lifecycle that the government's using.
Excellent. I guess if I think about where we are in the world, it seems to be like it's an easier way to cross-sell or upsell your existing customer base today versus going out and acquiring new logos. Can you talk about what GitLab is doing on the new logo front? Is there potential to actually accelerate it from where we sit today? If there is that potential, what would be a driver for that?
As a CFO, I always think there's room for acceleration. And so the answer to that would be yes. We have set a very simple commission structure for the sales team. And so when the sales team goes in and talks to customers, it's a very consultative sales approach. They're trying to understand what the customer needs are, what they're trying to solve, how they can help them. And so they aren't going in and saying, I want to sell you Ultimate. That's a SaaS product. And so we don't set our commissions by Ultimate versus Premium. We don't set them by net expansion versus first order. We allow the team to go in and basically sell whatever they can sell to the customer. It can be Ultimate versus Premium. It can be Duo Enterprise. It can be Duo Pro.
One of the things that I think is awesome about the GitLab model is once we land a customer, they tend to expand with us for many, many years and don't leave. We have a gross retention rate in the mid-90s%. If you look at cohorts from 10 years ago, they're expanding at roughly the same rate as cohorts from two years ago. I can't think of many other businesses with a single product or a couple of products where you can actually continue to expand 10 years with the same customer.
Can you talk about that dynamic for a second? I would think you see the value, you expand, and then maybe by the fourth or fifth go-round, you just roll it out to the, I guess, based on your comment, the customers are much more phased as far as how that rollout takes place then?
Yeah, so most of our customers buy from a bottoms up, and so it's a division department. They'll buy it. They'll use GitLab. Typically, we land with a developer, and then it will go to another division department, another developer, then another developer, then we'll change personas to operations, and then eventually, we'll change to security, and so when we land with a customer, we're landing more and more on Ultimate, but we typically would land on Premium, and then we would sell a number of seats, and then when we would have a tier upgrade, then they would have to buy Ultimate, so that was another way to get more wallet share, and so we see customers continue to expand. We have Agile Planning now, so people are adding new seats for Agile Planning. The AI SKUs are out there, so people are buying those.
And so there's just the ability to continue to expand with these customers because we land relatively small.
Excellent. OK, and on that point, as far as these new customers, are new logos landing larger today? I would think they are just because of how much traction you're seeing with Ultimate, but just to sound check it.
Absolutely. We are an enterprise company between enterprise and Pub Sec. And we classify enterprise as over 2,000 employees, which is sort of a higher bar than some companies. That comprises over 70% of our total business. But when you're really early and sort of new people adopting a platform, as you start to go up the curve, we're starting to see larger lands as we go into the customer base.
From that perspective, is that driven by, again, I know you're not necessarily incentivizing the sales work to go one way or another, but hey, larger organizations have deeper pockets, which might be stickier spend in this kind of environment? Or is it maybe a function of the maturity of the offering? What are the drivers for that move up market or landing larger today versus where we were a couple of years ago?
Yeah, I think us and our next closest competitor are really a small percentage of the overall market. It's a $40 billion TAM and growing, and so I think when you had a lot of first movers, you're buying for just really small segments of your company, and now people are buying for more division departments. We are a bottoms up sort of land and expand motion, but we do do some top-down sales. One company that we talked about is UBS. UBS, their CIO made a press release. You can Google UBS and GitLab years back and talked about how they're actually implementing. They implemented like 9,000 licenses in nine months, and then they're up to like 18,000 licenses in about 15 months, and so they've really taken a bunch of point solutions out, saved a lot of money, increased cycle times, and so forth.
They actually, several quarters ago, talked about it on their earnings call on how implementing this platform really saved them a lot of money and helped them with their software development.
So great on UBS and the expansion there. If you have seen that kind of success, is there a thought that maybe you can build out a stronger top-down selling motion to kind of supplement what you guys have? Or is that already in the works under the hood?
We do both, right? When we go in, obviously, reps have quotas, and they want to sell as much as they can as quickly as they can. But typically, it's a bottoms-up adoption. And when you do a wall-to-wall implementation, it's a much longer procurement process. And so when you're buying for a division department, if you look at the number of users that we say we have and the number of users that Microsoft says they have, it's like three-to-four X the number of knowledge workers in the world. And so everybody has a free GitLab account. Everybody has a free GitHub account. And so most people are familiar with GitLab already. And so when someone's buying it for a division department, they were GitLab users. And maybe they switch companies and are buying it for that division department.
Once we get in, it sort of spreads throughout the organization.
OK. And we spoke about the $40 billion TAM already, the fact that there's two platforms out there that are leading. It's you and GitHub. My question is, why haven't we seen consolidation faster in this market? There are so many point solutions out there. We're talking about these very quick paybacks and productivity gains. If someone says, hey, we're good, we'll take a pass on GitLab, what's the pushback on that?
Let me answer the question around consolidation.
Yeah, sure.
And then I'll talk about some of the reasons why people don't adopt. And so from a consolidation play within sort of the vertical, we've done some M&A, but all of our revenue has been organic. And so we haven't acquired revenue. We have a single code base for our SaaS product and the same code base for SaaS as it is for self-managed. And if we actually go out and acquire a big company, we actually deprecate the code and rewrite it in our platform. And so it wouldn't make sense to go out and pay a lot of money to acquire all these customers and say, hey, we're deprecating the code. We're no longer giving you that product, but you get the full platform. And so that's why there hasn't been a lot of company consolidation. We've actually built most of it.
We will look at a buy versus build to get to market quicker, but we've built most of the platform, most of the feature functionality that we had. When you're selling the product, if you're selling the product, our biggest competitor is not Microsoft. Our biggest competitor is DIY DevOps, and so it's the divisions of people that are actually with a bubblegum and baling wire and putting together all those solutions with all the interfaces and so forth, and so when you're going in and you're actually selling that product, you're trying to sell it to the people whose job is to do all those integrations and manage that. If you have a platform, you don't need those departments to do that, and so a lot of people, we don't claim our secret sauce is the fact that we have the entire software development lifecycle in a single platform.
It's a system of record. You can shift security left, but we don't claim to be best of breed in every single phase of the software development lifecycle, but it's good enough to allow people to actually be able to work on everything at the same time to create software better, faster, cheaper, more secure, and that really is the secret ingredient, if you will, to the platform.
Excellent. And I know that you guys put a big focus on security compliance, putting that checkpointing in place throughout the process, right? I think on the most recent quarter, you guys were talking about the launch of Advanced SaaS, that is now GA as well. And that was always the big pull for Ultimate. It feels like you guys are continuing to use packaging as a lever to help that upmarket adoption, right? We talk about Dedicated, you have to be an Ultimate customer. Duo, you have to be an Ultimate customer. Is that the philosophy we should think about as driving? It's carrot, not stick necessarily?
You're getting into sort of a pricing and packaging question. And so when we price our products, we really look at three things. We look at what the benefit we're bringing to our users, what the cost to deliver it is, and then what's the competition out in the marketplace. And a lot of the things that you talked about, Duo Enterprise and Duo Pro are per seat. And if you go into enterprise, you have to be on Ultimate because of all the advanced security and compliance and so forth. Dedicated, you have to be on Ultimate as well because it takes a certain amount of money to set up a single tenant SaaS, a private network. And so really, it's the number of licenses. It's the amount that they're spending with you.
It is helping drive Dedicated, but it's hard to put that just on Premium and say, hey, you can have Premium in all these things just because the Premium price point is much lower than Dedicated. I'm sorry, than Ultimate.
And with Premium, obviously lower price tag, customers more economically sensitive. What is it you guys are seeing on that front from a new logo perspective versus internal plan experimentation, again, just given the economic sensitivity of those customers?
Yeah, so since we don't have a commission for Premium versus Ultimate, we actually tell the reps to go out and sell what you can sell. When we report the numbers of the quarter, most of this is outputs, not something that we're actively managing to. And so we actually have seen Ultimate do really well since we managed to a total bookings number within a quarter or as big as we can. If you sell more Ultimate, by default, you're going to sell less Premium. And so Premium has been doing well within the enterprises. Where we've seen some weakness in Premium is really related to the price increase. And so we've seen some price sensitivity in the lower end of the market, lower in the mid-market in SMB. And so we've been running some promotions around that to see how price sensitive they are.
And one of the things that we did is if you had fewer than a certain number of employees, you could buy a certain amount of Premium licenses. And we saw trial signups increase by 20%. And we saw conversions increase by 20% when we did that. And so we're still playing with that. We may actually come out with something specifically for SMB in the low end and mid-market, or we may continue to run some promotions.
Excellent. And for Duo Pro and Duo Enterprise, how has the go-to-market playbook been built for those solutions? Does it require a little bit more of a sales specialist or a sales overlay? Is a sales engineer getting pulled in to help architect that? How is that introduced to the customer in the sales process?
Yeah, it's the same rep, and so there's not multiple reps. There's not like an overlay force that's going out and selling Duo, and so we've actually worked on the enablement materials. We've trained the field, and so the same rep is using Duo as a way to get into new accounts and also to go revisit the existing accounts, and so it's another tool in the tool bag, if you will, to say, "Hey, Mike, we have Duo Enterprise now. Would you like to try it and see what impact? Here's some of the customer testimonials that we have," and so it's the same rep. One of the things that we mentioned, I believe it was like two calls ago, is we've had to shift some of the go-to-market materials over the last couple of years.
It went from being a very technical sale to a payback financial sale. Now it's a combination of both with AI in it. We've changed a little how we've done that. Now on the technical side, for some of these larger enterprises, we have what we call Field CTOs. This is sort of like a very senior SAE with a Field CTO title. When they go into our enterprise customers, they can help talk about migration, benefit, the technical solution, and so forth. That's really helped out.
Excellent. And for Duo, you guys have spoken about some big customer wins there, Emirates, F5. Can you talk about, I guess, the adoption trends when that is discussed with the customer? Is that typically at renewal? Or is that, again, that's another opportunity for me to go out to that customer and engage them even before they're up for renewal?
Yeah, any time our reps are going out and talking to, it was five quarters ago. I think I got my quarters right. About five quarters ago, as we prep for earnings, I'll go to our CRO and CMO, and I'll say, what new trends are you seeing out in the marketplace? No one mentioned anything about generative AI. Four quarters ago, it was almost in every single sales conversation. And so if it doesn't come up, our reps are actually bringing it up proactively and getting people to try it. And so if you look at over the last four quarters, it's been super interesting. So this was the first time where I think media really led the hype cycle. Investors were trying to catch up, and companies weren't even allowing it within their enterprises because there's IP concerns. There's security concerns. There's a number of different concerns.
Some of those have been solved, and so then like two quarters ago, you would see people buy it, but they would buy just a small set of licenses, 10, 15, 20 licenses. They would test it in sort of sandbox environments. They would see what the impact was. But it was still not allowed from policy perspective in a lot of companies. One of the things I talked about on the earnings call was when I looked at Duo, and this was sort of a point of clarification I wanted to bring up as well. When we looked at Duo, the deals that we sold that had Duo involved in them, Duo was greater than 25% of the total ARR for those deals.
What we're seeing now is people are buying them to match the number of licenses that they have, but they're actually rolling them out sort of in slower batches, if you will. It's good to see that we're seeing more adoption. We talked about that it's exceeded our internal expectations. The logos that you mentioned are great logos. We're starting to see some bigger lands with bigger enterprise customers that are trying out Duo Enterprise. I also think it's really important to mention Duo Enterprise is very different than Duo Pro and also Copilot. Duo Enterprise is basically injecting AI throughout the entire software development lifecycle. A lot of what you hear about there with generative AI is just meant for coding, and it's code assistance. That's really what Copilot does.
We put out a DevSecOps report, and coding only makes up about 20% of the total process: actually plan, manage, deploy, write, test, secure. So it's a very small part of the whole process.
When Duo was introduced, so great on the 25% uplift, right? But when that's introduced, do you tend to see sales cycles elongate just because there are potentially more constituents, or now it's a larger price tag than it would have been otherwise? How does that play out?
Great question, and so if you go back four quarters ago, I think it was like three or four quarters ago, we talked about some elongation of the sales cycle. When I went back within Salesforce and looked at the phases, like phase one, two, three, we have the six phases in Salesforce, is really the evaluation phase expanded, and the later phases actually stayed almost identical, and so what happened was where people didn't have generative AI, they were actually going out and testing all the different products, seeing what impact it would make, and so their evaluation period got longer, but over the last three quarters, the sales cycles have remained relatively constant.
Excellent. And for Duo Enterprise specifically, you guys also just recently announced that now customers can self-host their models. Again, probably feeds very well into the public sector. Are commercial enterprises asking you for similar functionality as well? Or again, should I think about that as being much more of a public sector play?
No, it's for all of our customers. We have really happy on how SaaS revenue has grown. It's grown 44% year -over -year. I think it comprises about 29% of our total revenue today. We don't mind how people consume the product. Once again, going back to our sales compensation philosophy, we want to make it easy to sell and easy to acquire. So if we're in a sales process and we've sold them on Premium or sold them on Ultimate, and then we say, how do you want to consume it? Do you want us to host it? Do you want to host it? Do you want to put it Dedicated? And if we tell them that there's different prices for self-managed versus SaaS, they'll go back and say, well, how long is it going to take me to set it up? What's the environment going to cost?
We're like, same price. And so we're sort of agnostic on how people consume it.
And for, again, the core, let's say, Premium or ultimate offering, we understand, all right, if I'm not adopting, it's because Sid Sijbrandij has built out this tool chain. There's a lot of sweat and blood that went into that. And now we're just going to deprecate it because we're going with GitLab. What's the pushback you receive if someone decides, hey, we don't want to go with Duo? Is it really the privacy concerns or maybe compliance still needs to get over that hurdle?
We're a long-tail business, so we don't have any major customers that have big revenue concentrations. And we're across every industry vertical. And so there's a number of different use cases. But the primary reason why people don't adopt Duo now is around policy and security. And it's really trying to understand what impact it's going to have to their proprietary code in their platform. We don't allow other companies to train off of your code. And so there is some security benefits with going with our Duo product from an IP perspective as well. But there's a number of reasons why people don't adopt. I think right now it's really the newness of the product. It's trying to actually measure what the impact's going to be in a lot of internal policy.
Right. And I don't mean to mitigate what you guys have achieved either. Again, it's just interesting to see the spectrum because you do talk about new logos who are landing and buying GitLab with Duo, right? So not to take away from it, I'm just trying to figure out what that pushback would be, you're getting for the market. I think it'd be great to touch on Enterprise Agile Planning as well. I know a lot's been made as far as potential market share gains coming from competitors out there. How is that tracking?
So for those of you who aren't familiar with Enterprise Agile Planning, let me explain what it is. So you get all the benefits of planning when you buy the ultimate SKU. And so we were getting a lot of inbound requests for the non-developers to have a planning type functionality like Jira, but they didn't want to pay the $99 per month because they were non-developers. And so what we were doing was we were blending sort of the rates to be able to give them licenses, and they weren't using the rest of the functionality. And so we came up with the ability to have a planning-only functionality. And so you have to be on Ultimate. Once again, you have to have a certain amount of licenses. And then if you want, you can actually buy planning-only SKUs.
What gets me most excited about that is less from a financial standpoint because the number of seats that you're buying just for planning-only is relatively low. What I love about it is it's good enough for people to actually use that versus some of the competitive products. And what that means is it's a stickier platform. Total cost of ownership is actually improved. Paybacks improved. Time to value is improved. And so from a technology perspective and stickiness perspective, that's really why I like Agile Planning.
And also, correct me if I'm wrong, but I think that comes in as well as far as having that unified data store. I know you guys maybe don't say it as boldly, but I think of AI like this is kind of your spot to win right now, just given the unified data store. Can you take a second to just carve out and talk about that data store? What are the advantages that you guys have versus, again, some cobbled-together tool chain?
Yeah, I mean, if all your data is in one place and everybody's working sort of around that data store, it just makes the process much easier. One of the things I like to talk about is I'm brought in to take companies public. A number of those companies have been sold, so I've drafted a number of S-1s. And trying to draft S-1s in Microsoft Word is an absolute nightmare because you're playing air traffic controller the entire time. Cut, paste, send files out. Who has the right version and so forth? And so the system of records is really difficult. The last couple I've done, I've done in Google, and I've done it in Google Docs. And everybody can work on the document at the same time. You can put suggestions in there. You can track all the changes.
And so we are the Google Docs for software development. So you can see all that. And so when you have one data store, once you make a change, it goes through the entire set of data. You don't have to actually say, go update this over here, go update this over here, pull stuff together. And it just makes it more of a seamless process to do code development.
Right. And the greater insights generated just because you've removed those silos, right?
Correct. Exactly. No comment. No comment.
I think what would be interesting to the extent you can comment, but with your Duo customers or customers who are coming over and using you guys for AI Code Assist, do you have a segment as far as how many of those are net new to AI Code Assist versus maybe coming from a competing solution or continuing to use a competing solution and just tacking on GitLab as well? Is there a way to frame that out?
Yeah. So I don't have the data, so I can't answer it specifically. But I would say it is so new that there's not a lot of competitive wins. I can't think of one instance where we actually sold Duo because it's so new, and then they canceled it immediately and then went to Copilot and vice versa. I sit in all the forecast calls, and I can't think of one instance where they said, this is a Copilot ripout. They may be trying Copilot, and they may be trying Duo, but it's really net new across the board.
Got it. And I know I still have questions on my side, but I want to make sure I'm being true to my word. Does anyone have questions out there that they want to lob in? Otherwise, happy to keep going. All right. Let's go. On the competitive front, when you guys are coming in and let's say displacing one of those tool chains, are there more common competitors which are taken out first? Let's say there's some solution that just hasn't been invested in, or the solution's being deprecated by the vendor. Can you comment again from a competitive standpoint, not versus GitHub, but versus, again, that point solution effort?
Yeah. So in each sort of vertical of the software development lifecycle, there's a number of people who participate in each of those. Since we compete against every industry or we have customers in every industry, there's not one or two that we're taking out all the time. And so it just really varies by company.
No comment. For Dedicated and duo as well, again, just trying to think about potential benefits of the longer-term model. But newer offerings, are you starting to see enough data that would tell you that maybe these customers have better lifetime value economics or not necessarily? Again, just based on the nascency of the offering, tough to get there.
Yeah. Measurement around, you hear a lot of numbers up to 50% productivity improvement when you use AI. Customers have told us that as well. You really have to get into, is that around the coding part, or is that around the entire part? Where are you actually seeing it? I think GitLab's whole premise is to make the software development lifecycle more efficient, greater, quicker time to value, greater business outcome, and so forth. Pre-AI, we did a total impact study with Forrester, and the payback period, if you bought Ultimate, was less than six months. And the ROI was over 480% in three years. And so AI is another tool that our customers can use to become even more efficient.
So, I do believe when we get in pricing discussions, I'm like, if it really, really works well and you're really, really helping out the customer, you can charge a lot more. If it's not working well and you aren't helping out the customer, you probably can't charge a lot. And so, I think as we talk about these additional tools to make them more efficient, it's just going to help them the total cost of ownership, the benefit create more stickiness than what we already create today.
There is a demand from partners to work with you guys as well. We recently got news about the Amazon partnership. Could you discuss that as far as how the integration is supposed to work and what increased functionality or efficiencies you anticipate bringing to customers as a result?
Yeah. I mean, it's super exciting. Amazon, big company. They can work with anybody they want to work with. And so center stage, at re:Invent, they talked about a partnership with GitLab. And so we're going to take our Duo with their Amazon Q, put the best of both together. Then Amazon's going to invest in a Dedicated Salesforce, go out and sell that. Then we're also going to sell it to our customers as well. And so I view it as a real testament to GitLab and our thought leadership around the DevSecOps space that a company like Amazon would partner with us to take a product jointly to market.
Just to be super crisp on that, their Dedicated Salesforce for this, they'll be able to retire quota by selling GitLab then?
Correct.
Okay. Excellent. Excellent. Profitable growth you had discussed earlier. Obviously, margins have meaningfully increased in the last year plus. How do you guys think about growing this business? Is there a baseline yield you try to generate off each dollar? Again, how do you think about this margin expansion story for GitLab right now?
Yeah. Yeah. If you look at my operating history, when I was CFO at Verisign, we increased margins greatly. I've been brought into these companies to really help with the go-to-market focus. And I've been at GitLab now almost four and a half years. And every quarter -over the last four and a half years, sales and marketing as a percentage of revenue has improved every quarter. And so it's really helping the team understand that we can do more for less. And the number one thing we wanted to do was invest in growth. But if we don't see it, we're going to calibrate the expenses. And so we've been dropping that to the bottom line and happy with where we've gone. I don't have a formula per se.
When we do our internal planning, say, for every dollar of booking, this is what I want the incremental as a drop. We go through and look at every cost component within the business and say, how can you become more efficient? And so as we were getting ready to go public, Sid and I were on the road during the roadshow. And we said that our SaaS at that point was just a new offering. And we had 90%, I think, Non-GAAP gross margins in the roadshow. And we said, as SaaS continues to get bigger and bigger, you expect gross margins to drop down to 85% because of the cost to deliver it. But we really challenged our delivery team on how to actually maintain that margin. They've actually increased it with SaaS now 29% of total revenue and growing 44% year -over -year.
That's just an example. At 91% non-GAAP gross margins, we continue to push them and say, how can you become more efficient? What tools can you make for delivery? How do we negotiate our bandwidth costs? We're doing it in sales and marketing, doing R&D, and we're doing it in G&A.
Right. Right. And it's one of the things we're highlighting for folks too. I think if you go back the last two years, the incremental margins on the operating profit line have been somewhere in the neighborhood of like 50% plus. And even this year, I think it was Q1, you guys had a big live event that was somewhere around $10 million-$15 million as far as expense. But year to date, you're still well north of 40% on that incremental margin as well.
Yeah. So pre-COVID, GitLab was the largest remote company in the world. And so GitLab was started that way over 10 years ago and never had an office. So on our S-1, when it said physical address, we put N/A because we've never actually, I think we have a P.O. box where some stuff comes into, but that's really our only physical address. And it's really helped the company from hiring the best and brightest from around the world. Cost of labor, if you can hire in low-cost areas, is better. And so it's really helped us. And so we've been fortunate. So pre-COVID, the company would get together once a year so people could just have that connectivity. In person, the company was relatively small. And we actually had canceled it two or three times.
This year, in first quarter, we got over 2,000 team members together for a week. That was about a $15 million hit to the P&L.
Is that again? We're post-COVID now. Is that expected to be an annual thing then? Should we just?
No, we won't do an annual. And we're trying to. It was really beneficial to the team members. And so we decided not to do it next year. So we won't do it next year. We're trying to. What is the most impactful way to get people together? You probably remember when you're doing these conferences on a Zoom call, and there's much different, right? Being in person, being in a room, seeing body language, seeing people is much better. And so we may do regional events. We may combine it with SKO. There may be a couple other things that we may do, but we haven't made a final decision on that.
All right. I'll pulse check again. Any questions out there? All right. We'll close it out then. Just two more questions on my side. The first, management CEO. Obviously, we have Bill Staples, who's now appointed CEO. Can you just discuss his background, credentials, and what his expertise are as far as continuing to scale this business?
Yeah, absolutely. Let me first touch on Sid, and so Sid was the founder of the company and the CEO of the company for a very long period of time. About a year ago, Sid found out they had cancer, and that came back about two quarters ago, so the board stepped up its plan to look for succession planning to give Sid more time to focus on his health. And so Sid has moved to an Executive Chairman role. He's still an employee of the company. He's available really anytime to help out. As we were preparing for earnings, you could imagine how sad it was that Sid wasn't going to be the CEO. Sid looked at us and said, like, this isn't a goodbye. This is just the next chapter, and he was super energetic and excited about it.
So the board went out and was doing a search over several periods of time. And they met Bill. And Bill just really fit perfect. He has great experience at some big companies, Microsoft being one of them, Adobe being another, running billion-dollar-plus P&Ls. And then most recently, he's CEO of New Relic. His background is from being a product guy as well as a developer. So he really understood the space well. And when you're interviewing for that role, it's not about competency. It's about culture and fit. And Bill really fit well with the management team, with Sid, with the board. And so he's only been here five weeks. And it's been great to have him on board.
Excellent. Okay, and with that, we'll leave it there.