Q2 2023 CS Disco Inc Earnings Call

Participants

Kiwi Camara; Co-Founder, CEO & Director; CS Disco, Inc.

Lee Robinson; VP of IR; CS Disco, Inc.

Michael S. Lafair; Executive VP & CFO; CS Disco, Inc.

Brent John Thill; Equity Analyst; Jefferies LLC, Research Division

David E. Hynes; Analyst; Canaccord Genuity Corp., Research Division

Jackson Edmund Ader; MD of Technology Equity Research; SVB Securities

James Derrick Wood; MD of TMT - Software & Senior Software Analyst; TD Cowen, Research Division

Jeffrey Parker Lane; Associate; Stifel, Nicolaus & Company, Incorporated, Research Division

Koji Ikeda; VP; BofA Securities, Research Division

Mark William Schappel; MD; Loop Capital Markets LLC, Research Division

Scott Randolph Berg; Senior Analyst; Needham & Company, LLC, Research Division

Tyler Maverick Radke; Research Analyst; Citigroup Inc. Exchange Research

Presentation

Operator

Ladies and gentlemen, thank you for standing by, and welcome to the CS DISCO Second Quarter of Fiscal Year 2023 Conference Call. (Operator Instructions) I would like to now hand the conference over to your first speaker today, Head of Investor Relations, Aleksey Lakchakov. Please go ahead.

Lee Robinson

Good afternoon, and thank you for joining us on today's conference call to discuss the financial results for fiscal second quarter of 2023. With me on today's call are Kiwi Camara, DISCO's Co-Founder and Chief Executive Officer, and Michael Lafair, DISCO's Chief Financial Officer.

Today's call will include forward-looking statements within the meaning of the safe harbor provisions of the Private Securities Litigation Reform Act of 1995, including, but not limited to, statements regarding our financial outlook and future performance, our future capital expenditures, market opportunities, market position, product strategy and growth opportunities, and developments in the legal technology industry. In addition to our prepared remarks, our earnings press release, SEC filings and a replay of today's call can be found on our Investor Relations website at ir.csdisco.com.

Forward-looking statements involve known and unknown risks and uncertainties that may cause our actual results, performance, or achievements to be materially different from those expressed or implied by the forward-looking statements. Forward-looking statements represent our management's beliefs and assumptions only as of the date made. Information on factors that could affect the company's financial results is included in its filings with the SEC from time to time, including the section entitled Risk Factors in the company's quarterly report on Form 10-Q for the quarter ended March 31, 2023, filed with the SEC on May 10, 2023, and the company's upcoming Form 10-Q for the quarter ended June 30, 2023.

In addition, during today's call, we will discuss non-GAAP financial measures. These non-GAAP financial measures are in addition to and not a substitute for or superior to measures of financial performance prepared in accordance with GAAP. Reconciliations between GAAP and non-GAAP financial measures and the discussion of the limitations in using non-GAAP measures versus their closest GAAP equivalent is available in our earnings release. And with that, I'd like to turn the call over to Kiwi.

Kiwi Camara

Thanks, Aleksey. Good afternoon, everyone, and welcome to our earnings call for the second quarter of fiscal year 2023. Revenue for Q2 2023 was $34.3 million. Adjusted EBITDA was negative $7.4 million, and we ended the quarter with 1,431 customers, up 14% year-over-year. Both revenue and adjusted EBITDA were above the high end of our guidance range with adjusted EBITDA improving by over $5 million year-over-year and adjusted EBITDA margin improving by more than 1,700 basis points quarter-over-quarter. We are pleased with our results and our progress towards profitability.

On prior earnings calls, we discussed the various sales and marketing performance drivers that are improving. We discussed seeing continued customer count growth and improvement in pipeline metrics such as meetings, near-term opportunities, and wins. That improvement continued in Q2. Compared to Q2 of last year, total meetings were up 30% and near-term opportunities were up 20%. While at the same time, we have continued to enjoy the high win rates that we have experienced since our IPO.

Inbound leads have more than doubled year-over-year, in part as a result of our recent investments in marketing. While we have continued to see some reductions in usage due to cost optimization, especially from our largest customers, our accelerating sales activity is allowing us to offset these reductions by bringing new customers and matters onto our platforms. These positive signs tell us set our efforts to increase usage and accelerate product adoption are working.

Last year and early this year, we promoted several classes of sales development representatives, or SDRs, to quota carrying account executive roles on our inside sales teams. These internally promoted reps have seen strong early success with all reps producing revenue within their first 90 days in role and some reps already producing more than $1 million in annualized revenue. We believe that the experience SDR's gain through their time in our SDR program, experience that familiarizes them with the DISCO products suite, our customers, the legal industry and legal workflows, and the way successful sales leaders at DISCO sell, helps them achieve this kind of rapid success as quota carrying reps. We believe the pipeline of talent from SDR to inside sales, and eventually to other functions such as customer success and field sales, will be an effective and efficient way of growing our sales team going forward. Already, many of our top performers in these roles are alumni of our SDR program.

In our Review business, we have seen a rebound in activity this quarter, both in the scale of our total pipeline and the size of the deals in our pipeline. One of our key operational initiatives has been increasing the percentage of our sales team that sells Review successfully. In the first half of this year, more than 60% of our quota carrying reps generated Review revenue, which is almost double the percentage who did so in the first half of the prior year. We have also seen some increases in larger reviews with one customer that had multiple reviews in the quarter, generating more than $1 million in Review revenue in the quarter. This quarter, I would like to highlight a large international law firm that has been a DISCO Ediscovery user for more than 5 years. This customer has over 120 matters running simultaneously on DISCO and generated over $3.5 million in total revenue over the last 12 months. More recently, they have become recurring users of Review with 3 reviews active in Q2 and several more in the pipeline. Overall, this customer finds that our platform helps them complete legal document review more quickly and efficiently across a wide range of legal matters around the world.

Another customer, a global publicly traded shipping company, and their Am Law 100 law firm has been using DISCO for a large global antitrust matter. They began using DISCO in Q4 of 2022 and have since increased usage to more than $200,000 this quarter. This customer selected DISCO after a rigorous review of our capabilities against our peers. DISCO stood out as the company with the most robust technology for handling large datasets and complex review workflows quickly. It has been an exceptional partnership for us and a perfect example of DISCO's global reach and impact.

Now let me turn to R&D. Last earnings call, we discussed DISCO AI and how our AI capabilities are integrated into our platform and products. We discussed key features powered by AI, such as topic clustering with automatic indexing, predicted PAG and cross-matter AI. We also introduced Cecelia, our integrated AI chatbot for large-scale e-discovery. We continue to believe that our decade-long investment in our AI lab and in AI feature engineering, the private data that we have attracted to our platform, the fact that our platform is already the system of engagement for a variety of legal workflows that precede and follow the use of AI, and the trust we have earned in the market as a company that productizes cutting-edge technology for legal use cases in a reliable, user-friendly and secure way, all position us well to lead the legal industry's adoption of AI.

While investment in AI is important, it is also important that we continue to deepen our core non-AI product capabilities. We released 3 key features in this category this quarter, all of them long demanded by our existing customer base. First is Dynamic Threading. In a typical e-mail chain, later messages contain copies of earlier messages. When reviewing such an e-mail chain, it is useful to identify the e-mails that contain unique content and suppress the e-mails that are included in other e-mails so that lawyers do not waste time reviewing the same content multiple times.

For a long time, DISCO has had the capability to limit reviews to e-mails with unique content, often called inclusive e-mails. But DISCO has identified inclusive e-mails on a global basis, considering all e-mails in a given review database. In some reviews, customers prefer to identify e-mails with unique content from a subset of documents. For example, the documents that fall within a certain date range. Dynamic Threading allows customers to identify e-mails with unique content in arbitrary subsets of the documents and to do so dynamically. That is in real time as users set up review statements in DISCO Ediscovery. With Dynamic Threading, customers can for example define a review universe using complex searches by keyword, date, or AI predictive tags, then reduce that review universe by identifying e-mails that contain unique content and suppressing other e-mails. This allows customers to further limit review population and thereby accelerate the process of legal document review.

Another exciting product release this quarter is In-App Translation. This functionality allows customers to translate documents between over 60 different languages right in the DISCO platform. Today, lawyers tend to use either expensive third-party translation services or online translation features intended for consumer use cases that, while free and convenient, may leak private data to translation providers. With In-App Translation, lawyers can translate documents with the click of a button and do so securely with none of their data leaving the DISCO platform.

We also released the Cloud Connector for Office 365 e-mail. This allows customers to ingest e-mails directly from Microsoft Outlook with no intermediate collection or download steps. Today, lawyers frequently have to export data and then repeat the process for any subsequent collections. With the Outlook Cloud Connector, customers can ingest Ediscovery data themselves with the data flowing directly from Microsoft to DISCO. As more corporate data moves to the cloud, we believe the direct connectors like this will become a more and more common way of securely collecting data for legal matters. These connectors build on our Hold capabilities, which already allow for preservation in place in a variety of cloud systems including Office 365.

Now let me share our progress on Cecelia. We are on track to have Cecilia generally available to our customers by the end of this year. In Q2, we began piloting Cecelia with a select customer group. The feedback we have been receiving has been overwhelmingly positive. Cecelia is able to rapidly answer specific questions based on data contained in customers' private databases while citing specific documents and document experts to support her answers. In one instance, Cecelia found an important document that the legal team had not found in their original review using traditional review methods. An associate on a separate legal team mentioned that Cecelia would have saved their client thousands of dollars and saved the associate many hours of document review. This is the kind of impact we are aiming to have.

We believe Cecelia will be a big improvement for our customers and look forward to continuing to expand our Cecilia user base in the coming months.

We have noticed that our approach to AI is different from that of many of our competitors in that we are LLM agnostic. We believe that as the years pass and technology matures, there will be many winning LLMs with different strengths, weaknesses and optimal use cases. We have engineered our AI platform so that we can use different models for different tasks, including both internal models and third-party models. We believe this flexibility, rather than going all in on a single LLM or LLM provider, will give us an advantage in bringing the best product capabilities to market and doing so efficiently. We look forward to announcing our next set of AI capabilities in the coming quarters. We are happy with the progress we made this quarter and all of our team's accomplishments. With that, I will turn it over to Michael.

Michael S. Lafair

Thank you, Kiwi. In Q2 2023, revenue was $34.3 million. Our year-over-year growth in the quarter is attributable to a rebound in our Review usage. In Q2, we saw the highest quarterly Review revenue since Q1 of 2022. We are seeing improvement in our overall go-to-market activity, and we are optimistic about the second half of the year.

In discussing the remainder of the income statement, please note that unless otherwise specified, all references to our gross margin, operating expenses, and net loss are on a non-GAAP basis. Adjusted EBITDA is also a non-GAAP financial measure. Our gross margin in Q2 was 74%. As we mentioned before, our gross margins fluctuate from period to period based on the nature of our customers' usage. For example, the amount and types of data invested and managed on our platform. We expect gross margin to continue to be within the band we've historically seen.

Sales and marketing expense for Q2 was $16.2 million or 47% of revenue compared to 52% of revenue in Q2 of the prior year. This represents a decrease of $1.3 million in the quarter year-on-year. The decrease was primarily driven by a decrease in sales and marketing personnel. Research and development expense for Q2 was $10.5 million or 31% of revenue compared to 39% of revenue in Q2 of the prior year. This represents a decrease of approximately $2.6 million in the quarter year-on-year. This decrease was primarily driven by an increase in capitalized development costs associated with our AI investment efforts and a reduction in research and development personnel.

General and administrative expense in Q2 was $7.3 million or 21% of revenue compared to 24% of revenue in Q2 of the prior year. This represents a decrease of over $0.7 million in the quarter year-on-year. This decrease was primarily driven by lower costs related to reduced professional services fees through the renegotiation or termination of vendor contracts.

Operating loss in Q2 was negative $8.5 million, representing an operating margin of negative 25% compared to negative 39% in Q2 of the prior year. In total, our Q2 operating expenses were over $4.6 million lower than Q2 of the prior year, representing an approximate 12% reduction in operating expenses.

Adjusted EBITDA was negative $7.4 million in Q2 and adjusted EBITDA margin of negative 22% compared to an adjusted EBITDA margin of negative 37% in Q2 of the prior year. Net loss in Q2 was $6.5 million or negative 19% of revenue compared to a net loss of $13.5 million or negative 40% of revenue in Q2 of the prior year. Net loss per share for Q2 was $0.11 per share compared to $0.23 per share in Q2 of the prior year.

Turning to the balance sheet and cash flow statement, we ended Q2 with $178.9 million in cash and tax equivalents. Operating cash flow in the first half of 2023 was negative $21.8 million compared to negative $22.2 million in the same period of the prior year.

Now turning to the outlook, for Q3 2023, we are providing revenue guidance in the range of $33 million to $35 million and adjusted EBITDA guidance in the range of negative $8 million to negative $6 million. For fiscal year 2023, we are reiterating our prior revenue guidance of $135 million to $145 million and raising our adjusted EBITDA guidance to a range of negative $34 million to negative $30 million.

Now I would like to turn the call over to the operator to open up the line for Q&A. Operator?

Question and Answer Session

Operator

(Operator Instructions) Your first question comes from the line of Tyler Radke with Citi.

Tyler Maverick Radke

Good evening, Michael and Kiwi. Kiwi, I wanted to start off with your comments just on the pipeline generation. It sounds like you're seeing an uptick in leads. And I'm wondering if you could just kind of comment on the nature of those and how you're expecting that to progress over the coming quarters? Just how you see those, sales cycle length, and then any thoughts on conversion and could that potentially lead to more of an accelerating revenue growth profile as we head into 2024?

Kiwi Camara

We're really pleased by the performance of the pipeline. And I've talked about it on the past few earnings calls, looking at how well our new team members are ramping across each step of the funnel, from SDR and other kinds of lead generation through to our inside sales and field sales teams closing deals, and then our CSM team working to identify opportunities to expand existing customers. Now the numbers we shared today, on a year-over-year basis the meetings were up 30%, and near-term opportunities were up 20%. And we're continuing to enjoy, despite that broader pipeline, the same elevated win rates that we've been enjoying since the time of the IPO. Those deals are flowing through all the way to the end. I think you can also see this in the continued growth that we've posted in overall customer count, which is now up to 1,431. These are some of the things that give us belief that there are green shoots in terms of growth going forward.

Tyler Maverick Radke

Yes, thank you. And maybe a follow-up for Michael on the green shoots. If I take kind of the implied Q4 guidance with the updated guide on Q3 and the full year guidance, it does imply a pretty strong sequential increase into Q4, both on a quarter-over-quarter and year-over-year basis. Could you just talk about what's giving you the confidence in that Q4? Whether it's maybe some large Review cases or maybe it's a reflection of some of the pipeline commentary Kiwi just hinted at? Thank you.

Michael S. Lafair

Tyler, thanks, it's a good question. It's really related to the pipeline. And also, just -- we're providing full year guidance in the range of $135 million to $145 million and also Q3 guidance. We feel really good about where the business is and we feel really optimistic about the second half.

Operator

Your next question comes from the line of Jackson Ader with MoffettNathanson.

Jackson Edmund Ader

The customer count I think is growing faster maybe than we expected, but average revenue per customer, the other side of that coin, is a little bit weaker. I'm curious, is this because new customers that are coming on there today versus a few years ago are smaller? Or is it due to some of the larger customers may be putting pricing pressure on you? I'm curious what's driving that.

Kiwi Camara

Yes, it's 2 dynamics. First, starting with very big customers, that's where we have experienced most of the headwinds around customers seeking to reduce their overall spend. And as a consequence, reduce usage or change the nature of their usage on our platform to try to get that spend down. Now on the positive side, we've continued to add lots of new customers, but typically a DISCO customer starts relatively small and then grows over time. In the past, we've talked about kind of 3 to 5-year ramps to maturity for new logos. Sometimes it goes faster, especially when the new logo is somebody who has already adopted DISCO at a previous company. But it's still very much a land-and-expand business.

Jackson Edmund Ader

And I guess just following up on that, the let's say the type of customer that you're landing or adding as a net new customer today, do they still have the same type of 3 to 5-year opportunity as to customers that you were landing say 2, 3, 5 years ago?

Kiwi Camara

Yes. If anything, that has actually improved in our business. Over the last 3 to 5 years, we've moved steadily upmarket. We gave some examples in the customer stories part of our script. Some customers, including one very new customer who in less than a year got to material size. Many of our newer logos are large multinationals or large law firms that provide big expansion opportunity. We of course continue to sign up customers all along the spectrum. But if anything, I would say the average wallet potential of the new customer has gone up over 5 years, not down.

Operator

Your next question comes from the line of Derrick Wood with TD Cowen.

James Derrick Wood

Michael, I know you said gross margins bounced around a bit, but this was one of the first times we've had that kind of contracting quarter. And I'm just curious, does that reflect some of the investments on the generative AI and large language models that are starting that obviously have some added cost to it? Or are there other factors to call out? I guess now that you've had a little more time building Cecilia, any updated views on the potential impact on gross margins once a lot of this goes into production?

Michael S. Lafair

The margin profile that we've talked about in the past in terms of the bands that we've historically seen over the last 4 to 8 quarters is where we expect the future to be. The investments that we're making in AI are not having an impact on that band, and we don't expect that they will in the future.

James Derrick Wood

Okay. Kiwi, I'm curious your thoughts on vector search versus keyword search. And as you think about trying to get more contextual answers with natural language queries, are you thinking about it? Kind of embracing more vector search capabilities? Or what's your view on that technology?

Kiwi Camara

Absolutely. We're big believers in vector search technology. And this is one of those things that has suddenly become in vogue but is not at all new, right? This is something that we've worked on for many, many years, that's been around for many, many years, and that is a key component of some of the feature functionality that we're delivering to clients. We're big believers in vector search. I will say though, some people think that vector search will replace the historical kinds of search that are perhaps more precise or at least more transparent to the user about what the search is returning. We think vector search and technologies like it are going to be a supplement to, not a replacement of, more traditional ways of searching.

Operator

Your next question comes from the line of Scott Berg with Needham.

Scott Randolph Berg

I guess I've got a couple here. Kiwi, you talked about some of the cost optimizations that your customers still kind of had in the quarter. How should we think about maybe the magnitude or size of those reductions relative to what you've seen over the last maybe 2 to 4 quarters in this macro backdrop?

Kiwi Camara

I think they're consistent with what we've seen over the last 2 to 4 quarters. Neither larger nor smaller. The problem is that it's different customers. Different customers go through this process at different times depending on their internal budget cycles and prioritization and so on. We're still rolling through the customer base as different customers take a sharper eye on costs.

Scott Randolph Berg

Got it. Helpful. And then I know AI is a big topic. You guys are certainly making a lot of investments with Cecilia and the demo that I recently saw that was very compelling. But is it popping up yet in your sales cycles and your sales processes? Maybe not Cecelia specifically, although it could be, but just the use of AI in general within your platform?

Kiwi Camara

Absolutely. This is our core historical strength, with AI capabilities that have been built into our platform for many years. Those AI capabilities, they're a part of virtually every presentation that we do about the platform because they're a key part of the way that we deliver value to our clients by automating or greatly accelerating a growing range of legal work. The newer AI things, Cecelia as well as some other capabilities that we look forward to announcing shortly, we've already begun demonstrating those quite regularly to a range of clients. As we talked about in our prepared remarks, we now have Cecelia live in private access for real clients on actively litigated client matters, and we've got clients deriving tremendous value. One of the cool things we like to see is when clients use the share capability in our products. You can find a document, share it, and then put a note to your colleagues about what it is you found. And there's a ton of those where people are going in, they're sharing a document, they're saying, oh, I was trying this cool AI thing and look what Cecelia found. And so this is, I think, only going to grow as a focus area in customer conversations, both with new prospects and existing customers.

Operator

Your next question comes from the line of Parker Lane with Stifel.

Jeffrey Parker Lane

Kiwi, you referenced the recent marketing investments you've made as an accelerant to inbound lead volume. I think that doubled year-over-year. Can you take us under the hood of exactly where you're putting those incremental marketing dollars to work? And how much sustainability do you see in the inbound lead volume throughout the remainder of the year?

Kiwi Camara

It's a mix of things. I'm a big believer that marketing programs are interconnected. In the sense that for example when you do brand marketing, it improves the performance of your digital performance marketing, it improves the turnout at events, it improves the win rates and that sort of thing. The biggest new addition is of course the launch of our Law Better campaign and the Lady J ads that maybe you've seen. If not, I encourage you to take a look. That's got tremendous reception with our customer base. Anecdotally, when we talk to our reps, they've started to come into meetings where not only has the customer heard of us before, but they've seen the ad. They like the ad, they talk about the ad, and it can be the beginning of the customer to understand not just our products and offerings, but who we are as a company. That we're a company that cares about the law, that's trying to use technology to make the law work better, and so there can be that kind of alignment in spirit and intent with our customers before we dive into a more detailed discussion about the functionality we can provide and the results that we can drive. I think the big brand campaign is one contributor. But we've also done a lot of work optimizing things like performance marketing, digital marketing, social marketing, and our field events, both large at industry conferences and small local events that we do in different regions to provide opportunities for DISCOvians to connect with our customers. But also, for large existing customers to connect with prospects. That kind of old-fashioned social selling I think is still very effective today.

Jeffrey Parker Lane

And I think we're approaching the 18-month mark of bringing Hold and Request to the platform. Curious if you could just talk about the cross-sell efforts there, how much interest you're seeing in both of those tools and the growth algorithm, how much of that is baked in from these tools in the 2023 outlook and beyond?

Kiwi Camara

Yes, we're really pleased with the reception that those tools have for 2 reasons. One is just it gives customers more places to get started on our platform and more ways to expand. But second, it's that with enterprise customers where we sell directly to the General Counsel, as we've talked about on prior earnings calls, that's been a very big growth driver in terms of new logos and frequently a way those folks will want to get started is Hold and Request. Now they may not wind up buying Hold and Request first. They might start with Ediscovery or Review or something else. Many of them do start with Hold and Request, but very often, the conversation starts with General Counsel's desire to have a product like Hold or Request that can help them comply with incoming legal demands, conduct preservation in place, and also manage their organization's overall migration to the cloud. We've been pleased with Hold and Request. We're actively investing in those products. You'll hear us on earnings calls talk about future enhancements, integrations between those products and the rest of our product suite, and we are overall pleased with the trajectory these products have demonstrated.

Jeffrey Parker Lane

Understood. I appreciate you answering the questions here. Thank you.

Operator

Our next question comes from the line of Mark Schappel with Loop Capital.

Mark William Schappel

Kiwi, is it fair to assume that the Ediscovery product is still generating about 80% of revenues in this quarter?

Kiwi Camara

We don't disaggregate revenue by product line, but I can say that Ediscovery still represents by far the largest share of our revenue.

Mark William Schappel

Okay, great. Thanks. And then I just wondered if you could just talk a little bit about the just kind of relative growth rates of some of the other products, Case Builder or Legal Hold, that you're seeing, how they're coming along.

Kiwi Camara

Some of the other products enjoy smaller bases, and as a consequence, have higher growth rates. In terms of color, let's focus on Case Builder, which you mentioned. I'm really pleased with the growing adoption of Case Builder. We're seeing tremendous growth in terms of the number of matters on the platform, the number of users using Case Builder, and the number of each of the sort of the things that the platform stores. We track number of depositions in Case Builder, number of timelines and chronologies that have been built, number of events that have been created. And all of those metrics for usage are showing very, very strong growth. In addition, we're going to have some announcements coming shortly around building some of this new generative AI technology into Case Builder and other parts of our product suite, and we think those will only accelerate the growth of those products.

Operator

Your next question comes from the line of DJ Hynes from Canaccord.

David E. Hynes

Kiwi, what has the feedback been on proposed pricing mechanisms for Cecilia in your early customer conversations? And I'm wondering kind of how that's informing your strategy as you think about general availability.

Kiwi Camara

We're going to start having the kind of great bulk of those conversations in the coming quarter. In particular, there's a big industry conference coming up called the ILTA at which some of these conversations will start happening. I think it's early days to provide feedback on pricing. Our goal is, over the course of Q3 and Q4, to settle on what the at least call it first year's pricing for the AI platform and AI capabilities will wind up being, and we'll provide an update on our annual earnings call about our thoughts on what AI pricing will be. At a high level, we believe the right mix, and I think I talked about this a bit on the last earnings call, is some mix of a platform-based AI fee with usage-based billing on top for certain AI capabilities.

David E. Hynes

Okay. Got it. And then, Michael, just given the timing of your Q2 restructuring, is it fair to assume that the magnitude of cost savings are not fully reflected in Q2 run rate OpEx?

Michael S. Lafair

Yes.

Operator

Our next question comes from the line of Koji Ikeda with Bank of America.

Koji Ikeda

Listening to the prepared remarks, it sure sounds like you had some Review cases that came in better than expected into Q2. Just kind of taking a step back, and when you originally guided to Q2 on the last earnings call, was there any Review incorporated in that guide? Just trying to understand kind of the puts and takes to the Review performance.

Kiwi Camara

Yes.

Koji Ikeda

Thanks, Kiwi. And then when looking at -- I guess as a follow-up to a previous question and the annualized ARPU, maybe ask the question in a different way. And I think it would be super helpful to -- super helpful to us to start understanding the underlying kind of new versus expand opportunity. But I do realize that you guys only give net revenue retention on an annual basis. Maybe if you could give it in a qualitative aspect, it would be really helpful. The question is, the net revenue retention in the quarter, was it above or below 100?

Kiwi Camara

Well, we only, as you point out, we give the number on an annual basis. But I'm more than happy to provide color, Koji, that may be helpful. I think a couple of things are going on in terms of expansion and retention. First, we still have the falling away of the really big reviews that we had for a sequence of 4 quarters and that now have left. Now we are seeing some green shoots, and this goes to your first question, where we did see Review outperformance. In particular, we had one customer that across a number of reviews it added up to more than $1 million in the quarter. And that's something that we didn't have certainly in the previous quarter. There's that, so Review has shrunk, right? And that is a take from retention. The other thing that's happening on retention is what we talked about on last earnings call, and I think one of the earlier questions here, which is big customers seeking to manage their overall spend in our platform. Which they do by changing the overall volume of usage or changing the way they use our platform. For example, more aggressively using features like ECA and Vault as opposed to active Review in the Ediscovery product as a way of reducing costs. Offsetting that, we still have the more normal dynamic, especially among our smaller and more medium-sized customers, where justice has been proved throughout the history of DISCO, those customers tend to start relatively small, call it 5 figures of spend, and then grow their spend over time as they grow their usage of the first product they started with and as they adopt more products across the product portfolio.

Koji Ikeda

Got it. Thank you, Kiwi, that's super helpful. Thanks so much.

Operator

Our final question comes from the line of Brent Thill with Jefferies.

Brent John Thill

Thank you. This is Luv Sodha on for Brent Thill. I wanted to ask one on the Review side. Kiwi, you mentioned that you now have 60% of reps selling Review products. I guess, are these new sales? Are they starting at a lower volume and then other ramps baked into that? Just any color around the new green shoots around Review.

Kiwi Camara

Sure. I think the way to think about this is, if you have an established sales team, and especially if like DISCO, you were a single product company for many, many years, your sellers get good at selling that one product. And going from 1 to 2 is harder than going from 2 to 3, right? There's a sort of step function to get sellers and to get the market and customers to think of us as a multiproduct company. Or to think of there being many different ways of using our platform rather than a single way of using our platform, that being Ediscovery. Our success at making that jump from single product to multiproduct is what I'm talking about when I share this stat that now more than 60% of our quota-carrying reps generated Review revenue, almost double the percentage for the first half of last year. This I think is good, not only because it's green shoots for the Review business, but also because it's indicative of our sales team and the market and our customers beginning to make that transition from thinking of DISCO as single product to thinking of DISCO as a platform that can be applied across many different kinds of legal work. And we've definitely seen improvements in the multiproduct attach rate if you will, the percentage of customers who are using more than one product.

Brent John Thill

Got it. That's super helpful. And then just one quick follow-up for Michael. Michael, if you look at the EBITDA guidance for the year, I guess the embedded margin for Q4 is now in the low teens versus negative mid-teens guide previously. Just wondering, are you baking in additional leverage from certain line items and what might that be? Thank you.

Michael S. Lafair

Thanks for the question. As you're aware, we've improved our adjusted EBITDA guidance for the full year and also have guided to Q3 improvement on where you expect that we were going to come out in your models. We obviously -- there's 4 levers that are going to hit those numbers, and it's a combination of revenue, the COGS optimization that we've been working on in terms of working with AWS, also both not just for COGS, but also below the line. We talked about this in the prepared remarks, but just a very aggressive managing our kind of non-staff costs around things like software. And then the fourth is really, and we've talked about this before, globalization and attracting the best talent on a worldwide basis, but on a per unit basis, that is more attractive than what we've had in the past. Those are the levers that will impact our overall numbers.

Operator

I will now turn the call back over to Kiwi Camara, Co-Founder and CEO for closing remarks. Your line is open.

Kiwi Camara

Thank you for joining us today, and thank you for your interest in DISCO.

Operator

Ladies and gentlemen, that concludes today's call. Thank you for joining. You may now disconnect.