Ladies and gentlemen, thank you for standing by. Welcome to 4Q and fiscal year 2020 financial results conference call. At this time all participants are listen only mode, after speaker presentation there will be question-and-answer session to ask question during the session you will press star and one on your telephone please be advice that this conference is being recorded. I would like to hand conference over to your first speaker today, Ms. Fionna Chen. Thank you. Please go ahead.
Thank you, operator. Good evening, and good morning, everyone. My name is Fionna. I'm the Investor Relations at Agora. Thank you for joining Agora's fourth quarter 2020 earnings conference call. Joining me today are Tony Zhao, Founder, Chairman, and CEO, and Jingbo Wang, our CFO. Our earnings results press release and a slide deck can be found on our IR website at investor.agora.io. Reconciliations between our GAAP and the non-GAAP results can be found in our earnings press release. During this call, we will make forward-looking statements about our future financial performance and other future events and trends, including guidance. These statements are only predictions that are based on what we believe today, the actual results may differ materially. These forward-looking statements are subject to risks, uncertainties, assumptions, and other factors that could affect our financial results and the performance of our business.
We will discuss them in detail in our filings with the SEC, including today's earnings press release, and the risk factors and other information contained in the final prospectus relating to our initial public offering. Agora assumes no obligation to update any forward-looking statements we may make on today's call. With that, let me turn over to Tony.
Thank you, Fionna, and welcome everyone. Reflecting on the year, the global pandemic has brought tremendous changes and challenges to every one of us. Despite the year's trauma, we have seen a significant acceleration in demand for online ways of living and working. There is a growing need for contextual real-time video or voice engagement in almost any app we use. We are proud to be there when the world needed the real-time engagement more than ever before. On earnings, we finished the fourth quarte with revenue of $33.3 million, up 74% year-over-year. We reached more than 270,000 registered apps at the end of December, adding nearly 9,000 per month in the fourth quarte. Our number of active customers reached nearly 2,100, up 101% year-over-year.
For the full year of 2020, our revenue was $134 million, which represents an increase of 107% year-over-year. Besides numbers, 2020 was a year of innovation. We crossed a number of noteworthy milestones, such as powering 40 billion minutes of engagement per month. We continue to create industry firsts, such as our experience level agreement or XLA. We further broadened our product offerings, optimized our technology, and helped developers around the world to create many innovative use cases. Out of the many new use cases we empowered, I want to elaborate on two of them here. The first one is interactive lecture hall. A large size classroom with more than 100 students is widely considered as the most economical model for live online education.
Previously, such classes are taught through one-way video broadcasting powered by traditional content delivery technology, where latency is anywhere between three and 10 seconds or even more. Given the high and uneven latency, there is very little room for students to interact with teacher or among themselves. Recently, we have been working with leading online education institutions on redefining the lecture hall experience. Here, the interactive lecture hall runs entirely on our real-time engagement platform and switches between one-way broadcasting session and fully interactive breakout sessions. During the one-way broadcasting sessions, students can enjoy real-time non-video interactions, such as voting and posting emojis. During the breakout sessions, students form small groups to work on assignments together. We believe that interactive lecture hall combines the engaging learning experience of small class and the cost-effectiveness of large classroom, and it is an emerging trend in online education.
The second one is audio live cast. I believe most people are familiar with audio podcast, but what is audio live cast? Podcast is like radio channels. There is one speaker, or sometimes multiple speakers in the same physical room, and all the audience simply listen to the speakers. Live cast, on the other hand, create a virtual in the same room experience. Here, multiple speakers can join from different locations worldwide and discuss spontaneously, thanks to ultra-low latency audio. Instead of just listening, the audience can raise their hand and come on stage to join the speakers anytime. We believe Live Cast represents a new way for people to connect and learn, and we are particularly proud that our proprietary NOVA audio codec further enhance the experience by offering crystal clear HD audio.
These are just two examples out of the many new use cases we see, such as virtual events, remote assistance, remote collaboration, and virtual tours. We are really encouraged by the enthusiasm and creativity from our developers. We also made solid progress on new products in this quarter. Recently, we released the first version of Agora Flexible Classroom, a low-code application PaaS solution for education providers. It combines video, voice, messaging, whiteboard, and recording functionalities into one cloud-based solution that allows developers to build an online classroom in a matter of minutes. Compared with our standard SDK, Agora Flexible Classroom can significantly simplify software development and shorten time to market. Compared with out-of-the-box SaaS solution, Agora Flexible Classroom allows developers to use their own brands and have complete control of their own user data.
In addition, Agora Flexible Classroom is modular, which means developer can customize their application to create differentiated user experience. Agora Flexible Classroom represents an important step in our low-code effort, which we believe will reduce the friction of adoption and expand our addressable markets. Moving on to the M&A slide. We recently agreed to acquire Easemob, a leading instant messaging API provider, and complete the acquisition of Netless, a leading interactive whiteboard API provider. Both companies share our developer-first philosophy, they bring highly complementary APIs to our product portfolio, which our developers and customers often ask for. As a pioneer and longtime market leader, Easemob also has a large and vibrant developer community. We believe that with these two acquisitions, we are uniquely well-positioned to help developers build more immersive Real-Time Engagement use cases. Last but not the least, security, compliance, and privacy protection are critical to our success.
Recently, we received SOC 2 Type 1 certificate as certified by Deloitte. Our network penetration, application vulnerability, and compliance assessment were also recently completed by Trustwave Holdings, a global security specialist. Looking forward, we will continue to work with leading experts to ensure that our security practice remains best in class. Overall, 2020 was a milestone year for us in many ways. We launched Agora seven years ago in a garage in Silicon Valley, and it's amazing how much we have accomplished in the past seven years. I want to take this moment to thank all the developers, customers, partners, and investors for your trust in us. Above all, to all Agorans, including new Agorans from Easemob and Netless, thank you for your hard work and commitment to serving our developers and customers. There is a long journey ahead of us toward ubiquitous real-time engagement.
Let's start fresh, just like seven years ago. Let me turn things over to Jingbo, who will reveal our financial results.
Thank you, Tony. Hello, everyone. Let me start first by reviewing our financial results for Q4, and then I will discuss our outlook for the fiscal year 2021. Total revenues grew 74% year-over-year to $33.3 million in the fourth quarter of 2020. Total revenues for the fiscal year 2020 were $133.6 million, which represented 107% year-over-year growth and exceeded the high end of our guidance by $3.6 million. In 2020, we have powered more than 500 billion minutes of real-time engagements in total. As we mentioned in previous earnings calls, our revenues in Q1 and Q2 this year were positively impacted by the spike of usage due to COVID-19 in China. In order to help investors better understand our organic growth, excluding such short-term impact, we calculate Q1 and Q2 adjusted total revenues as follows.
For each customer in China, we use its revenue in Q4 2019 as a starting point, and revenue in Q3 2020 as the endpoint, and calculate its adjusted revenue in Q1, Q2 as an arithmetic progression from the starting point to the endpoint. Revenues from customers outside China remain unchanged. This would lead to adjusted total revenues of $22.2 million in Q1, and $27 million in Q2. The adjusted total revenues for the full year of 2020 were $113.3 million. Our trailing 12 months constant currency dollar-based net expansion rate is 179%. If we use adjusted total revenues, the adjusted expansion rate would be 149%. Now turning to cost, expense, and margin. I will focus on non-GAAP results, which exclude share-based compensation expense. Non-GAAP gross margin for the fourth quarter was 60.5%, which was 5.9% lower than Q4 last year, 2% lower than Q3 this year.
The decrease in gross margin was mainly due to our international expansion into regions with higher infrastructure costs, such as Southeast Asia, South America, and Oceania, as well as capacity expansion in anticipation of future usage growth. Non-GAAP R&D expense were $13 million in Q4, up 88% year-over-year, as we continue to build our R&D team. Non-GAAP R&D expenses were 39.2% of total revenues in the quarter, compared to 36.4% in Q4 last year. Looking forward, we'll continue to invest significant resources in our R&D capabilities in order to further strengthen our technology leadership and broaden our API portfolio. Non-GAAP sales and marketing expenses were $6.9 million in Q4, up 34% year-over-year, mainly attributable to team expansion and the increased advertising and event expenses. For example, our RTE virtual conferences in October.
Sales and marketing expenses represented 20.8% of total revenues in the quarter, compared to 27.1% in Q4 last year. We believe this has again demonstrated the efficiency and scalability of a developer-centric go-to-market model. Non-GAAP G&A expenses were $5.7 million in Q4, up 163% year-over-year, mainly due to team expansion and professional service fees. G&A expenses represented 17% of total revenues in the quarter, compared to 11.2% in Q4 last year. Non-GAAP operating loss was $4.8 million, translating to a 14.4% non-GAAP operating loss margin fourth quarter, compared to a net loss margin of 8.1% in Q4 last year. Turning to cash flow. Our operating cash flow was positive $2 million in Q4 compared to positive $2.1 million last year. Free cash flow was negative $1.4 million compared to positive $0.9 million last year.
Net cash outflow in Q4 was mainly due to the purchase of servers and networks equipment as we continue to scale our business. Moving on to balance sheet. We ended Q4 with $635.4 million in cash equivalents, and short-term investments, compared to $635 million at the end of Q3. Subsequently, we raised $250 million by issuing ordinary shares to an accredited investor earlier this month. Now turning to guidance. COVID-19 is still an unprecedented variable to our business model, where historical experience may not apply. Our guidance on full-year revenue reflects a number of assumptions that are subject to change based on the uncertainties related to the impact of COVID-19. With that, we currently expect total revenues for the fiscal year 2021 to be in the range of $178 million-$182 million, which would represent approximately 35% year-over-year growth at the midpoint.
If we use adjusted total revenues for 2020, 59% year-over-year growth at midpoint. In closing, we executed very well and are proud of our strong performance we delivered in our first year as a public company. We'll continue to empower developers and meet customer needs around the world. Thank you to the entire Agora team for your hard work this past year and hope you're healthy and safe. Let's open up for questions.
Ladies and gentlemen, we will now begin the question-and-answer session. To ask a question during the session, you will need to press star one on your telephone. If you wish to ask a question, please press star one on your telephone and wait for a name to be announced. If you wish to cancel your request, please press the pound or hash key. Your first question comes from the line of [Yang Yu]. Please ask your question.
Thanks for the opportunity to ask questions. Three questions from my side. The first one is, can management share the revenue upside from penetrating the audio live broadcast use cases? How should we think about the revenue upside and how to quantify that? Do we have other customer beyond Clubhouse? The second question is, how much is Easemob revenue contribution in the full year 2021 guidance? What is the organic revenue growth in the guidance? The last question is gross margin outlook, because we have new use cases, new acquisition, et cetera, do you think the fourth quarter 2020 gross margin is sustainable going forward? Thank you.
Thank you. I guess I'll take all the three questions. On the first question, first of all, we have internal policies on customer disclosure, and we cannot comment on particular customers other than what's already disclosed on our website. We'll not comment on any particular customer names. Regarding this use case, it's actually not an entirely new use case. There have been previously what we call the audio chat room, was a existing use case for a long time. Obviously recently there have been a lot of new developments in the area. We do see a significant pickup in usage. With that said, because audio price per minute is much lower than video price per minute, so audio live cast, as a whole, the revenue contribution is not that significant if you think about the overall picture.
We wouldn't guide a very significant revenue upside from this one use case alone, given we have so many use cases on the platform. On the second question, currently Easemob has a revenue run rate of around $1 million a month. We are currently in the closing stage of the acquisition. We expect the acquisition to complete in the next few weeks. With that, I would say the full year contribution would be something around $10 million. Lastly, on gross margin. Our current view is we think there will be short-term pressure on gross margin in the near term. We do not believe it will be very different from what we have already seen in Q4. There are mainly two factors at play. The first factor is international expansion.
As we mentioned before, currently given smaller scales in certain markets and higher infrastructure cost in those markets, our margin is relatively lower in those new markets. As our revenue contribution from these markets increase, that will cause a drag on the overall GP margin. The second factor is all the technical optimizations we are implementing. In the short term, it's likely that the first factor will be stronger, so we'll have near-term pressure on the margin. We do expect things should recover in a few quarters time as more and more optimizations take effect gradually.
Thank you.
Your next question comes from the line of Rich Valera from Needham.
My final one is on OpEx. I was wondering if you could give us any help for modeling in terms of how to think of OpEx in 2021 versus 2020, either in absolute dollars or in percentage of revenue.
We do not provide specific guidance on exact operating margin. As we mentioned earlier, we'll continue to invest very heavily on R&D. R&D will remain the main component within OpEx. G&A, the ratio we saw in Q4 is high. We do think it will be lower than that given we had a few professional fees projects related to security audit, as Tony mentioned earlier in his opening remarks, completed in Q4. Q4, we have a few non-recurring items. The G&A expense would be lower as percentage of revenue. Sales and marketing would be around similar levels as percentage of revenue compared to 2020.
Got it. Thank you very much.
Your next question comes from the line of Emerson Chan from Bank of America. Please ask your question.
Hi. Thank you, management. I have three questions. My first question is about our overseas market. How much revenue is contributed from overseas in Q4, and what we see differently now versus, let's say, one year ago in terms of customer acquisition in the overseas market, in terms, of course, our brand awareness and the developer outreach? My second question is regarding the RTE use case in the audio live cast we mentioned before. How should we think about our technology leadership in the audio use case, whether it is less significant on audios than the video use case, given audio traffic may look less demanding? What are our competitive advantage on the audio use case that our competitors do not offer? My last question is about the usage of the audio live cast use case.
I just wonder if listen-only audience also need to use RTE, and what are the alternative technology that can accommodate those listen-only audience, and what are their quality and cost difference versus our RTE? Thank you.
Thanks, Emerson. I will take the first question, and Tony will take the other two questions. In terms of overseas revenue mix, in the most recent quarter in Q4, actually revenue from U.S. and other markets, so non-China markets, was already close to 30%. It's a little less than 30%. It's significantly more than, let's say, one year ago or two year ago. In terms of the overall situation in the non-China market, we have seen developer sign-up continue to be very strong worldwide. The monthly developer registration, so new apps registered on the platform outside China right now is almost three times as many as one year ago. If you just do some simple search engine search on the main keyword, like video call, video streaming, you can see that a lot of times we are already ahead of major competitors.
We have come a long way, and we are in a much stronger situation compared to one or two years ago. There's still a lot of work to do. We recently added quite a few development frameworks such as React Native, Flutter, Unreal, and we are working on a few more. We recently launched the startup program. We'll better help startups to get free credit, get technical support so they can better build their innovative use cases. Overall, we think we are on the right track, but we still need to be more proactive and be more strategic to really solidify our position as the go-to platform for the RTE APIs worldwide.
I'll address the question around our technical or technology leadership, whether it's less significant on audio-only use cases and about the advantages. It's actually very clear to us that our technology leadership is not less significant on audio-only use cases. I'll explain why. Video is commonly known for more demanding than audio in terms of traffic and CPU usage. That doesn't mean audio is easy to do, because people are more sensitive to audio quality. For example, in a video call, if the video frame freeze for a second or so, or has some color blocks for a few seconds, it's less than ideal, but you won't feel too uncomfortable or hard to continue the conversation.
If the audio breaks up for a second or a few seconds, or there's a sharp noise happening, like in this call, then you would very much likely to feel really uncomfortable or not able to continue the conversation. In terms of our advantage in audio use cases, there are at least three of those advantages. We have a proprietary codec to provide higher quality, and we have a much better network to cover global audience. Also we leverage AI technology to further enhance audio experience in our services. We have a proprietary codec called NOVA, which is widely used in today's live cast use case, where it provide full band audio with very low bit rate, which is the killer feature for audio live cast use cases.
Again, our network, which is called SD-RTN, provides the best global coverage in this market, in any country, especially across long distance. It's hard to really ensure the experience across region and various type of devices. We also use artificial intelligence to improve various aspect of audio quality. For example, AI-based noise reduction and cancellation. On the third question were about listening-only audience use cases on RTE. I think, it's important to understand that listen only use cases is so-called podcast. Listen most of the time, but can jump in any time to discuss is hugely different from strictly listening only. You need to understand, because this is a new use case, where psychologically, the audience or the participant feel really different from strictly listening only. Having the ability to interact gives people the feeling of being there or being together.
Whether they make any response or not means a lot in that environment. People do come on stage to interact from time to time when they hear something they are really passionate about, which makes the whole discussion more spontaneous. That's very different from strictly listening only. That's why we see it's necessary for such use case to use RTE.
Thank you.
Your next question comes from the line of Colin Liu from China Renaissance. Please ask your question.
Good morning, management. Thanks for the opportunity to ask the question. I have just one on the acquisition of Easemob. With the IM capabilities to be added to Agora's platform, what do you think of the new business cases we should actually focus on? Are these new business cases going to bring some new customers that we may never across in the past before? Also, I noticed that, actually before the acquisition and the cooperation with Agora, Easemob also outlined some overseas expansion plan, particularly in the ASEAN market. How do we see the potential opportunities in overseas markets regarding the cooperation with Easemob in the future? Thank you.
I would say the acquisition of Easemob and the addition of instant messaging API, is less about penetrating into customers we can never penetrate before. It's more about offering a more complete and more seamless solution to customers. Actually, this is not something we imagined. It's during our experience with developers and with customers. In fact, instant messaging API is one of the most frequently requested feature from our developer community. If you imagine all the apps you have used, if an app has a video or voice engagement functionality embedded, actually, there's a very high chance it also need instant messaging function. Now with IM APIs on the platform as well, we can offer these developer and customer a more seamless integration of our features. If you think about it, actually, the primary way for people to interact in an app, basically video, voice, and messaging.
Now we have all of these covered. That's really the logic. We think this is a very natural move.
I think to add a little bit to that, the instant messenger APIs falls under a bigger umbrella of RTE APIs, which we've been working on for a long time. It does give us complementary features and more APIs to serve our customer base, which will help customer to more easily to build their actual use cases or their apps. This will definitely help us a lot in growing our developers. On international expansion, there are past effort for Easemob to expand internationally. I think with acquisition, they can leverage our existing operation globally, which proven to be successful in past to further penetrate to serve global developers instead of in the past, more so in China developers.
Very clear. Thank you.
Your next question comes from the line of Eric Wen from Blue Lotus. Please ask your question.
Oh, good morning. Thanks for answering my questions, and congratulations on the good quarter. My question is regarding this other revenues. We noticed that other revenues has grown quite a bit this quarter. Does this represent trial customers or more customization requirements, or it represents a new class of revenues we plan to continue to develop in the future? If you can comment on what is the gross margin profile of this other revenue, it'd be very helpful. Thanks.
Sure. This other revenue is mainly due to one enterprise customer where we provide essentially the same product and same service, but the contract is structured as a license as requested by the customer. It's booked as other revenue. Essentially, it's really the same service. Given this is one case with one particular customer, we actually don't think this will constitute a new line of business. It seems like this will happen from time to time, but we don't consider this as anything different from what we already do.
Okay, thanks.
Once again, ladies and gentlemen, if you wish to ask a question, please press star one on your telephone and wait for your name to be announced. Your next question comes from the line of Akida, Eric Kim from China Securities. Please ask your question.
Thank you, management, for taking my question. There are three questions from my side. The first one is about online education. I was wondering, could management update us in terms of progress of penetrating online education customers? At present, what is the proportion of revenue contributed by small classes versus large classes? How they see the potential market size of RTE in online education sector in China. My second question is about the strategy of product development and R&D. What is the focus of R&D and product development currently, and what's the most challenging part? Apart from education and entertainment sectors, are there any innovative RTE application emerging from other fields such as finance and IoT and so on? My third question is about overseas business.
For 2021 and also for the mid to long term, what is the expected proportion of revenue contributed by overseas business? It would be appreciated if you can talk a bit more about your strategy to develop the global market. Thank you.
Thank you. In terms of revenue contribution from different class formats. The small class is still the majority format. We have about half of the revenue from small class. The other half, it's between one-on-one use cases and then our large class or the lecture hall. We believe lecture hall use case, as Tony discussed earlier, has a significant potential, because as Tony explained, that class format is most economical. You can have one teacher or few teacher teaching hundreds or even thousands of students. The model itself has great economics. We do expect the contribution from large class or lecture hall to increase substantially in the future. With that said, we think even for small class and one-on-one class, there is still a lot of room to grow.
Currently, the overall online education market is growing very rapidly, not just in China, but globally. Secondly, there are still significant volume, significant usage currently served by in-house solutions. We are working hard to convert more and more customers to switch from in-house solutions to professional third-party solution, like our own solution. That's the first question. On the second one, we actually don't think the technology is perfect. In fact, we think the technology is far from perfect at the moment. If you actually ask some of our customers, they will tell you they like our quality, relative to competitors. They don't think the quality is good enough in an absolute sense. If our quality can be 20%, 30% better, they are more than happy to pay us more.
If you think about use cases like high definition 4K or VR/AR, or for a more challenging use case across geography, there's still a long way to go. If you think about the use cases, I would say, people talk about social and education. Actually, under these two umbrella, there are numerous new innovative use cases, like audio live cast, right? That's still social, but that's probably something people didn't do before. Even within the two verticals, we think there is still a lot of room to grow. It is hard to predict what new use cases will emerge, because by definition, right, it's new because we don't know it yet. We haven't developed it yet, so it's hard to predict. As history has shown, right, people will always innovate, and the need to learn and the need to interact will not go away.
That's on the verticals. We will not just stay at the PaaS or SDK level and wait for customers to build apps. We'll also, at the same time, try to verticalize ourselves. The education aPaaS, Agora Flexible Classroom, is one of the first efforts from us in terms of verticalization. Because aPaaS is a low-code solution, basically it will reduce friction of adoption and allow companies with a smaller development team, or let's say a more traditional enterprise rather than a pure technology company, will allow them to be able to adopt our technology and build apps much more easily. That will expand the addressable market, addressable developer or customer base, and that can also create new revenue opportunities. That's the second question. Third question on overseas revenue.
This year, it's hard to say, given a lot of things are moving, including the development of the COVID-19 situation, right? We expect random contribution from outside China will be similar or slightly higher than what we saw in Q4, so around 30%. We do think the non-China market has huge potential given all the new use cases are starting to really ramp up and gain popularity. We do expect to have revenues from outside China to rival revenues in China, from within China in three to five years.
Thank you.
Once again ladies and gentlemen if you wish to ask a question, please press star one on your telephone and wait for your answer to be announced. You have a follow-up question from the line of Emerson Chan from Bank of America. Please ask your question.
Hi, management. May I have one follow-up questions on the overseas market? I know we have a global pricing plan, but as far as I know, large customer may have some discount. I'm curious on what are the ASP difference in the overseas versus China after the discount we given to our client, and how should we think about the competition intensity from the overseas market versus China? Thank you.
In fact, in terms of ASP, in the most recent quarter, the two markets are not that different. Non-China markets still has a slightly lower ASP, it's not drastically different. I'm talking about real ASP, the after-discount ASP. As we mentioned earlier, the challenge is really the infrastructure cost. The scale in China, the network capacity in China is much higher, we are able to get a relatively efficient procurement cost. In other China, although the overall scale is not as small, but that's distributed across many different geographies, many different countries. In one particular country, the scale could be quite small, that cause the infrastructure cost to be higher. That's the challenge at the moment.
In terms of competition, I guess people have been saying that recently there have been new entrants into the market, including some large companies which we actually think, again, confirms the potential of this market. In terms of difference, I would say the non-China, U.S., and rest of the world market is less well-defined and still shaping up compared with China. In China, the market has been well-defined primarily by us, and there are competitors who compete with us in a more direct way. Also in China, I think we also play a key role in defining the market. We see competitors from various different backgrounds, and they try to offer some overlapping products. Very few are competing with everything we do. Some of them do a little bit of video call, some do a little bit of voice, some do streaming.
We don't see a lot of competitors really try to do the same thing we do, that is, to power all kinds of immersive engagement use cases. I would say the market of China overall is still less developed.
Once again, ladies and gentlemen, if you wish to ask a question, please press star one on your telephone. There are no further questions at this time. I would like to hand conference back to today's presenters. Please continue.
Thank you everyone. Feel free to email us or contact us through investor.agora.io. Thank you again.
Yes, ladies and gentlemen, this concludes today's conference call. Thank you for participating. You may all disconnect.
Thank you. Goodbye