"The Year of Efficiency" delivered good execution results, while Reels showed great potential (Meta 1Q23 conference call minutes).

The following is the summary of the Q1 2023 performance conference call information for $Meta. For financial report analysis, please refer to "Meta: Reborn After the Tribulation, Fully Revived".

I. Overview from Management

This was a good quarter. We saw strong momentum in both our products and business growth. Our community reached a milestone of over 3 billion people using at least one of our applications every day. Facebook reached a milestone of 200 million daily active users in the US and Canada, following global daily active users of 2 billion in the previous quarter.

First, I would like to share the latest on our efficiency efforts. Our efficiency work is aimed at making us a stronger tech company that builds better products faster, and improves our financial performance to give us space to execute on our ambitious long-term vision in challenging environments. I still believe that slowing down our hiring pace, flattening our management structure, increasing our tech composition percentage, and rigorously prioritizing projects will increase the speed and quality of our work. I also believe that a stronger financial position will enable us to weather the uncertain environment while continuing to focus on our long-term priorities.

So far this year, we've gone through two of the three planned restructurings and headcount reductions in our recruiting and tech teams. In May, we plan to conduct the third wave of headcount reduction in our business group. This is a tough process. But after completing this work, I believe we will provide a more stable environment for our employees. Then for the remainder of the year, I hope we can focus on improving our distributed work mode, providing AI tools to improve productivity, and eliminating unnecessary processes throughout the company.

Products and Business

Our app engagement and monetization efficiency have made good progress, driving good business results. Reels continues to grow rapidly on Facebook and Instagram. Reels has also become more social, with people forwarding Reels over 2 billion times each day, and doubling over the past six months. Reels is also increasing overall app engagement, and we believe our share in short video is increasing.

Now, a key topic I want to discuss today is AI. There are two major tech waves driving our roadmap today, a massive AI wave, and the future metaverse wave.

AI:

We have two main areas of focus in our AI work.

First, a massive ranking and recommendation infrastructure supports all of our products, from feed to Reels, to our ad and integrity systems, which we have been investing in for many years.

Second, new generative foundational models are enabling entirely new categories of products and experiences. Our investment in ranking and recommendation systems is driving many of the results we see today in our discovery engine, Reels, and ads, as well as the content people see from friends and family. Now, more than 20% of the content in your Facebook and Instagram feeds is recommended by AI from people, groups, or accounts you do not follow. This is approximately 40% of the content you see across Instagram. Since we launched Reels, AI recommendations have increased people's time spent by more than 24%. On Instagram, our AI also works to improve monetization. Reels' monetization efficiency has increased more than 30% on Instagram and more than 40% on Facebook. Compared to the previous quarter, daily revenue for our Advantage Shopping Ads series has grown sevenfold over the past six months.

Our work on delivering business messaging as the next pillar of our business is also making progress. Last quarter, I shared that ad revenue had reached an annual run rate of $10 billion. Since then, the number of businesses using our other commercial messaging services (pay-to-message delivery on WhatsApp) has grown by 40% on a quarter-over-quarter basis. Our success here depends on delivering results for other businesses and generating macroeconomic impact through our scale. We recently partnered with economists at UC Berkeley to understand the impact of our service. Their conclusion was that every dollar spent on our ads generated $3.31 in revenue for US advertisers. This means that over $400 billion in economic activity is connected to and dependent on our platform in the form of supply chains, supporting more than 3 million jobs.

Apart from recommendations, another major focus of our AI work is base models to support many new use cases, including generative AI. It has been an incredibly exciting year of progress in this area, and the ongoing work will impact every one of our applications and services. I'm very excited to announce more about what we're building in the coming months.

I believe there's an opportunity to introduce AI agents to billions of people in useful and meaningful ways. We’re exploring chat experiences, WhatsApp and Messenger, posts, Facebook and Instagram's visual creation tools, video, and multimodal experiences, and more. I hope these tools are valuable to everyone from regular people to creators to businesses. Once we establish these experiences, I expect there will be strong interest in AI agents for business messaging and customer support.

Over time, this will also extend to our work in the metaverse, where people will be able to create avatars, objects, worlds, and code them together more easily. Next, let's talk about the technology platform to achieve this. Most companies training large language models have a business model that leads them to adopt closed development approaches. I believe there’s an important opportunity in the industry to help create an open ecosystem. If we can help be a part of that, then I believe the majority of the industry will standardize on these open tools, helping to drive further improvement, making it easier for other companies to integrate with our products and platforms. Due to the fact that we have enabled more integrations, this will help keep our products in the lead.

Our attitude towards artificial intelligence and infrastructure has always been quite open. We have open sourced many of the most advanced models so that people can experiment and build with them. This quarter, we released our LLaMA LLM to researchers. It has 650 billion parameters, but performs better than large models and has proven to be very popular. We have also open sourced three other groundbreaking vision models along with their training data and model weights, shard data, dyno V2, and our animation drawing tool. We have also received some positive feedback on all of these.

Finally, let's talk about artificial intelligence infrastructure and capital expenditures. A few years ago, this was a major investment and I asked our infrastructure team to develop an ambitious plan to build enough capacity to support not only our existing products, but also enough buffer capacity for major new products. This has been the main driver of our capital expenditures over the past few years.

At this point, we are no longer lagging behind in building artificial intelligence infrastructure. Instead, we now have the ability to take a leadership position in the field on a large scale. As these new models and use cases continue to expand, we will need to continue investing in infrastructure, although once we can measure the usage of some new products, we will better understand the trajectory of our investment later this year.

Metaverse:

Now, in addition to artificial intelligence, another major technological trend we are focused on is the metaverse. For years, we have been focused on both artificial intelligence and metaverse, and we will continue to focus on both, as they are also related fields.

The breakthroughs in computer vision have allowed us to launch the first standalone VR device. Augmented reality is built on a range of AI technologies for understanding the physical world and fusing it with digital objects. The ability to programmatically generate worlds is critical to offering immersive experiences at scale. Our vision for AR glasses involves an AI-centric operating system, which we believe will be the basis of next-generation computing. Metaverse technology will also help to advance AI. For example, AI avatars will leverage the deep investments we have made in avatars over the past few years. Building the metaverse is a long-term project, but its underlying principles remain the same, and we remain committed to it.

Now in the short term, we have already achieved some milestones that I think are worth mentioning. Since last year, over a billion Meta Avatars virtual avatars have been created. The number of games in the mission store that generate at least $25 million in revenue has doubled. More than half of daily active users are now using their devices for over an hour.

The next milestone is that we are preparing to launch the next generation of consumer virtual and mixed reality devices later this year. About three years ago at this time, we launched the Quest 2. This was a big step forward for VR and I am really excited to show the world all the improvements and new technologies we have developed since then, at a price that can be used by many people. 2Q23 Guidance:

We expect total revenue for the second quarter of 2023 to be between 29.5 and 32 billion USD.

Our guidance assumes that the proportion of unfavorable foreign exchange effects and year-over-year growth in total revenue in the second quarter will be less than 1%, based on current exchange rates.

We expect total expenditure for the full year of 2023 to be between 86 and 90 billion USD, including restructuring costs of 3 to 5 billion USD related to facility merger costs, severance payments, and other personnel costs.

We continue to expect the operating loss of Reality Labs to increase year over year in 2023.

We expect capital expenditures to be between 30 and 33 billion USD, consistent with our previous estimate. This outlook reflects our continuing efforts to enhance our artificial intelligence capabilities to support advertising, feeds and reels.

We estimate the effective tax rate for the full year 2023 to be approximately 20%.

2. Analyst Q&A

Q1: What is your outlook for the impact of generative AI on the advertising business?

A1: I can start with the generative AI question, although I won't share too many details on that front at this point. As we start to ship more of that in the coming months, that will be a focal point, but I do think there's a huge opportunity there. You specifically asked about advertisers, and I think it will also help to create more engaging experiences, which should create more opportunities for engagement, which itself creates more opportunities for advertisers. But I think there's a lot of opportunities on the visual side to help advertisers create different kinds of creative. Over time, eventually we'll have tools that can do that. We've been trying to empower advertisers to tell us what their objectives are and then enable us to do as much of the work for them as possible. Now being able to do more creative work there for those who want it, I think that could be a really exciting opportunity.

I also think there's going to be a really interesting convergence between some AI agents and commercial messaging, where now we see a lot of commercial messaging places where many businesses can basically afford to have people answer a lot of questions and chat with them. Obviously, once you light up the ability for AI agents to act on behalf of those millions of small businesses, you have more businesses that have the ability to chat with customers. I think that could be a fairly significant opportunity as well. And again, I think this will touch every product and service that we have in multiple ways. It's just a really big wave and a new technology that's available, and the entire company is researching it.

Q2: What are your thoughts on recruiting trends for 2024 and beyond? How does the use of AI internally impact long-term recruiting ideas?

A2: Regarding the statistics on employee growth of 1% to 2%, I think it's worth clarifying - this stems from Mark mentioning in a recent internal employee question that he expects the number of company employees to increase by only 1% to 2% per year. I'd like to clarify that this is actually a reference to the employee budget that has already been allocated to employees. We have previously stated that we have been in a widespread hiring freeze for the past six months. With the completion of our layoffs in April and May, we will resume hiring, and as we ramp up our recruitment channels, we expect employee growth to exceed 1% to 2% by 2024. However, our long-term focus is on efficiency. As resumes come in, you will find that our largest areas of hiring will be in support priority areas, such as generative AI advertising and structural reality experiments, which are areas we have been discussing all along. You also inquired about the use of AI internally and how it affects our long-term hiring plans. Certainly, we do not yet have enough visibility to know how AI will increase our employee productivity, but that is what excites us, and I believe that with more and more tools being developed to improve employee productivity across the industry, we will become clearer about this.

Q3: From a priority or investment perspective, what might the company want to do to become more of a driving force for open systems and artificial intelligence, as we saw in Web 2.0?

A3: I think there is an important distinction between the products we offer and many of the technological infrastructures. Therefore, for the software we write to support this. Historically, whether it is the open computing projects we have completed or simply open-sourcing many of the infrastructures we build, we have been open-source throughout our history - there are many infrastructures, even if the product itself is not ours.

We have code for an open-source core product or something similar. I think we do this because, unlike some of the other companies in the field, we are not selling cloud computing services, we are trying to maintain the proprietary nature of the different software infrastructures we are building.

For us, it would be better if the industry standardizes on the basic tools we are using. Therefore, we can benefit from the improvements made by others and others can use these tools to reduce the cost of things like open computing, making our business more efficient. So I think to some extent we are just playing a different game on infrastructure, and then it is Google, Microsoft or Amazon that creates different incentives for us.

So overall, I think this will lead us to do more work in open sourcing some of the lower-level models and tools. However, of course, a lot of product work itself will be concrete and specific and combined with what we are doing. So not everything we do is open. Obviously, a bunch of them need to be developed in a way that creates unique value for our products. But I think in terms of basic models, I hope we can push here and help build an open ecosystem, which I think will be an important thing.

Q4: How do you view artificial intelligence as a medium- to long-term driver of data center architecture, and what does this mean for capital intensity?

====== A4: There are two parts to this answer.

Firstly, we announced some updates in our data center architecture last quarter. We are building there, but the new data centers being constructed are actually for capacity a year from now, so they won't come online for a few years. Therefore, we don't have much to share beyond the fact that work is underway.

Secondly, artificial intelligence will drive capital intensity over the next few years. You can really think of our capital expenditures as being invested in three big baskets.

The first is non-AI computing needs that we have discussed before. We do have ongoing general computing and storage needs to support existing businesses, but we have become more efficient in capital intensity in this area and we are very focused on continuing to do so over time.

The second area is our core AI investments, which is actually most of our AI investment today, supporting building discovery engines, ranking unconnected organic content, ranking ads, and we focus on measuring the returns on these investments and investing and ensuring that we are satisfied with the return on investment we have made there. This will really drive our plans for core AI spending in the future.

The third is actually around capital expenditure investment in support of generative AI, which is a new opportunity for us. We are still in the early stages of understanding various applications and potential use cases. I do think that this could be an important investment opportunity for us, in comparison to some of the other AI work we are doing, it is earlier on the return curve. It is too early to say how this will affect our overall capital intensity. Therefore, if we spend more on capital expenditure in the near future, I expect it to be primarily to drive this work.

But what I want to emphasize is that we remain focused on striking the appropriate balance between building the AI capabilities we need and efficiently using our costs.

Q5: How do you view the relative competitive position of digital advertising?

A5: We have talked a lot over the past few years about the landscape post-ATT (App Tracking Transparency) and our position in the advertising business. We are definitely making progress in mitigating the direct impacts of the ATT platform changes. But I think that is the reality of the online advertising environment we are in right now. We have invested not only in the direct mitigations we discussed in the past, including conversion APIs, and so on.

But really, now we are focused on improving ad performance in the current signal environment through our investments, both by increasing ads with on-site targets, and secondly, by our AI investments, which are improving the effectiveness of ads and enabling us to introduce new tools and features for advertisers.

So, I think we are in a pretty advantageous position for on-site conversions. We have a lot of products that are growing well for us. I think we have discussed a lot about the messaging ads and what that means for driving our business growth and new business opportunities. I think other formats that have performed well in terms of website conversion are lead ads and shop ads. Therefore, we are exploring different ad formats. Additionally, we have been investing in the use of artificial intelligence for a long time to improve our ad system. This includes ranking to improve our modeling and measurement, as well as supporting the introduction of new AI-driven tools and products. So I think we feel good about our relative position. Furthermore, this is something we have been doing for a long time, and I think that combined with some stability in the macro environment, allows us to see our work pay off.

Q6: Regarding revenue guidance outlook, what are the advantages and disadvantages that we have seen so far this quarter and beyond?

A6: In terms of revenue performance for the first quarter and acceleration relative to the fourth quarter, we have certainly seen a stronger demand, including the impact of the Ukrainian war that began in the first quarter of 2022. In particular, we have seen Chinese advertisers accelerate their targeting of users in other markets. We believe that this is partly due to the decrease in transportation costs for these advertisers and the relaxation of COVID-19 restrictions. Additionally, the decline in the value of the U.S. dollar also eased growth pressure. Looking ahead to the second quarter and the second half of the year, the growth will certainly be easier to compare as the baseline is low.

Starting from the period of weak demand and entering the first complete quarter, there was no Russian revenue compared to 2Q22. We certainly expect that the headwinds of currency will be less than in the first quarter.

That being said, our second quarter outlook includes broad expectations as we still believe that the macro environment remains volatile. In the past year, the market has absorbed many new developments, including inflation, rising interest rates, and bank instability. It is difficult to fully understand how these dynamics will affect the broader economy, especially the advertising market for the second quarter and the remainder of the year. This is also still a challenging regulatory environment, so this is something that needs to be closely monitored. On the other hand, we are satisfied with the performance we are providing to advertisers and the return on these investments. The core engagement trend remains strong, and Reels is today's revenue headwind. We expect to achieve revenue neutrality by the end of this year or early next year. Similarly, we expect forex to be a tailwind and again impress the relatively weak comparisons. There are many factors to consider when we look to the future, when we look to the second quarter and beyond.

Q7: How does the company's business architecture consider long-term sustainable financial goals and the restructuring efforts made for it?

A7: I think this has several components. We have been conducting a lot of efficiency work, especially this year, not only out of financial need but truly focused on improving operational efficiency. So I think this is one of the most important goals, and a lot of work is built around that. This actually includes a more careful review of the roadmap, gradually reducing projects that are no longer on our priority list, and redefining investment priorities. I think this is really a muscle we have exercised a lot in the past six months. I hope that we will carry this discipline into the way we evaluate future product roadmaps. Similarly, I believe that much of the work we do is focused on improving operational efficiency and ensuring that we are satisfied with our management span and hierarchy. I think all of this work is also geared towards making us a company that can build and ship more and faster.

We want to continue to ensure that we think our organizational posture is really serving us in building and transporting products. As we move forward, including the downsizing and direct restructuring work we have been implementing, we will continue to focus on efficiency work, including other areas such as helping investors/creators build and deliver quality products faster.

We will invest in tools where possible to increase the speed of product development.

We will continue to streamline cross-functional or process burdens that may slow down our speed or hinder development processes, and ensure that we are truly focused on improving the quality and frequency of collaboration, and personally go online to establish relationships throughout the organization and get work done faster. So that's what I think I expect, and all of this will continue in the years to come.

Q8: Regarding Facebook user growth, can you talk about the driving factors in detail?

A8: As far as Facebook's core user growth is concerned, we are satisfied with the growth of the community and the participation trends we see, especially, I think in the first quarter, the growth of DAUs benefited from strong product execution, but I should also point out that the first quarter was a strong seasonality participation quarter for us. So, this has some benefits in terms of quarter-on-quarter growth.

You mentioned whether the work in Reels is driving incremental growth or driving Facebook engagement. What I want to say is, more broadly, the AI discover engine works for us, and our investment in ranking unconnected content goes beyond Instagram, and I think this is definitely a part of driving Facebook growth.

Q9: Regarding the new Opex guidelines, how does the downgrade affect growth expectations after 2023?

A9: We certainly haven't shared any cost guidance for the next few years. Frankly, this is something we are working on internally ourselves because we are making long-term plans, as we are engaged in long-term planning. But I do believe that many of the strengths we have established in efficiency will translate into our long-term roadmap. So, I expect that we have indeed improved our cost structure in the past 6 months, which will form the foundation, while also focusing more on it.

Regarding future operational efficiency. I think this will form the basis of our future outlook and ensure that we continue to evaluate our investment roadmap to ensure that we invest in the top priority work, and reduce investment priorities and gradually reduce them where we do not see return on investment or compelling opportunities.

So, we don't have any guidance for the next few years yet, but I do hope that the overall spirit of efficiency will be a part of the next few years.

Q10: Regarding regulation, will the changes in transatlantic data transfer rules have an impact on European business? A10: First of all, I want to emphasize that we still hope the new EU-US privacy framework will be implemented before the deadline. However, if it involves this, we do not know the details of the final order and how long the intervention order will last, which will be an important variable in determining the overall impact.

What we do know is that about 10% of global advertising revenue comes from ads targeted at Facebook users in EU countries/regions, but we need to know more details, including the impact on advertisers and EU countries, before we can truly provide more accurate or comprehensive estimates.

Q11: Can you explain the restructuring cost guidance of 3-5 billion in detail?

A11: The first quarter restructuring costs are around $1.1 billion. Then we provided guidance of $3-5 billion for the full year. Therefore, we expect to recognize further restructuring costs for the full year.

Q12: Are you considering making more efforts in message advertising?

A12: What are the opportunities for click-to-message ads, especially for WhatsApp? This is an area that we are very excited about. We have invested a lot in it. We see continued opportunities to expand click-to-message, especially in the WhatsApp and Instagram direct areas. Both are in the early stages of their growth curves. We are investing in several areas to make it easier for people and businesses to connect.

First, we are indeed investing and working to simplify ad creation. Therefore, in the past year, we have invested in improving the ad manager experience and providing a native ad creation experience directly in WhatsApp.

We are also expanding our partnerships. So, with the help of cloud-hosted APIs, we have now hosted API solutions on WhatsApp, Messenger, and Instagram so that our developer ecosystem can help build best-in-class experiences for businesses.

Then, we are also working to improve the effectiveness of these ads by driving more actions downstream of the funnel. Therefore, we are working hard on ranking and optimization to improve the conversion rate of these ads and introducing new ways to improve their effectiveness. We have introduced new features such as thread payments and other business tools. So, we believe there is a huge opportunity here. We are working hard to make every part of the advertiser experience easier, better, and more effective.

We are pleased to see growth outside of core markets. In Southeast Asia and Latin America, messaging has become a common way for businesses to interact with customers and potential customers. But we hope to see this situation take off in other markets as well, seeing strong growth in the United States and Canada, but that's still too early today.

Q13: Regarding the open-source of generative artificial intelligence models and infrastructure, is this an opportunity for developers to participate and help improve your application series to generate revenue? Is there a possibility of other forms of business cooperation? A13: You know, when it comes to AI tools, we have a history here, right? So if you look at the work we've done with PyTorch, it has become the industry standard as a tool for many people who are building AI models and different things in that space. And that's something that's been valuable for us to provide because now all the best developers in the industry are using those tools, we're using them internally as well, so the toolchain is the same.

So when they create something innovative, we can easily integrate that into what we're doing. And when we improve something, it improves other products as well because it's integrated with our tech stack. And when there's an opportunity to integrate with products, it's much easier for developers and others to work in a systematic way that's compatible with what we need.

I prefer to think of it as a backend infrastructure advantage that has potential integration with products, but one that allows us to stay ahead and integrate more broadly with the community, and make running all this infrastructure more efficient over time. I just used PyTorch as an example. Open compute is another way in which we're able to effectively integrate innovation and scaling efficiency into our own infrastructure. So I know our incentives, and I think they're basically pointing in that direction.

Having said that, there's still a lot to figure out. All of this is early in development. The better we do in the foundation work, I think the more opportunities will arise and be showcased. So I think these are all things we need to figure out. But at least at the foundation level, I think we're generally incentivized to move in that direction, and we need to figure out how to do that over time.

I mentioned LLaMA earlier and I also want to clarify that when I talk about helping contribute and opening up ecosystems, LLaMA is a model that we only truly offer to researchers, and there's a lot of great things happening there. But a lot of the work we're doing, I think we're eager and hoping to be more open than that. So we need to find a way to do that.

Q14: How is Instagram considering increasing investment in Reels to maintain market growth among users?

A14: We don't have a quantitative expectation of user engagement metrics, but we're very pleased with the boost we've seen from Reels so far in terms of incremental engagement on the platform.

We've talked about how the incremental growth has been increasing over time, and that is important as we continue to focus not only on the absolute growth of Reels but also the level of engagement. That's where we're very pleased with the trend that we're seeing, with all the metrics that we're watching like plays, watchtime, shares, and all the metrics that we are focused on. And on these metrics, we're seeing significant growth from Reels. It's obvious that people value short videos and recommendations. This opens up a whole new pool of content, which we believe creates engagement opportunities because we can help people discover more interesting posts. And this is really valuable on top of the social graph. We see shared engagement with Reels taking off with a doubling of growth over the past six months.

To more widely introduce how the discovery engine is driving incremental engagement across the entire platform, our feed recommendations certainly go well beyond Reels. They cover all types of content, including text, images, links, group content, etc.

On Facebook, we see AI-driven recommendations continuing to grow and the increase in application engagement being praised. So again, we haven't quantified this, but I think it's a place where we're all pleased with our progress and see an important opportunity to do better.

Q15: Thank you for sharing your confidence in the future profitability of Reels. Can you share some factors that can accelerate the growth of Reels? Can you share some data on ad loading rates and ROAS advertising ROI?

A15: We're addressing a growth bottleneck. That is, Reels consumes time from Feed and Stories, which in turn leads to a decrease in revenue. Since Reels' monetization efficiency is relatively low, we're balancing the flow of traffic from these three sources. This may make it less easy for Reels to exceed revenue expectations this year.

The economic model of Reels depends on two factors that affect each other: the improvement in user participation brought about by Reels and the low monetization efficiency of Reels. We're executing against the goal of balancing revenue and expenses by the end of the year.

But I also want to remind you that Reels and Feed/Stories have completely different structures. This makes it impossible to compare the monetization value per unit time between them. However, since it does bring about an increase in user engagement, it can still improve our overall commercialization capabilities.

Q16: Can you share some new developments regarding messenger advertising?

A16: Messenger commercial advertising is a relatively important part of our overall advertising strategy. It involves first-hand user data feedback. Of course, this is one of the ways we profit from messaging behavior that occurs on the platform and creates a powerful communication channel for businesses to directly connect with consumers, and we hope to expand this channel through a wider range of commercial messaging products over time. So we're very excited. We're excited about the progress we're making with click-to-message ads and the business opportunities we have there.

Risk disclosure and statement for this article: Dolphin Disclaimer and General Disclosure