Keep saying 'efficiency' non-stop, Mark Zuckerberg has learned to be cooperative. (Meta Q4 2022 Performance Conference Call Summary)

The following is $Macys(M.US)eta.US 22Q4 earnings call summary, financial report interpretation can refer to Good News Stacked, Meta's Gorgeous Turnaround? .

1. Management Overview

2022 is a challenging year, but we have made good progress on our main priorities, with over 3.7 billion people using our app family every month. On Facebook, our DAU reached 2 billion and MAU approached 3 billion. The number of people using Facebook, Instagram and WhatsApp every day reached an all-time high.

The management theme for 2023 is the year of efficiency. Last year, we made some layoffs and adjustments to our business lines. Since then, we have taken some additional steps, such as:

1. Working with our infrastructure team to study how to achieve our development roadmap and reduce capital expenditures.

2. Simplify our organizational structure and eliminate some middle-level management so that we can make decisions faster.

3. Deploy AI tools to help our engineers work more efficiently and be more active in cutting projects to focus more on improving our productivity, speed, and cost structure.

2. Drive the two waves that ultimately lead to our roadmap - AI and Metaverse

(1) AI

Let's talk about our AI discover engine. Facebook and Instagram are shifting away from organizing around people and accounts you follow to increasingly showing related content recommended by our AI system.

Reels are growing so fast that we are paying close attention to short videos. In the past year, Reels' view on Facebook and Instagram has more than doubled, and the social sharing part of Reels' growth has increased even faster. In the past six months alone, the number of Reels views on these two apps has more than doubled.

The next direction we are focusing on for Reels' continued growth is to improve its monetization efficiency - how much revenue Reels generates per minute. Currently, Reels' profit efficiency is much lower than Feed. So even if Reels takes up a portion of Feed's time, we are still losing money due to its time increase. However, people want to see more Reels content.

The key to achieving this is to increase our monetization rate so that we can display more content without losing more money. Our monetization rate on Facebook has doubled in the past six months. Given the revenue headwinds, we expect Reels to remain neutral in profitability by the end of this year or early next year. But after that, we should be able to meet the demand for Reels and achieve profitability.In our broader advertising business, we continue to invest in artificial intelligence and see our efforts paying off. In the last quarter, advertising clients' conversion rates increased by more than 20% compared to last year, coupled with a decrease in acquisition costs, resulting in higher advertising return on investment.

We remain excited about the monetization of Business Messaging. Facebook and Instagram are the two pillars of our business, and in the next few years, we hope to make Messaging Online the next pillar. One way is click-to-message advertising (currently with annual revenue of about $10 billion), and the other is paid messaging. We will continue to support more businesses to the business platform, where they can answer customer questions, send updates and sell directly in chats. For example, Air France has started sharing boarding passes and other information in 22 countries and four languages using WhatsApp, and the fact is that they have achieved better results on WhatsApp than on other channels.

Generative AI (Generative AI/AIGC) is a very exciting new field with many different applications. Meta's goal is to be the leader of generative AI, building on research and leading work in recommender AI.

(2) Metaverse

Regarding Mixed Reality (MR), we delivered Quest Pro at the end of last year. It is the first mainstream MR device, and we are committed to setting standards for this industry using our technology. The value of MR is to provide a better social experience than on a phone, enabling you to experience the immersion and presence of VR while remaining grounded in the physical world around you.

Developers have already created some impressive new experiences in the field, such as Nanome for 3D molecule modeling and drug development, Arkio for architects and designers to build interior environments, and others. The MR ecosystem is still relatively new, but it will have significant development in the coming years. Later this year, we will launch our next-generation consumer headphones with MR capabilities. I hope this will establish this technology as the foundation for all future headphones and ultimately, AR glasses.

Outside of MR, the broader VR ecosystem continues to grow. Currently, more than 200 applications on our VR devices have generated revenue exceeding $1 million; there has also been continuous progress in Avatar. Last quarter, we launched Avatars on WhatsApp, and more than 100 million people have created Avatars in the app. Among these people, about one-fifth use their Avatars as their WhatsApp profile pictures. This example demonstrates the trend of combining application families and the metaverse. Even though the majority of our Reality Labs investment is for future computing platforms (glasses, helmets, and operation software), most people will experience Metaverse for the first time on their mobile phones as technology progresses.And begin building their digital identities in our app.

Guidance for Next Quarter and Full Year Performance

Revenue: We expect total revenue for the first quarter of 2023 to be between $26 and $28.5 billion. We assume current exchange rates will have a negative impact of approximately 2% on the year-over-year growth of total revenue in the first quarter.

Expenses: We predict total expenses for the full year of 2023 to be between $89 and $95 billion, lower than our previous outlook of $94 to $100 billion as expected growth in wage expenditures and revenue costs slows. We now anticipate recording restructuring costs of an estimated $1 billion in 2023 related to integrating our office facilities, down from our previous estimate of $2 billion as we recorded some costs in the fourth quarter of 2022. As we make further progress in improving our efficiency, we may incur additional restructuring costs.

Capital spending: We expect capital spending to be between $30 and $33 billion, down from our previous estimate of $34 to $37 billion. The lowered guidance reflects new plans to lower spending on data center construction in 2023 as we shift to a new data center architecture that is more cost-effective and supports AI and non-AI workloads. In large part, all of our capital spending continues to support our suite of apps.

Tax-related. If there are no changes to US tax law, we expect our tax rate for the full year 2023 to be below 20%. Additionally, as mentioned on prior conference calls, we continue to monitor the feasibility of cross-Atlantic data transfers and their potential impact on our European operations.

II. Analyst Q&A

Q1: Regarding generative AI, how do you view its potential use cases for both users and advertisers? When do you expect to see it promoted at the platform level?

A1: Regarding generative AI (specifically applications like AIGC), I think this is a very exciting space. The two biggest themes that we're focused on this year are operational efficiency and then new businesses - generative AI being one of the most interesting challenges here in terms of how we make it scale and make it more efficient so that we can bring it to a wider audience of creators to increase their productivity and creativity within our app. I think once we do that, there will be a lot of very exciting use cases that arise.

We have a lot of different things coming out this year and we'll be sharing updates on them. I do hope this business can develop quickly and we can quickly ascertain who's right and who's wrong.

Q2: Regarding the new data center architecture, how do you view the long-term capital spending intensity for that business? How much long-term benefit and overall cash flow gain can that business bring?A2: The new data center architecture we discussed has led to lower capital expenditure guidance for the future. We are moving our data centers to a new architecture that can better support both AI and non-AI workloads. As we learn more about our AI needs over time, we will have more options available to us. Additionally, we expect the new design to be cheaper and quicker to build than the previous data center architecture. With the new data center architecture, we will optimize our approach to building data centers. Therefore, we have adopted a new phased approach that allows us to make basic plans with less initial capacity and less initial capital expenditure. However, we can quickly increase future capacity if needed, and we are still planning to significantly enhance our artificial intelligence capabilities.

Regarding long-term capital expenditures, we have lowered our capital expenditure guidance for 2023, and the lowered guidance can reduce the proportion of capital expenditure to total revenue, which is also an issue that we will continue to monitor.

The current surge in capital expenditure is actually due to the construction of AI infrastructure, which began last year and continued this year. We will measure the return on investment of these AI investments, and their returns will continue to affect our future spending plans. Our goal is still to reduce the proportion of capital expenditure to revenue, but short-term capital expenditures will depend to some extent on the revenue outlook and our need to further build AI capabilities in the future.

Q3: How does the company plan to monetize Reels? Can you share some use cases for advertisers to use it to deliver their brand and direct messaging? Can you quantify the difference in monetization efficiency between Reels and other short video platforms, and how we plan to narrow the gap over the next few years?

A3: For these consumer products, building and expanding products is somewhat different from making profits. Therefore, building these new products is a very difficult problem. So let's start by doing something useful for people. Once we have hundreds of millions or billions of people using it, we will focus on improving monetization, which has always been a useful formula for us. This is a general approach, and we do have a lot of people using Reels now.

So I think the question at this point is whether it is more strategically valuable to expand its scale of use now than to monetize it immediately. I don't think there is a clear answer to this at the current scale. But I think that, given our current user base, the correct approach now is to improve profitability efficiency.

We know that users now have a demand to see more Reels. Therefore, I am confident in the trend of Reels naturally increasing monetization. Reels' monetization rate on Facebook has doubled in the past six months, and I think we can continue some of these trends. Then we will naturally unlock the ability to display more and more Reels. You can expect us to continue doing this in the future, including new generative artificial intelligence products and some new Metaverse content we are working on where we will use the same methodology.Regarding the advertising industry's response to Reels. We continue to work towards getting more advertisers involved in Reels and have made good progress. Currently, more than 40% of advertisers in our application use Reels. Both performance and brand advertising have improved, but performance advertising remains the focus of advertisers. For example, a German fashion retailer, Outlook City, developed creative campaigns specifically for Reels to test the impact on conversions. They found that Reels increased their ROI by 19 times, reduced their cost per purchase by 89%, and increased their sales by 9 times. Overall, they achieved very good results.

We have not quantified the monetization gap between Reels and other services. It took us several years to narrow the monetization gap between Stories and Feed Ads in the past, but we expect it will take longer for Reels. That being said, we still expect to reduce the negative impact on Reels revenue to neutral by the end of this year or early next year. We plan to achieve this goal by improving Reels' monetization efficiency and increasing user interaction.

Q4: How do you consider the negative impact of IDFA on the company's business?

A4: We are continuously mitigating the impact of ATT changes, but this is the reality of the online advertising environment we operate in today. Therefore, we will continue to build tools that alleviate the impact of these changes, and we see these tools being widely adopted. We also invest in methods that bring conversions online, and we have many ad formats that have been instrumental in driving click-through and lead generation. In the long term, we will continue to invest in our privacy-enhancing technologies to fundamentally enable us to offer advertisers better performing and privacy-safe ads.

Regarding our strategy: First, continue to invest in AI, which is where we see significant improvements in ad relevance. As Mark mentioned, we've seen a more than 20% increase in ad impressions year-on-year, coupled with a decrease in cost per acquisition. We also apply it to advertisers' automation experience, improving measurement standards, and enabling advertisers to make better decisions. Second, bring more effective conversions to our website.

Q5: With the efficiency year coming up, how do you consider the long-term investment in Reality Labs?

Regarding the investment consideration for Reality Labs, I think two analysis approaches are applicable. First, I think it's important not to view Reality Labs as one thing. It can be broken down into three main areas, with long-term AR work being the biggest area, but there are still significant research challenges there, and there's a lot of work we haven't really started to tackle.

VR is on the rise, and I think Quest 2 is doing well. Our Quest Pro has multiple product lines. The smallest scale for now is Metaverse software, which has not yet fully demonstrated its importance.I believe that software and social platforms may be the most critical parts of what we are doing, but the level of capital expenditure on software is much lower than that on hardware. We are doing many different things in every field, so just like any project we are running, we constantly look at the performance of the products we deliver, the overall development of the market, the performance of competitors, and their reactions to different things, as well as the experiments we are conducting.

We are constantly adjusting our roadmap. Obviously, some of them are long-term things. So far, I have not seen any signals that we should change the long-term strategy of Reality Lab. But we will constantly adjust the details and consider how to execute them. So, I think we will definitely regard it as part of the efficiency work in progress.

Another part is just different strategies, such as trying to flatten the organization, etc., which will be applicable to the entire company. Therefore, we hope that in the roadmap we are trying to implement, whether it is in Reality Labs or Family of Apps, we only want to focus on improving the efficiency of all our work. Many times, when people talk about efficiency, many people focus on priorities and which major projects can be cut. However, I actually think that over time, what makes you a better company is being able to execute and do more things, because your operational efficiency is higher, and you can complete work with fewer resources, so I want us to enter this mode more. Our company has developed so fast in the first 18 years. When a company grows rapidly, it is difficult to really improve efficiency. And I think we are now in a different environment, and we have many areas to explore and develop further, so I think the correct strategy is to focus on revenue growth.

Q5: Can you share some areas where efficiency can still be improved? How should we view the level of recruitment in 2023 under the new spending guidelines?

A5: The 2023 spending guidelines mainly involve reducing expenditures in three areas:

1. The first is the slowing wage growth. We will continue to review how we allocate resources throughout the company. We are currently freezing HC across the board, and we expect the recruitment rate to continue to slow down this year as we assess the positions we will be opening. Reducing expenditures is a continuing process for us, and we currently have no end-of-year recruitment targets to share.

2. The second is cost reduction. We expect growth to slow down because in Q4, we extended the depreciation life of non-AI servers.

3. The third component is that our current guidance reflects an estimated $1 billion in equipment integration costs, which is lower than the $2 billion expected in the previous quarter.

Q6: How should we address the issue in Europe about Meta's use of first-party data for advertising and how should the company consider the risks involved?

A6: If you are referring to the ruling of the EU DPC, which requires us to change our practice of relying on contract necessity as a legal basis for European advertising, we disagree.We believe that our current approach complies with GDPR and we are appealing the substance of the rulings and fines. We expect these decisions will not affect our ability to provide personalized ads in the EU, and advertisers should be able to continue using our platform to reach customers and grow their businesses.

Q7: In terms of privacy and data usage, is Q1 24 still expected to be unfavorable? Has the impact of IDFA or ATT passed? Looking ahead to the next year, what progress will be made on the Digital Markets Act in Android or the EU?

A7: ATT still has a negative impact on revenue. We are adapting and reducing this part of the impact by using different advertising client tools, offline conversion ads, and long-term investments in AI technology for privacy protection. As for Android and other adverse factors, it is too early to know where the Act will land. I think Google is taking a cooperative approach with the industry, which is crucial, and we will keep up-to-date.

Regarding the EU regulatory issue, we have been preparing for some time to comply with DSA and meet the compliance deadline expected to take effect this year. These are significant but manageable cost items, which are already included in our total expense guidance.

Q8: In terms of efficiency, what is the relationship between revenue growth and operating expenses and capital expenses? Will there be a compression of profit margin after 2023? What is the attitude towards the overall economy after Q4?

A8: Lower expense and capital expenditure prospects put us in a better position in our financial performance this year. I believe that we will still focus on the goal described by Mark in the previous quarter, which is to achieve comprehensive revenue growth while actively promoting the development of future technology. This still remains the principle that guides us in formulating our financial plans.

Attitude towards the economic environment: The situation of the fourth quarter shopping season is basically within our expectations. For us, the trend of online commerce has slightly improved, which is encouraging, but the year-over-year growth is still negative. So, the global economy is still in a very unstable macro environment, and it is too early to predict how 2023 will develop.

Q9: Can you talk about the signals Meta sees from the AI discovery engine (i.e., content algorithm distribution) of Reels, which brings users more relevant content they expect to see, creating this more relevant content between Reels and Story to Feed, even Messenger?

A9: Overall, we have gained a lot from the AI discovery engine, mainly used to achieve the integration of all different content types on Facebook and Instagram through AI recommendation systems. Reels is a special case because it grows the fastest, but there is no specific data type that is useful. We have seen great improvements in time and engagement we have gained, but it is difficult to predict how much we will be able to adjust and improve. Based on our experience so far, we believe there is still significant room for growth.This applies to Reels and other discovery engines.

We did spend most of our time talking about Reels, but the rest of AI discovery engines have done well, and they contribute more to other businesses, because if people can eventually discover more photos, links, groups, or similar things on Facebook or just find interesting content on Instagram, they are just more invested in the product, and we know how to profit from this content. This is ultimately very helpful for overall engagement and profitability.

We expect the percentage of recommended content in the total video volume on Facebook and Instagram to continue to grow. It may reach about 30% or even 40% by the end of this year, and it will continue to grow, because we can find content that people like, interested in, but may not originate from accounts they directly follow.

Q10: How does Reality Labs see the 23-year profit situation? Will the loss peak this year? How is the current progress of store advertising?

A10: Regarding RealityLabs, losses are expected to increase in 2023. Given the significant long-term opportunities in this field, we will continue to make meaningful investments in this area. This long-term investment is supported by the need to drive overall operating profit growth as we make these investments.

In the fourth quarter, we continued to test store ads in the US, improving store performance by guiding consumers to where they are most likely to buy. The overall product is very early, but we see market adaptability in testing. We saw triple-digit growth in revenue and adoption in the fourth quarter, and we expect this growth to normalize in 2023. The revenue of the store ad beta is in the hundreds of millions of dollars. Currently, this revenue scale is relatively small, but it is growing rapidly.

Q11: Currently, click-to-message revenue has reached 10 billion US dollars. How do you view its future growth path and development prospects?

A11: As it stands, click-to-message ads are one of our fastest-growing ad products, with over half of click-to-message advertisers exclusively using click-to-message ads on our platform. Ways to scale:

1. On the demand side, making it easier for more businesses to adopt click-to-message ads by creating more entry points and simplifying the creation process. We are working hard to integrate with partners who can help small businesses expand.

2. In terms of product performance, continue to focus on improving the return on investment that advertisers get from click-to-message ads, enabling them to do more offline optimization, create better experiences, simplify the process, and help them drive conversions.

3. In terms of supply, we aim to expand the commercial communication ecosystem by providing people and businesses with more ways to connect messaging applications, such as direct search on WhatsApp. We have invested heavily and seen very healthy growth.

Q12. Regarding efficiency management, it has been reflected since last year, how to understand this year as the "year of efficiency"?

A12: Regarding the issue of improving efficiency, we have the following considerations:

1. Transitional needs for the company's development style (the end of high growth). Just like the company's development history, in the first 18 years, we grew at a rate of 20%, 30%, or even higher every year. However, this changed greatly in 2022. For the first time in history, revenue declined negatively last year, which was a significant drop. We do not expect this situation to continue, but we do not believe that it will necessarily return to the previous manner. This forces us to step back and consider that we cannot treat everything as super-fast growth. Some business areas are growing very quickly, but now there are also some businesses that only have a lot of people using them, and we should operate them in a different way.

2. Make the company better. I am actually quite optimistic and have a good roadmap to make us more productive and able to build what we want. This may not reduce costs, such as focusing on artificial intelligence tools to help increase the productivity of engineers. In the long run, this may also enable us to employ fewer people. But I believe that reducing the number of management levels can help information flow better in the company and make decisions faster. Ultimately, this will help us not only produce better products but also attract and retain the best people who want to work in a rapidly developing environment.

As we began to delve into this issue, the company began to become better. I don't know how long this will last, but I do believe that we have a lot to do. I want to continue to emphasize these dual goals: making the company a better technology company and improving profitability. But I also think that focusing on the first goal is very important (making it a better company). Even if our performance exceeds our business goals this year, we will stick to it because in the long run, this will make us a better company.

Q13: What is the potential expectation of market conditions in 1Q23 guidance? What is the assumption behind this guidance? How to consider the seasonal cycle factors?

A13: The 1Q23 guidance is between 26 billion and 28.5 billion US dollars, YoY -7% to 2%. Based on the current overall macro environment, this figure reflects overall uncertainty. It is expected that the adverse impact of foreign exchange on YoY growth will decrease compared to the fourth quarter. We will continue to closely monitor advertising demand and ongoing macroeconomic instability.

Risk disclosure and statement of this article: Dolphin Research Disclaimer and General Disclosure