The following $ Google-A.US 22Q4 earnings conference call summary, financial report interpretation can be reviewed in《 Short-term pressure is not small, Google needs to learn from Meta 》
I. CEO Remarks
After the surge in digital spending during the epidemic, the macro economy has become more challenging. We continue to have outstanding business and provide very valuable services to our partners. For example, during the World Cup final on December 18th, Google search reached its highest ever rate of queries per second.
In addition to our advertising business, we have strong momentum in cloud services, YouTube subscriptions, and hardware. However, our revenue this quarter was impacted by advertising client spending adjustments and foreign exchange effects.
Today I will focus more specifically on two main things, and then I will provide a brief quarter-by-quarter outlook for our business than usual.
1. How we leverage artificial intelligence to bring incredible opportunities to consumers, our partners, and our businesses.
Artificial intelligence is the most profound technology we are researching today. Our talented researchers, infrastructure, and technology position us favorably in the field of artificial intelligence at this inflection point. In the coming months, you will see us discovering opportunities in three major areas--
2. Modeling. We have published a large number of articles on LaMDA and PaLM, which are the largest and most complex models in the industry. Coupled with a lot of work in the coming weeks and months, we will start offering these language models from LaMDA, so that people can interact with them directly. This will help us continue to receive feedback, test and improve them safely. These models are particularly amazing in composition, construction, and summarization. When they provide the latest and more accurate information, they will become more useful to people. In search, language models such as BERT and MUM have undergone four years of search result improvements, resulting in significant ranking improvements and multi-threaded search like Google networks. Soon, people will be able to interact directly with our latest and most powerful language models, serving as a companion for searching in experimental and innovation bases.
- We will provide new tools and APIs for developers, creators, and partners. This will enable them to innovate and build their own applications and discover new possibilities for artificial intelligence. In addition to our language models and other artificial intelligence models, our artificial intelligence is a powerful driver for enterprises and organizations of all sizes.
Google Cloud is providing our customers with our leading technological position in AI through our cloud AI platform, including infrastructure and tools for developers and data scientists, such as vertex AI. We also provide specific artificial intelligence solutions for industries such as manufacturing, life sciences, and retail, and will continue to launch more workspaces. Users benefit from AI-powered features, such as intelligent canvases for collaboration and intelligent writing tools for creation - and we're striving to introduce LLM (large language models) into email and documents. We will also offer other practical productivity tools, from coding to design and beyond.
Moreover, AI continues to dramatically enhance Google's other products. We'll continue partnering with companies outside of Google - including joint research and collaboration to apply AI to society's biggest challenges and opportunities. For example, DeepMind's protein database - which contains all known 200 million proteins - is now being used by one million biologists globally. We're continuing to invest comprehensively in AI - Google AI and DeepMind are an essential part of AI's future. In recent years, DeepMind has increasingly worked between teams within Google and other departments. To reflect this progress, Ruth will share more about our financials in our comments section.
4. How we focus investments and make efficient decisions as a company.
We're committed to investing responsibly and identifying areas where we can operate more cost-effectively. We're focused on building financially sustainable, vibrant, and growing businesses across the company in a deliberate and measured way. For example, we're working to improve economies of scale and hardware, as we focus on the Pixel line of business and our overall cost structure. Cloud computing remains very focused on its path to profitability, and we have many opportunities to build on progress made in years of YouTube, beginning with short-term earnings.
Overall, I believe this is an important journey of reimagining company costs over a sustained period of time. Several aspects are already underway, including priorities for investments we're making across the company, our hiring needs, our technology infrastructure efficiency, and improvements to the productivity of our AI tools. We've conducted a rigorous review of products and functions to ensure employee roles are aligned with our company's top priorities. We announced layoffs and thank departing Google employees for their contributions and hard work.
II. A Quick Review of Business
Just this week, we launched revenue sharing for YouTube videos and now have an average of over 50 billion daily views, up from the 30 billion announced on our Q1 2022 earnings call. This will reward creators and improve everyone's video watching experience.
Our subscription business continues to grow, but YouTube Music and Premium had more than 80 million subscribers (including trial). Along with subscriptions to YouTube TV's YouTube Gold channels, we're doing well in this area. YouTube's NFL Sunday tickets will help drive subscriptions, bringing new audiences to YouTube's paid and ad-supported experiences, creating new opportunities for creators to accelerate this goal. On the hardware side, many media and commentators have rated Pixel 7 Pro as the phone of the year. The fantastic features like magic eraser, photo, and blur are unbelievable and help differentiate pixel from other phones. Pixel 6, 7, and 7 Pro of 2022 are the best-selling generations of phones that we have launched, and we have gained shares in every market operated this year.
As for Google Cloud, we saw a 32% revenue growth in Q4. Differentiated products and a focused market strategy continue to drive customer momentum. Starting from real-time data analysis and AI, more and more customers are choosing BigQuery because we unify data drop, data warehouse, and advanced AIML into one system, and now analyzing over 110TB of data per second.
Regarding infrastructure, our global network and advanced TPU v4 for AI supercomputing helps Snap triple its throughput of the critical business advertising ranking workload while significantly reducing costs. Our machine learning infrastructure has Cloud TPU v4 Pods that can run large-scale training workloads, 80% faster than alternative solutions benchmarked by third parties. Our reliability advantage and open-edge cloud provide critical support for Deutsche Telekom's 5G network.
As I mentioned, our AIML solutions across vertical industries are a key differentiator. We help the Rich State Bank automate the customer service experience for mobile users and continuously improve their customer experience. In 2022, mandian, which we are now integrating, has more than 1,800 customers preparing for or recording the most critical cybersecurity incidents.
Internally, the innovations mentioned above help drive new victories and expansion across regions. In other businesses, we also focus on sustainable investing throughout the portfolio and building good businesses. Verily, for example, recently adjusted its strategy and structure to make its product development focus clearer.
3. CBO Speech
We start with Google's service performance in Q4 and dig deeper into our areas of strength.
Google's service revenue was $68 billion, down 2% YoY, with significant negative foreign exchange effects. In terms of Google advertising, search and other revenues declined 2% YoY, while YouTube advertising and network revenue saw high single-digit declines. Google's other revenues grew 8% YoY, with strong growth in YouTube, non-advertising, and hardware revenues but a drop in gaming revenues.
I will focus on two additional factors that affected our Q4 advertising business. In terms of search and other revenues, we saw moderate growth YoY (excluding the impact of foreign exchange), reflecting growth in retail and travel, partially offset by declines in finance. At the same time, we saw some advertisers reduce their spending on search in Q4 compared to Q3. The YoY revenue decline is due to the broadening of the Q4 advertising customer spending rebound. Now, I'm going to narrow down and share with you our investment areas and see clear opportunities for long-term growth.
1. Google AI. It's important to recognize that our advertising business has clearly benefited from the transition to mobile over the past decade. Recently, our advertising revenue has grown significantly during the pandemic, with ad revenue in 2022 up $90 billion from 2019. Looking ahead, we're focused on further driving this high base through AI-driven innovation. Breakthroughs in every aspect, from natural language understanding to generative AI, are pushing our ability to deliver results, which brings meaningful performance to advertisers that is practical for them. For example, we use intelligent bidding and AI to predict the future conversion and value of ads, helping businesses stay agile and respond quickly to changing demands. In 2022, advances in AI have improved bidding efficiency, allowing us to increase ROI and use budgets more effectively.
In search query matching, large language models like MUM match advertisers to users. Combining this understanding of human language intent with prepayment and bid prediction is why, in using target cost per conversion ad campaigns, when merchants upgrade from exact match keywords to broad match, they can see an average of 35 or more conversions.
Google AI is also the basis of our creative products, such as text suggestions in Google Ads, as well as ad asset optimization and adaptive search ads. We're excited to start testing our auto-generated asset resources better, and once advertisers opt-in, AI can seamlessly generate ad headlines and descriptions for search creatives. And with maximum performance, it offers our clients the best combination of our AI components. But we're not stopping there. AI has been the backbone of our advertising business for the past decade and will continue to bring cutting-edge advancements to our products to help businesses and users.
2. Retail. We have three pillars that underpin long-term value.--
(1) We have been on a mission for many years to make Google the centerpiece of consumers' shopping journeys and a valuable place for merchants to connect with users. This means constantly improving our consumer experience, starting with more intuitive, immersive, and browsable searches.
(2) We're engaging more merchants in our non-paid product details and ad experiences. In 2022, we've seen an increase in the number of merchants, particularly small and medium businesses, and product inventory on Google. Adding more value to merchants remains a top priority.
(3) To further drive retail performance, we're focused on excellent ad products - from auto insights to bidding tools, as well as omnichannel solutions to marketing campaigns for Pmax (maximum effectiveness ads). We're helping retailers achieve their goals and connecting with them when customers shop. Two quick insights on Pmax: first, merchants who upgrade from SSC to Pmax see an average of a 12% lift. Second, Pmax is a success story during the holiday and Cyber 5 periods. In the turbulent retail season, it has been able to expand and adapt to constantly changing traffic, prompting many retailers (especially mid-market advertisers) to achieve strong performance.
Regarding YouTube, although fourth-quarter revenue continued to decline, we are confident in YouTube's long-term development trajectory. The creator ecosystem is the lifeblood of YouTube. In 2022, more people are creating content on YouTube than ever before - long-form, short-form, audio, podcasts, music, live streaming. What sets YouTube apart is that we provide creators with more ways to create content and connect with fans, as well as more ways to make money. On any other platform, more creators mean more content, which means more viewers, and this brings more opportunities for advertisers.
Four, YouTube Investment Direction.
The creator ecosystem and our diverse development strategy will continue to drive YouTube's long-term growth. To support this growth, we focus on:
(1) Increasing short videos. The viewing rate of short videos is rapidly increasing, with more than 50 billion views per day. We are also still pleased with our continued progress in early profitability. On the creator side, seeing creators using short videos to introduce their content and innovative ways to expand existing channels is impressive. We focus on providing creators with the best content creation and monetization tools, richer new features and analysis capabilities, and helping them personalize and optimize content strategies. It is still early for short videos, but we are confident that this road is long.
(2) Accelerating Big Screen Engagement. More and more users are watching their favorite creators on the big screens in their homes. According to Nielsen, YouTube is the leader in streaming viewing time, and advertisers are adopting AI solutions. We are helping brands increase coverage and return on investment, and addressing pain points such as frequency and measurement.
(3) Investing in our subscription products. Then there are our subscription products. Obviously, the future of online video is to help users seamlessly discover and watch content that supports and provides high-quality services with ads. Our goal is to become a one-stop shop for multiple types of video content. That's why we first offer YouTube Music and Premium, where more than 80 million paid subscribers and fans can enjoy their favorite content and music ads for free. Then we expanded into YouTube TV, significantly improving the traditional TV experience. Last fall, the prime time channel was launched, offering streaming subscription services on YouTube on a per-item charge basis. Given the potential we see in subscription products, we recently announced a multi-year agreement on NFL Sunday tickets, and we are excited about the opportunities this will bring.
(4) Long-term efforts to make YouTube more shopping-oriented. YouTube Shopping is still in its infancy, but we see many potential opportunities to make it easier for users to shop from their favorite creators, brands, and content.
CFO Speech - 2024 Spending Outlook
I would like to share with you the key elements of our efforts to provide a sustainable cost base for our operations and to slow down the pace of expense growth. We expect this impact to become more evident in 2024.
Firstly, in terms of the number of employees in the company, we are meaningfully slowing down the pace of recruitment for 2023 while still investing in priority areas. In Q4, we added 3455 new employees. As in previous quarters, the vast majority of hires were for technical positions. Regarding our recent announcement to lay off about 12,000 employees, the majority of the impact will be seen in 1Q23. We will incur severance costs of approximately 1.9 to 2.3 billion, which will be reflected in the company's cost next quarter. We will continue to recruit in priority areas, with a special focus on top engineering and technical talent and on the global footprint of our workforce.
Secondly, we are making longer-term efforts to redesign our cost bases in three major categories.
(1) Using artificial intelligence and automation to increase the productivity of operational tasks across the company and the efficiency of our technical infrastructure.
(2) Managing our spending with suppliers more effectively.
(3) Optimizing our ways of working and locations.
In 1Q23, we expect to reduce about 500 million of expenses related to existing leases to align our office space with our adjusted global workforce size. This will be reflected in the company's cost, and we will continue to optimize our real estate footprint.
In 2023, we expect capital expenditure to be basically consistent with 2022. There will be an increase in investment in technical infrastructure and a decrease in investment in office facilities. Our continued investment in technical infrastructure is obviously a key part of supporting our long-term growth opportunities.
We have noticed that we have adjusted the estimated service life of servers and some network equipment starting from 1Q23. We expect these changes to have an impact of about 3.4 billion on our 2023 operating performance in terms of assets and services (reducing depreciation costs).
Significant cost adjustments will be visible in 2024.
Q&A
Q1: How should we consider the potential impact of AI on capital expenditures? Will the higher computing power of these tools potentially affect profit margins in the coming years?
A1: Regarding the question of AI and capital expenditures, as both of the previous speakers have pointed out, AI has already been incorporated into many of our products such as performance, performance maximization, intelligent bidding, and cloud services. AI has higher computing power, but it has also opened up more services and products for our users, creators, and advertisers. Having said that, we are very focused on further optimizing computational costs. We will continue to invest with a keen eye on the return on these capital expenditures. As I mentioned in my opening remarks, we expect capital expenditures to stay flat in 2023 compared to 2022, but there will be significant structural changes (less investment in office facilities and more in infrastructure). We are increasing investment in technical infrastructure, and this is not just for AI. To support cross-company investments, especially in cloud-related investments, we are reducing capital expenditures on office facilities.
Q2: A discussion of long-term efforts to improve efficiency. How should we view the potential impact of these efforts? Has the type of savings in the income statement been calculated for this period?
A2: Regarding overall efficiency opportunities, I am very focused on three areas I have noticed, one key factor being the use of AI and automation to improve the productivity and efficiency of our technology infrastructure. We want to focus on persistently improving our cost basis. If you browse through the work we are doing, you will find that they take longer to implement and execute. They are now in progress and will continue to provide additional upside over time. That's why I said that you will see a greater impact on 2024 than 2023, but we are continuing to address these issues.
Q3: You have both talked about NFL and the opportunities it presents. Can you define what you see as the long-term opportunity? Why is having NFL so important?
A3: We believe that there are many great opportunities to differentiate the user and creator experience by leveraging our unique abilities. This basically means that every YouTube audience interested in the NFL can now access full Sunday Ticket products with just one click, as an add-on to YouTube TV subscriptions and independent products on premium channels.
This will be the first time Sunday Ticket has really offered fans ordering in Los Angeles. For example, we are building a multi-screen viewing feature for subscribers on YouTube TV. On YouTube CTV, we will add new features specific to the Sunday Ticket experience, such as commenting, chatting, voting and more. On the creator side, imagine all the innovative ways they can create exclusive NFL content, behind-the-scenes activities, and more. We're really excited to see what they do in long-form, short-live, and beyond.
Q4: Regarding the profitability of YouTube short videos, what is the stickiness factor? What will really attract advertisers as time goes on? What problems have been seen and what has been solved to make it a faster profit product?
A4: In terms of short videos, as I said before, viewership is rapidly growing and we are still satisfied with our sustained growth in monetization. It is a big project for us to narrow the monetization gap between short and long videos, and of course, we are very focused on continuing to build a great creator and user experience. Ads on short videos are now available, and we have provided some progress signals for you before. Video actions, app discovery, and maximizing the effectiveness of ad campaigns can be purchased for short videos.
Again, we are the only platform where creators can make various types of content, spanning multiple formats, across multiple screens, and truly make a living in multiple ways. Just yesterday, we brought revenue sharing for short videos to the YouTube Partner Program. Our ultimate goal is to make YouTube the best place for short videos and creators, which is exactly what we're focused on now.
Q5: Can you talk about how you would push AI products to the market and do so without sacrificing quality or audience trust?
A5: This is an exciting time for AI. I think we've been investing for some time and it's clear the market is ready and consumers are interested in trying new experiences. I'm satisfied with all the investments we've made in ensuring we can develop AI responsibly.
In certain cases, we'll launch beta features and slowly scale them up. Obviously, we want to make sure we iterate under public feedback. This field is rapidly changing and the cost of services needs improvement. So I think this is a very early stage, but we're committed to providing both experiences, either in new products or experiences, bringing direct large language models into the research, providing APIs for developers and businesses to learn from - something we've always done, and that's something to look forward to.
Q6: "Redesigning cost structure in a persistent way" and everything that goes with it is different from what we've heard in the past. Can you quantify your thoughts on these efforts?
A6: When we talk about focusing on providing sustainable financial value, it clearly means cost growth cannot precede revenue growth. We're focused on the upside of revenue as well as on the persistent changes in cost and space that truly secure our ability to invest in such growth, which is obviously the focus or revenue growth.
In Google's services, we have a lot of exciting things in front of us. All advances in AI are improving the return on investment for advertising customers and the search user experience.
More broadly, as we've discussed in key product areas, we have a very firm commitment to cost growth - persistently redesigning our cost basis, which will benefit all alphabet departments and key businesses. Slowing down hiring is a starting point, and product priorities at Google are key. For example, when we focus on pixel and cost structure, we increase hardware efficiency and then use AI and automation to increase the productivity of operational tasks and the efficiency of our tech infrastructure.
We have many workflows. As I said, managing our expenses with vendors and suppliers, and then optimizing workflows and locations. Part of the reason is real estate consolidation because employee growth has slowed. All of this benefits not only Google's services, but also many of the services that improve company efficiency. So we're still very focused on profit, revenue, and profit-driven factors. In other businesses, we also focus on sustainable investment. So, this persistent change is underway, you may see the impact in 2023, but it may not be until 2024 that you really see the benefits of full-year operating efficiency and this change. Q7: Regarding the theme of artificial intelligence, considering all of the investments made in the past 5 years, we have been talking about a lot of opportunities without focusing on the possibility of potentially disrupting existing products' user experience or revenue models. How do you philosophically approach seizing opportunities while finding the right balance between opportunity and self-destruction?
A7: Firstly, we have created such a foundational technology that we have invested not only in research but have actually been ready for full-scale production. We have already deployed models like BERT and MUM and if you look at their impact on search quality (enabling multiple models in search, driving the use of Google Lens, and more), I think we have done a very good job scaling Vertex AI in Google Cloud today.
We have brought AI APIs into enterprises, and they are on a very healthy growth trajectory. Therefore, we do see long-term opportunities in both rolling out these APIs and ensuring that every developer and organization in the world can use them. As I said earlier, we are still in very early stages, and I think there is a lot of headroom in search as well.
We can now integrate more direct experiences with LLMs (large language models) into search, and I think that will help us expand into and serve new types of use cases. So, I see this as an opportunity to rethink and reimagine what drives search in service of solving more use cases for our users. It is still early stages, so you will see us be bold and put things out there, get feedback, iterate, and make it better.
Q8: Regarding the question and debate on cost structure, how should we think about rationalizing cost structures in the medium to long term and making the costs fit the opportunities of certain departments within the company?
A8: With regard to investment levels in other businesses, as we have discussed in previous earnings calls, our goal for Other Bets is to use our deep technology investments to drive innovation with true value creation potential. At the same time, we are very focused on overall investment pacing and financial returns. What we really see here is where are the opportunities for profitability and commercialization? There is no single metric for the entire investment portfolio, but we are very focused on what the pacing of investments is and what the opportunities for profitability and commercialization are.
With regard to DeepMind, one thing is very clear, and that is we will only integrate it when it is an effort to support the products and services within the company. You see us do this already, as for example when we integrated Chronicle with our cloud business and the AI proved to be effective for network security products that are now in the cloud, so we do see huge opportunities. DeepMind's research is the future core and spans across product areas of the entire company portfolio. So, this shift in reporting reflects DeepMind's strategic focus of supporting every department's business internally (and will be spun out of Other Bets as a cost item next quarter). This is why I pointed out that DeepMind's finances will report in our company's cost department from the first quarter.
Q9: Regarding the low single-digit YoY growth rate for search advertising under a fixed exchange rate, could you elaborate on the growth pressures for search advertising volume and price? What caused the decline in growth rate? The negative growth is similar to the recession level of 2009. Are there any signs of approaching the bottom or stability?
A9: So overall, as we've indicated, we're very excited about everything we're doing in search, and it's practical for all of us. That's why you hear so many comments about the application of AI and what that means for continued opportunities. You have a lot of different questions in there. I think one is around quantities.
In our upcoming annual report, you'll see that the single click cost for 2022 decreased by 1% YoY, as we discussed in previous quarters, the change in cost per click can reflect many different factors. For example, the click volume in 2022 increased by 10% YoY, which reflects various factors, including increased interaction on mobile devices and advertising format improvements.
I think overall, the most important thing is the technological innovation and our cost restructuring that truly allows us to achieve this long-term, sustainable value creation.
Q10: The hardware business is related to the company's overall AI plan. Could you talk about the importance of having a strategy with a wide range of hardware products?
A10: First of all, we're very pleased with how Pixel is performing in a challenging macro environment. I think our part combination process is very important. This allows us to invest and push innovation forward. If you think of it through an overall ecosystem lens, the user engagement is surpassing the phone itself. For example, we're involved in Pixel Watch, integrating Fitbit as part of the strategy and bringing it into our ecosystem, as well as working closely with Samsung on the operating system. These business combinations are driven by a 300% increase in activity on the Android observer ecosystem.
So we're very thoughtful about how we approach deepening our hardware presence. As Ruth mentioned, in all these areas, we're also working to drive our investment portfolio to be more focused and cost-effective. Clearly, we're working in a challenging supply chain environment and under demanding customer challenges. So we will continue to focus on improving all these (cost-effectiveness) in a sustained way.
Q11: Regarding the profitability of Google Cloud, is there anything to share?
A11: As we've discussed on our Google Cloud conference call before, given the overall opportunity and our desire to ensure we're positioned and our teams have the continued momentum, we need to make meaningful investments to be ready to support customers in all the different market segments globally. We're very focused on the path to profitability and some of the projects I mentioned that are beneficial to Alphabet are definitely related to cloud computing. We're very excited about all the efforts they're making in terms of efficiency in our technical infrastructure.
With regard to the profitability progress of Google Cloud, where we are now is that we have significantly narrowed the gap with profitability, but we are still working on it (because our investment continues to grow).
Risk disclosure and statement for this article: Dolphin Analyst Disclaimer and General Disclosure
