Youssef Squali: Great. Thank you. Two questions for me, please. First, maybe one for Susan. Just staying on the theme of the efficiency and with all the adjustments to your cost basis lately, can you maybe just speak to the relationship you are trying to build between revenue growth and OpEx, CapEx growth over time? I think the last time we saw them move in tandem or close to each other was back in 2017. So, are you at a point where you would want them to grow much closer to each other, or are we still in an investment mode and therefore, potentially margin compression mode beyond just 2023? And then maybe, Mark, can you just provide an update on kind of the health of the broad digital ad space, especially for the SMB side and DR. Just curious if coming out of Q4, you are incrementally bullish or you are still as cautious as that you are three months or six months ago? Thank you.
Susan Li: Thanks. I am happy to take the first question. So certainly, the lower expense outlook and CapEx outlook puts us in a better position in terms of financial performance for this year. We are I think still €“ we are focused on the goal that Mark outlined, I think last quarter of delivering compounding earnings growth, while enabling aggressively in the future technology. And that continues to be the principle by which we are €“ by which we are guiding our financial plans. The second question is on update €“ oh, on the health of the broader digital ad space. I can take €“ I can start with this and Javi, you should feel free to jump in if you want to add more color. Q4, for us, we saw that the holiday season, for us, fell mostly within our range of expectations.
Trends in online commerce modestly improved for us, which is encouraging, but again, the growth was still negative year-over-year. So, overall, I think this is still a pretty volatile macro environment. It’s early in the year to know how this will shape up for 2023. And if there is anything, Javi, you would like to add, you can jump in.
Javier Olivan: No, I think you covered it well, Susan.
Operator: Your next question will come from the line of Ron Josey with Citi. Your line is open.
Ron Josey: Great. Thanks for taking the question. Mark, I wanted to talk a little bit more around the progress on the AI discovery engine in Reels. And we are seeing it in our own usage in terms of content and categories and just getting more insights and content that we would like to see. But can you just talk about the added signal that Meta is seeing and gaining you to produce this more relevant content across Reels to Stories to Feed and maybe even to Messenger? Thank you.
Mark Zuckerberg: I am not exactly sure what would be useful to share here. But in general, a lot of the gains that we are seeing on the discovery engine overall, which we basically used to refer to our AI recommendation system across Facebook and Instagram across all different content types of which Reels is sort of a special case, and that’s growing the fastest with short-form video. But a lot of this is €“ I mean there is not like one specific data type that’s useful. A lot of the trends that we are seeing here is, we are using larger models, which require more computation. We have shifted the models from being more CPU-based to being GPU-based. We have seen big improvements in the amount of time and engagement that we have gotten.
And we €“ it’s a little bit hard for us to predict exactly how much we will be able to continue tuning those and improving. But from the experience that we have had so far, I would bet that there is still pretty significant upside there. I know that you kind of asked about specific data points, but I think that that’s really the theme that we are seeing. And that applies across Reels and the rest of the discovery engine. The €“ one thing that I would add, it’s a little bit separate from your question, but we do spend most of the time talking about Reels, so maybe it’s worth giving some color on the rest of the discovery engine work, which is, it has also been doing quite well and is much more incremental to the rest of the business because if people end up being able to discover additional photos or links or groups or things like that in Facebook or just interesting content across Instagram, then they are just more engaged in the product.
And we already know how to monetize that content. So, that ends up being really helpful both for the overall engagement, not very cannibalistic at all and already profitable. So, that’s €“ we have spent less time talking about that because I get that Reels is sort of the faster-growing area. But we do still expect that as a percent of the overall feeds in Facebook and Instagram, recommended content will continue growing. I don’t know if it will be a majority by the end of this year, but maybe it will be 30%-ish, 40%, something in that zone, and continuing to grow because we can just find content that people are going to be interested in that may not be from accounts that they have followed directly. So, hopefully, that’s some useful color on what we are seeing across those efforts.
Operator: Your next question comes from the line of John Blackledge with Cowen. Your line is open.
John Blackledge: Great. Thanks. Two questions. Just a follow-up on Reality Labs, should we continue to expect accelerating losses at Reality Labs in €˜23? And if so, should we expect Reality Labs to be a peak up losses this year? And then second, Susan mentioned shop ads in the bucket of early monetization. Just any color there on how we might see that scale this year. Thank you.
Susan Li: Yes. Sorry, I will go ahead and take the first question about Realty Labs, and Javi, you can take the second question on shop ads. On Reality Labs, we still expect our full year Reality Labs losses to increase in 2023, and we are going to continue to invest meaningfully in this area given the significant long-term opportunities that we see. It is a long-duration investment, and our investments here are underpinned by the accompanying need to drive overall operating profit growth while we are making these investments. I will turn it to Javi on the shop ads.
Javier Olivan: Yes. So, in Q4, we continued to test shop ads in the U.S., and we are seeing increased performance by helping direct the consumer to the place where they are most likely to purchase. So, it’s early to know, but we really finally saw product market fit in the test. It’s off a small base. But to just give you a sense, we saw triple-digit growth in both revenue and adoption across Q4. And we expect this growth to normalize to a lower level in 2023. And the shop ads beta has a revenue run rate in the hundreds of millions of dollars. So, that gives you a small base. It’s a small revenue run rate yet, but it’s growing rapidly. And we expect it to continue, but normalize to more lower levels in 2023.
Operator: Your next question comes from the line of Mark Mahaney with Evercore ISI. Your line is open.
Mark Mahaney: Thanks. Two questions, please. The click-to-messaging is now about $10 billion revenue run rate. How do you think about the growth path for that going forward? And do you find that that’s bringing in brand-new advertisers to Meta that you hadn’t seen before? Like who is coming in on that? And does that give you a new growth path? And then, Mark, if I could just ask you. This year of efficiency, it’s almost like there has been a journey going on since early last year when you talked about €“ talking about driving the business for growing operating profit. And I guess I just want to ask the why question. I mean it’s €“ the markets obviously like what they are hearing from you today and the changes. But why the much greater focus on efficiency, not just tonight, but kind of over the last 9 months or 12 months with maybe a few hiccups.
Like what €“ is it just the maturity of the business? Is it just trying to take on advantage of a crisis, but there is a crisis out there in terms of the economy, and maybe that forced minds to think this way. Just a little bit more color on the why. Thanks a lot.
Susan Li: Thanks. This is Susan. On the click-to-messaging ads, so we are €“ this is one of our fastest-growing ads products. And we believe that they are bringing incremental demand onto our platform. I mentioned in my script that over half of click-to-message advertisers exclusively use click-to-message ads on our platform. In terms €“ you asked how we are going to scale and I think there are a couple of dimensions. In terms of demand, I think the biggest piece here is getting more businesses to adopt click-to-messaging ads via creating more entry points, simplifying creation flows. We are trying to integrate with partners who can help smaller businesses scale. So, there is a lot of work going on there. On the performance side, we are continuing to focus on just driving up the ROI that advertisers get from click-to-messaging ads.
We are trying to give advertisers the ability to do more down-funnel optimization, create better in-thread experiences and simplify flows that help them drive conversion. And then ultimately, we are always focused on growing supply and in this case, growing the business messaging ecosystem by creating more ways for people and businesses to connect across our messaging app. So, I think an example of that would be something like business direct research on WhatsApp. So, it’s an opportunity we are very excited about and we have invested a lot in and we have seen very healthy growth. I will turn it to Mark on the efficiency question.
