Applovin Corp (APP) 2025 Q2 法說會逐字稿

Welcome to AppLovin's earnings call for the second quarter ended June 30, 2025. I'm David Hsiao, Head of Investor Relations. Joining me today to discuss our results are Adam Foroughi, our Co-Founder, CEO and Chairperson; and Matt Stumpf, our CFO.

歡迎參加 AppLovin 截至 2025 年 6 月 30 日的第二季財報電話會議。我是投資者關係主管 David Hsiao。今天與我一起討論我們業績的有我們的共同創辦人、執行長兼董事長 Adam Foroughi 和我們的財務長 Matt Stumpf。

Please note, our SEC filings to date as well as our financial update and press release discussing our second-quarter performance are available at investors.applovin.com. During today's call, we will be making forward-looking statements, including, but not limited to, the future development and reach of our platform including the expected timing of product launches, our expected growth opportunities, the efficiency of our operations, the expected future financial performance of the company and other future events.

請注意，我們迄今向美國證券交易委員會 (SEC) 提交的文件以及財務更新和討論第二季度業績的新聞稿均可在 investors.applovin.com 上查閱。在今天的電話會議上，我們將做出前瞻性陳述，包括但不限於我們平台的未來發展和覆蓋範圍，包括預期的產品發佈時間、我們預期的成長機會、我們營運的效率、公司預期的未來財務表現以及其他未來事件。

These statements are based on our current assumptions and beliefs, and we assume no obligation to update them, except as required by law. Our actual results may differ materially from the results predicted. We encourage you to review the risk factors in our most recently filed form 10-Q for the first quarter ended March 31, 2025. Additional information may also be found on our quarterly report on Form 10-Q for the fiscal quarter ended June 30, 2025, which will be filed today.

這些聲明是基於我們目前的假設和信念，除非法律要求，否則我們不承擔更新它們的義務。我們的實際結果可能與預測的結果有重大差異。我們鼓勵您查看我們最近提交的截至 2025 年 3 月 31 日的第一季 10-Q 表中的風險因素。您也可以在我們今天提交的截至 2025 年 6 月 30 日的財政季度的 10-Q 表季度報告中找到更多資訊。

We will also be discussing non-GAAP financial measures. These non-GAAP measures are not intended to be superior to or a substitute for our GAAP results. Please be sure to review the GAAP results and the reconciliation of our GAAP and non-GAAP financial measures in our earnings release and financial update available on our Investor Relations site.

我們也將討論非公認會計準則財務指標。這些非公認會計準則 (non-GAAP) 指標並非旨在優於或取代我們的公認會計準則 (GAAP) 結果。請務必查看我們投資者關係網站上的收益報告和財務更新中的 GAAP 結果以及 GAAP 與非 GAAP 財務指標的對帳。

This conference call is being recorded, and a replay will be available for a period of time on our IR website. Now I'll turn it over to Adam and Matt for some opening remarks. Then we'll have the moderator take us through Q&A.

本次電話會議正在錄製中，重播將在一段時間內在我們的 IR 網站上提供。現在我將把時間交給亞當和馬特來致開場白。然後主持人將帶領我們進行問答。

Thanks, everyone, for joining us today. We appreciate your time and interest. Q2 2025 was another great quarter, driven by continued strength in gaming advertising. Our growth comes from improved technology, increased demand as well as from supply side expansion. The MAX marketplace creates the supply that drives our growth as well as the growth in the market.

感謝大家今天加入我們。我們感謝您的時間和關注。2025 年第二季度又是一個出色的季度，這得益於遊戲廣告的持續強勁增長。我們的成長來自於科技的進步、需求的成長以及供應方面的擴張。MAX 市場創造了推動我們以及市場成長的供應。

As marketing technologies in the industry continue improving, we expect the supply will keep growing quickly. While we don't disclose exact MAX marketplace growth rates, it has consistently been double digits, far outpacing growth in the in-app purchasing gaming market. The ongoing improvement in our models drive sustainable growth rates beyond the market growth rates, while we continue to expand our dominant leadership position.

隨著行業行銷技術的不斷改進，我們預計供應量將繼續快速成長。雖然我們沒有透露具體的 MAX 市場成長率，但它一直保持兩位數成長，遠遠超過應用程式內購買遊戲市場的成長速度。我們模型的持續改進推動著永續成長率超越市場成長率，同時我們繼續擴大我們的主導領導地位。

Based on all the opportunity in front of us in our core market, we are confident we can sustain 20% to 30% year-over-year growth driven by just gaming. However, what gets us more excited now than ever in our history before is the opportunity to really expand outside our core market.

基於我們核心市場面臨的所有機遇，我們有信心僅靠遊戲就能維持 20% 至 30% 的年成長。然而，現在讓我們比以往更興奮的是，我們有機會真正向核心市場之外擴張。

Recently, we took the first step towards opening up our platform broadly, quietly launching our new AXON ads manager, our self-service portal, which will serve as the foundation for our next decade of growth. Our ads manager has many benefits. It puts day-to-day controls directly in advertisers' hands, reducing friction. It enables credit card billing, eliminating the hassle of monthly invoicing. It provides the architecture for agents that can eventually automate every workflow.

最近，我們邁出了廣泛開放平台的第一步，悄悄推出了新的 AXON 廣告管理器，即我們的自助服務門戶，它將成為我們未來十年成長的基礎。我們的廣告管理器有很多好處。它將日常控制權直接交到廣告商手中，減少了摩擦。它支援信用卡計費，免去了每月開票的麻煩。它為代理提供了最終可以自動化每個工作流程的架構。

It establishes the framework for automatically generated ads. It simplifies onboarding through our recently launched Shopify app. It deepens integrations with attribution providers, giving customers more accurate reporting. With the rollout going smoothly, we were ready to widen access.

它為自動產生的廣告建立了框架。它透過我們最近推出的 Shopify 應用程式簡化了入職流程。它深化了與歸因提供者的整合，為客戶提供更準確的報告。隨著推廣工作的順利進行，我們已準備好擴大存取權限。

On October 1, 2025, we plan to open the AXON ads manager on a referral basis, perfectly timed for the holiday season. Feedback from these partners will guide our global public launch in the first half of 2026. To date, web advertising campaigns have been limited to the United States.

2025 年 10 月 1 日，我們計劃以推薦方式開放 AXON 廣告管理器，適逢假期。這些合作夥伴的回饋將引導我們在 2026 年上半年的全球公開發布。迄今為止，網路廣告活動僅限於美國。

On October 1, we plan to open our platform to most major international markets. Now stepping back, we have spent the last decade assembling the pieces, reach of more than 1 billion users, best-in-class optimization and now a self-service interface. Together, they position us to help any business of any size anywhere in the world grow profitably.

10月1日，我們計劃向大多數主要國際市場開放我們的平台。現在回顧一下，我們花了十年的時間整合這些組件，覆蓋超過 10 億用戶，實現一流的優化，現在還擁有自助服務介面。它們共同幫助我們幫助世界各地任何規模的企業獲利成長。

That is good for our partners. It's good for economies around the world, and it's great for job creation. The opportunity is so big that we will be launching the platform under its own brand, AXON. Once AXON is fully open next year, we plan to begin paid marketing to recruit new advertisers, which will drive predictable compounding growth. We have been building performance-driven advertising products longer and better than most anyone.

這對我們的合作夥伴來說是件好事。這對世界各地的經濟都有好處，而且對創造就業機會也很有幫助。這個機會如此之大，我們將以自己的品牌 AXON 推出這個平台。一旦 AXON 明年全面開放，我們計劃開始付費行銷來招募新的廣告商，這將推動可預測的複合成長。我們打造以效果為導向的廣告產品的時間比大多數人都要長，品質也更好。

Operating at our current scale with an incredibly small amount of advertiser relationships highlights the magnitude of the opportunity ahead. Our strategy is simple: build world-class products, launch them when they meet our high bar and compound from there.

以我們目前的規模經營，與數量極少的廣告商建立關係，凸顯了未來機會的巨大。我們的策略很簡單：打造世界級的產品，當它們達到我們的高標準時推出，然後從那裡開始復合。

Patient, disciplined execution produces durable success, and we hope our track record gives you the same confidence we have in our future. We're incredibly excited about what's ahead. With that, I'll turn it over to Matt for a closer look at the numbers.

耐心、嚴謹的執行將帶來持久的成功，我們希望我們的業績記錄能讓您像我們一樣對我們的未來充滿信心。我們對未來感到無比興奮。說完這些，我將把它交給馬特來仔細研究這些數字。

Thanks, Adam, and thanks to everyone for joining us today. Q2 was another exceptional quarter for AppLovin. At the end of the quarter, we closed the sale of our Apps business to Tripledot Studios. This quarter, the financial results for the Apps business are included within discontinued operations, and we will keep our commentary limited to the advertising business only.

謝謝，亞當，也謝謝大家今天加入我們。對於 AppLovin 來說，第二季又是另一個非凡的季度。在本季末，我們完成了將應用程式業務出售給 Tripledot Studios 的交易。本季度，應用程式業務的財務業績包含在已終止經營的業務中，我們的評論僅限於廣告業務。

During the quarter, revenue increased by a very healthy 77% from last year to approximately $1.260 billion, while adjusted EBITDA nearly doubled to an impressive $1.020 billion, achieving an 81% adjusted EBITDA margin.

本季度，營收較去年同期大幅成長 77%，達到約 12.6 億美元，而調整後的 EBITDA 幾乎翻了一番，達到令人印象深刻的 10.2 億美元，實現了 81% 的調整後 EBITDA 利潤率。

The majority of our revenue growth in the quarter was driven by our core gaming business. While e-commerce continues to perform well, we limited onboarding of new customers to focus on the preparation for the self-serve launch in Q4.

本季我們的營收成長主要來自於核心遊戲業務。雖然電子商務繼續表現良好，但我們限制了新客戶的加入，以專注於第四季度自助服務的準備。

Quarter-over-quarter flow-through from revenue to adjusted EBITDA was a very strong 81%, illustrating our continued dedication to operating lean. At the end of the second quarter, we had $1.2 billion in cash and cash equivalents, which includes $425 million in net cash received from the sale of the Apps business.

營收與調整後 EBITDA 之間的環比流轉率高達 81%，顯示我們持續致力於精實營運。截至第二季末，我們擁有 12 億美元現金和現金等價物，其中包括出售應用程式業務所獲得的 4.25 億美元淨現金。

In the second quarter, we generated $768 million in free cash flow, up a staggering 72% year-over-year. Our free cash flow was slightly lower than last quarter due to the timing of payments for interest on our bonds, which are semiannual and certain taxes associated with the prior year.

第二季度，我們產生了 7.68 億美元的自由現金流，比去年同期成長驚人的 72%。我們的自由現金流略低於上一季度，這是由於我們債券利息的支付時間為半年，並且某些稅款與上一年有關。

This quarter, we repurchased and withheld approximately 900,000 shares for a total cost of $341 million funded through free cash flow. As a result of our ongoing strategic share management activities, we were able to reduce our weighted average diluted common shares outstanding this year from 346 million in the fourth quarter to 342 million this quarter. Finally, turning to our financial guidance for next quarter.

本季度，我們回購並扣留了約 90 萬股，總成本為 3.41 億美元，透過自由現金流提供資金。由於我們正在進行的策略性股份管理活動，我們能夠將今年的加權平均稀釋普通股從第四季度的 3.46 億股減少到本季的 3.42 億股。最後，談談我們對下個季度的財務指引。

In the third quarter of 2025, for the Advertising business, we anticipate delivering between $1.320 billion and $1.340 billion in revenue, with adjusted EBITDA between $1.070 billion and $1.090 billion, targeting an adjusted EBITDA margin of 81%. We're confident these targets position us to continue driving strong growth and value for our partners and shareholders. Now with that, let's move to Q&A.

2025 年第三季度，對於廣告業務，我們預計營收在 13.2 億美元至 13.4 億美元之間，調整後的 EBITDA 在 10.7 億美元至 10.9 億美元之間，調整後的 EBITDA 利潤率目標為 81%。我們相信，這些目標將使我們能夠繼續為我們的合作夥伴和股東帶來強勁的成長和價值。現在，讓我們進入問答環節。

(Operator Instructions) Matthew Cost, Morgan Stanley.

（操作員指示）馬修·科斯特，摩根士丹利。

Hi everybody. Thanks for taking the question. I guess on your plans to start doing paid marketing next year, talk to us a little bit about how you arrived at the decision to -- market to acquire advertisers because I think historically it's mostly been growth by word of mouth and sort of virality within the gaming industry that people become aware of your of your products.

大家好。感謝您回答這個問題。我想問一下，您計劃明年開始做付費行銷，請跟我們簡單談談您是如何做出這個決定的——透過行銷來吸引廣告商，因為我認為從歷史上看，遊戲產業的成長主要是透過口耳相傳和病毒式傳播，讓人們了解您的產品。

So I guess why is now the right time to start marketing? What channels are you going to market through? And how are you going to evaluate the return there? And then I have a follow up.

那我想為什麼現在是開始行銷的最佳時機？您將透過哪些管道進行行銷？您如何評估那裡的回報？然後我有一個後續行動。

Matt, good to see you. So look, we've got big aspirations with our platform. The advertising solution right now to web commerce advertisers and more broadly, other categories we've talked about is looking really strong.

馬特，很高興見到你。所以，看吧，我們對我們的平台抱持著遠大的期望。目前，針對網路商務廣告商以及更廣泛的我們討論過的其他類別的廣告解決方案看起來非常強勁。

Obviously, we have very small penetration in terms of advertiser base against what our global advertisers count. And the way we look at our business is if the model works this well at this small amount of penetration, what happens when we can really open it up and go service all the small businesses in the world.

顯然，與我們的全球廣告商數量相比，我們的廣告商基礎滲透率非常小。我們看待我們業務的方式是，如果該模式在這麼小的滲透率下運作良好，那麼當我們真正開放它並為世界上所有的小型企業提供服務時會發生什麼。

Our aspirations are to help any business of any size be able to acquire customers profitably. And if we can do that, we'll achieve the goals that we've set for ourselves. So the value of us being able to go out in performance market, the platform, is that we've got one of the most lucrative financial models the world has ever seen. I mean, obviously, you can see the amount of cash that we print. And we're very good performance marketers.

我們的願望是幫助任何規模的企業都能獲利地獲取客戶。如果我們能夠做到這一點，我們就能實現我們為自己設定的目標。因此，我們能夠進入表演市場和平台的價值在於，我們擁有世界上迄今為止最有利可圖的財務模式之一。我的意思是，顯然你可以看到我們印刷的現金數量。我們是非常優秀的績效行銷人員。

So we're very good performance marketers. It's plausible that we will be using our own models to recruit advertisers off of our own inventory. There's plenty of moms with small businesses and dads with small businesses sitting in games, playing games all day that could use our platform to market themselves.

所以我們是非常優秀的績效行銷人員。我們可能會使用自己的模型從自己的庫存中招募廣告商。有很多經營小生意的媽媽和爸爸整天坐在遊戲中玩遊戲，他們可以利用我們的平台來推銷自己。

You can imagine us running ads on Facebook, on LinkedIn, on TikTok. But it really does come down to the fact that we've got such a lucrative financial model, and we try to run lean and automate every step of that process.

你可以想像我們在 Facebook、LinkedIn 和 TikTok 上投放廣告。但事實上，我們擁有如此有利可圖的財務模式，並且我們試圖精實營運並實現流程每一步的自動化。

And we believe that if we can automate the onboarding of advertiser flow from when an advertiser can find out about us from an ad all the way through to going live and then scaling on our platform, the model will be very lucrative on an LTV to CAC basis, and we won't have to staff up a large sales force.

我們相信，如果我們能夠實現廣告客戶流程的自動化，從廣告客戶透過廣告了解我們，到上線並在我們的平台上擴展，該模式在 LTV 到 CAC 的基礎上將非常有利可圖，而且我們不需要配備龐大的銷售團隊。

Great. And then on the supply side, you made some comments in the prepared remarks about some strong supply growth that you're seeing. Given that MAX is kind of already starting from such a strong position already, is the further growth, is it a function of just taking even more share in mediation? Is it about ad load increasing to the customers you already have? I guess where is the supply growth coming from?

偉大的。然後在供應方面，您在準備好的發言中對所看到的強勁供應成長發表了一些評論。鑑於 MAX 已經從如此強大的地位起步，進一步的成長是否僅僅意味著在調解中佔據更多份額？這是否會增加您現有客戶的廣告負載？我猜想供應成長從何而來？

Yeah. Look, we talk a lot about the market in mobile gaming and people fixate on the in-app purchasing market. And what we're trying to do is just highlight the fact that our business grows from improvements in technology and demand, which drives CPM and expansion of supply.

是的。你看，我們談了很多關於手機遊戲市場的事情，人們關注的是應用程式內購買市場。我們試圖強調的事實是，我們的業務成長源於技術和需求的改進，這推動了 CPM 和供應的擴張。

The MAX mediation platform is really high penetration into the market, so we're not going to be able to grow it particularly quickly by taking more mediation from other platforms at this point. What we were highlighting was that the base platform, just the audience inside MAX, is growing swiftly, double-digit growth. So multiples greater than what you would expect when people try to size up the gaming market growth rates of 3% to 5%.

MAX 中介平台的市場滲透率確實很高，因此我們目前無法透過從其他平台獲得更多中介來實現其快速成長。我們強調的是，基礎平台（即 MAX 內部的受眾）正在快速成長，實現兩位數的成長。因此，當人們試圖估計遊戲市場的成長率為 3% 至 5% 時，這個數字將高於預期。

The other thing that is super valuable to understand in that is that inside that marketplace, we historically have always talked about how there's not a share gain concept. As the marketing platforms improve, that platform grows. The eyeballs playing games play more games every single day.

另一件非常有價值的事情是，在那個市場中，我們歷來都在談論如何不存在份額收益的概念。隨著行銷平台的改進，該平台也在不斷發展。玩遊戲的眼球每天都會玩更多的遊戲。

Inventory goes up. That drives growth on our platform. It drives growth on the platforms that are the other bidders inside that ecosystem. And then we benefit the most because we're taxing every transaction for the most part that happens outside of us at the MAX fee. And then obviously, our own DSP is a super lucrative business model when we win the inventory.

庫存增加。這推動了我們平台的成長。它推動了該生態系統內其他競標者平台的成長。然後我們受益最大，因為我們對大部分在我們之外發生的每筆交易都按最高費用徵稅。顯然，當我們贏得庫存時，我們自己的 DSP 是一種超級有利可圖的商業模式。

Great, thanks so much.

太好了，非常感謝。

Ralph Schackart, William Blair.

拉爾夫·沙卡特、威廉·布萊爾。

Good afternoon, Adam. Matt. Thanks for taking the question. Just two if I could. Just on the self-serve platform, Adam, maybe you could sort of frame this for us. The launch is coming up, public launch coming up, which is exciting. But could this be a pretty material impact to the overall business? Obviously, it's of good scale today and growing pretty fast, but maybe kind of frame that opportunity. And then I have a follow-up, please.

下午好，亞當。馬特。感謝您回答這個問題。如果可以的話就兩個。就在自助服務平台上，亞當，也許你可以為我們設計這個。發布會即將舉行，公開發布會即將舉行，令人興奮。但這會對整體業務產生重大影響嗎？顯然，它目前規模很大，而且成長很快，但也許可以建立這樣的機會。然後我還有一個後續問題，請回答。

Yeah, sure. I mean let's look back over the last year at our numbers, and then I'm going to give you qualitatively sort of what was going on with growth in the business. In Q4, we saw a huge ramp-up in the e-commerce category when we went into this pilot state, recruited hundreds of advertisers that we disclosed. And that was a big ramp-up in Q4, where, in Q1, we got the full benefit of that run rate. And if you look back at the quarter-over-quarter growth, Q4 and Q1 were really high growth rates.

是的，當然。我的意思是，讓我們回顧去年的數據，然後我會從品質上告訴你業務成長的情況。在第四季度，當我們進入這個試點州並招募了數百名廣告商時，我們看到電子商務類別的大幅成長。這是第四季度的一個巨大增長，而在第一季度，我們充分享受了這一運行率帶來的好處。如果回顧季度環比成長情況，就會發現第四季和第一季的成長率確實很高。

Now, the reason for that is that the new type of customer that comes into our platform is extremely incremental to our business like we've talked about in the past. Then, since then, we knew that we had highlighted a whole bunch of things that we had to go build into the platform to be able to service advertisers at the level that we like to do. And we like to set a really high bar for the products we deliver.

現在，原因在於，正如我們過去談到的那樣，進入我們平台的新類型的客戶對我們的業務具有極大的增量作用。從那時起，我們就知道，我們已經強調了一大堆必須在平台上建立的東西，以便能夠以我們喜歡的水平為廣告商提供服務。我們希望為我們交付的產品設定一個非常高的標準。

We wanted to build the ads manager, which we released and we will continue to innovate and iterate through. We wanted to build dynamic product ads that came out in the last couple of quarters. We wanted to do better integrations with attribution companies, which happened inside the quarter. And we wanted to launch a Shopify app and other apps that allow for seamless integration amongst the advertiser base.

我們希望建立廣告管理器，我們已經發布了該管理器，並將繼續進行創新和迭代。我們希望打造在過去幾季推出的動態產品廣告。我們希望與歸因公司進行更好的整合，這在本季度內就已實現。我們希望推出 Shopify 應用程式和其他允許廣告商群體無縫整合的應用程式。

So we ended up constraining the advertiser onboarding process for a couple of quarters while we made sure that the product was at the level that we wanted to get it to. Now we're looking at, in Q4, talking about a referral-based opening. That, in itself -- I'm going to define that for you so you understand what we're talking about, but accounts on the platform, they obviously like our solutions because they're spending substantial amounts of money on our platform. We'll get the chance to refer their colleagues into our platform and have them go live in a self-service way.

因此，我們最終在幾個季度內限制了廣告商的入職流程，同時確保產品達到了我們想要的水平。現在，我們正在考慮在第四季度討論基於推薦的開局。就其本身而言——我將為您定義這一點，以便您了解我們在說什麼，但平台上的帳戶顯然喜歡我們的解決方案，因為他們在我們的平台上花費了大量資金。我們將有機會將他們的同事推薦到我們的平台並讓他們以自助服務的方式上線。

We expect that will increase the advertiser count quite quickly and also allow us to go through live examples of advertisers coming in self-service all the way to scale on our product. Assuming all that goes well, then we talked about opening up the platform entirely to the world in first half of next year. We think as advertiser count grows on our business, especially in categories outside of gaming, you're going to see a lot of upside in the numbers that we're able to report.

我們預計這將很快增加廣告商的數量，同時也使我們能夠透過廣告商的自助服務實例來擴大我們的產品規模。如果一切進展順利，我們將討論在明年上半年向全世界全面開放平台。我們認為，隨著我們業務的廣告商數量的增長，特別是在遊戲以外的類別中，您將看到我們能夠報告的數字有很大的上升空間。

Great. Maybe one more if I could. I think historically, you've talked about e-com being around 10% or so of the business for this year. Maybe just sort of an update if that's current thinking. And then I think in your prepared remarks, you talked about limited onboarding to e-com customers.

偉大的。如果可以的話也許再多一個。我認為從歷史上看，您曾談到電子商務將佔今年業務的 10% 左右。如果這是目前的想法，那麼可能只是一種更新。然後我認為在您準備好的發言中，您談到了對電子商務客戶的有限入職培訓。

Did that sort of limit the growth rate in the quarter had you not limited those customers being onboarded?

如果您沒有限制新客戶的數量，這是否會限製本季的成長率？

Yeah, for sure. I mean, I'll take it in reverse. By constraining the advertiser count, we're limiting growth. Now, we're focused on growing the cohort that's live, which we saw growth from the cohort that's live, but Matt is prepared remarks that majority of our growth came from gaming.

是的，當然。我的意思是，我會反過來做。透過限制廣告商數量，我們限制了成長。現在，我們專注於擴大線上玩家群體，我們也看到了線上玩家群體的成長，但馬特準備表示，我們的大部分成長來自遊戲。

So if you assume a large, large amount of the growth in the quarter came from gaming and you're roughly 9% quarter of a quarter, gaming is still a 30% to 40% grower for us. So well above our 20% to 30% long term goals that we've stated.

因此，如果你假設本季的大部分成長來自遊戲，每季的成長率約為 9%，那麼遊戲對我們來說仍然有 30% 到 40% 的成長。遠高於我們提出的 20% 至 30% 的長期目標。

Now e-commerce, we went from a state where we ramped it in Q1 and then controlled, as I said, on the advertiser on boarding while we got these tools ready to go. And so where we were at was that sort of 10% range that we were targeting for the year. There's no reason to expect it to have grown above that because gaming is growing so quickly, and for two quarters now we've mentioned that gaming was a big contributor to growth.

現在，對於電子商務，我們從第一季開始大力發展，然後，正如我所說，在我們準備好這些工具的同時，對廣告商進行控制。因此，我們現在所處的水平正是我們今年所設定的 10% 的目標範圍。沒有理由期望它會成長超過這個數字，因為遊戲成長如此之快，我們已經提到兩個季度遊戲是成長的一個重要貢獻者。

As we go into Q4, that's a huge holiday shopping season. So not only are you going to see the cohort that we have live spend a lot more, you're also going to have new on boarding happening for the first time in our history at a rate that's much higher than we will have ever seen before.

進入第四季度，這是一個重要的假期購物季。因此，您不僅會看到我們現有的客戶群花費更多，而且還會看到我們歷史上第一次以比以往任何時候更高的速度迎來新客戶。

So we fully expect that e-commerce will see a pretty substantial ramp up through that what you can call a soft launch period. And then obviously, as we go. Into a broader global release, the impact from that. And then the last point to remember is another one of my prepared remarks highlighted the fact that we have constrained the advertisers we even have live today by not allowing them to buy our audience that's international. The vast majority of our user audience is outside the US. We will be releasing almost all markets once we go into this October 1 release.

因此，我們完全預期電子商務將在所謂的軟啟動期實現相當大的成長。然後顯然，隨著我們前進。放眼全球，影響由此而來。最後要記住的一點是，我的另一篇準備好的發言強調了這樣一個事實：我們限制了我們今天的廣告商，不允許他們購買我們的國際觀眾。我們的用戶絕大多數都在美國以外。一旦進入 10 月 1 日發布，我們將發布幾乎所有市場。

Omar Dessouky, Bank of America.

美國銀行的奧馬爾‧德蘇基 (Omar Dessouky)。

I just wanted to take the conversation in a slightly different direction if I could and ask about game engine data. And one of your competitors, specifically Unity, which owns a game engine, which 70% or so of mobile games are built on plans to use game engine data for ad targeting purposes sometime in 2026.

如果可以的話，我只是想讓談話稍微轉向不同的方向並詢問遊戲引擎數據。您的競爭對手之一 Unity 擁有一個遊戲引擎，大約 70% 的手機遊戲都是基於該引擎構建的，該公司計劃在 2026 年某個時候使用遊戲引擎數據進行廣告定位。

Do other companies besides them have access to game engine data? Do you know if your customers own their own game engine data and you're able to access it and use it at some point in the future? And do you think it matters?

除了他們之外還有其他公司可以存取遊戲引擎數據嗎？您是否知道您的客戶是否擁有自己的遊戲引擎數據，並且您是否能夠在未來的某個時間點存取和使用它？您認為這很重要嗎？

Yeah. Omar, we can't speak to other people's data. I mean we don't know what game engine data even means. But when you're integrated inside an application, both as a publisher and as an advertiser, you have a lot of data points that you're able to extract that are behavioral data points of how the consumer is playing into a game.

是的。奧馬爾，我們不能談論其他人的數據。我的意思是我們甚至不知道遊戲引擎數據意味著什麼。但是，當您作為發布商和廣告商整合到應用程式中時，您就可以提取大量數據點，這些數據點是消費者如何玩遊戲的行為數據點。

We're obviously very good. Our models are cutting edge. AXON 2 has shown phenomenal growth for, I don't know, eight, nine quarters now since release. You have to assume we're pretty good at using the data we have available to us. So what we don't know, as you didn't define game engine data is, is there some magical data out there. But our market penetration inside gaming is really large at this point.

我們顯然非常優秀。我們的模型是尖端的。自發布以來，AXON 2 已經經歷了八、九個季度的驚人成長。你必須假設我們非常擅長利用我們掌握的數據。因此，由於您沒有定義遊戲引擎數據，我們不知道是否存在一些神奇的數據。但目前我們在遊戲領域的市場滲透率已經非常大。

It's materially higher than 70%. We've got very good visibility into what matters in the gaming category given how large we've become. What's really going to matter for our business as we go forward is whether we can go expand out this offering the way we expect to across all businesses of all sizes in any category. If we're able to do that, we will have a much better sense of the consumer on the other side.

它明顯高於70%。考慮到我們的規模，我們對遊戲類別中的重要事項有了很好的了解。對於我們未來的業務來說，真正重要的是我們能否按照我們期望的方式，將這項服務擴展到所有類別、各種規模的所有企業。如果我們能夠做到這一點，我們將對另一方的消費者有更好的了解。

And I do want to remind everyone, we have 1 billion-plus users and they're not gamers. These are human beings doing a whole bunch of things, but the share of wallet from gaming is going to be a minority of the dollars that these people are spending outside of gaming.

我確實想提醒大家，我們有超過 10 億用戶，他們不是遊戲玩家。人類做了很多事情，但從遊戲中獲得的收入只佔這些人在遊戲之外花費的一小部分。

And that's what's really compelling to us as we go to this next chapter in the company. If our models that are already as good as they are with such a small amount of advertiser penetration now go get visibility into consumer behavior across every category, we're going to be way more predictive about the advertisement that we could show the consumer.

當我們進入公司的新篇章時，這對我們來說是真正有吸引力的事情。如果我們的模型在廣告商滲透率如此低的情況下就已經很好了，現在能夠洞察各個類別的消費者行為，那麼我們就能更好地預測可以向消費者展示的廣告。

Thank you very much.

非常感謝。

Chris Kuntarich, UBS.

瑞銀的 Chris Kuntarich。

Thanks for taking the question here. Maybe just on the referral program and how this is going to work functionally. Are advertisers that are currently on this platform, are they going to be given referral bonuses? Is there going to be any restriction in advertisers that they could refer?

感謝您在這裡提出這個問題。也許只是關於推薦計劃以及其功能如何運作。目前在這個平台上的廣告商是否會獲得推薦獎金？他們所推薦的廣告商會受到什麼限制嗎？

You've talked about opening up international inventory. Will international advertisers be available to participate in the referral program? And just as we think about kind of this line out the door of advertisers you've been talking about for a while, are they going to get any special look here potentially to come in here maybe if they don't get a referral?

您曾談到開放國際庫存。國際廣告商可以參與推薦計畫嗎？正如我們思考您剛才談到的這種廣告客戶排隊現像一樣，如果他們沒有得到推薦，他們是否會在這裡得到特別的關注並有可能進入這裡？

Chris, I can't say we have all the answers to all those questions you asked. We're going to definitely go through an iterative period, too, on the release of the referral program. But fundamentally, we don't think a ticket into our system is really worth paying for. If someone is a client of ours, they have a lot of benefit from being on the platform.

克里斯，我不能說我們已經回答了你提出的所有問題。在推薦計畫的發布過程中，我們也肯定要經歷一個迭代期。但從根本上來說，我們認為進入我們系統的門票並不值得花錢購買。如果有人是我們的客戶，他們會從這個平台上獲得很多好處。

People love using social media. You've already seen a whole bunch of organic posts about us from influencers. We think as we give them the capacity to invite their peers, they're going to do it on their own because the platform has been restrictive and exclusive for so long. If we see that happen, it's much better organically. And beyond that, line out the door, we'll still need to most likely get an invite into the platform from one of our customers to be able to get automatically cleared.

人們喜歡使用社群媒體。您已經看到了大量來自有影響力人士的關於我們的自然帖子。我們認為，當我們賦予他們邀請同伴的能力時，他們會自己去做，因為這個平台長期以來一直是限制性和排他性的。如果我們看到這種情況發生，那麼從有機角度來看情況會好得多。除此之外，在門外排隊，我們仍然很可能需要從我們的一位客戶那裡獲得進入平台的邀請才能自動清除。

Right. Okay. Just as we're kind of thinking about the 3Q guidance here maybe a little bit faster than what we've seen in the past couple of quarters, is there anything kind of reflected here in the faster sequential growth associated with the App portfolio divestiture? And any sort of quantification around that would be helpful.

正確的。好的。正如我們在思考第三季的指引可能會比過去幾季的指引要快一點一樣，與應用程式組合剝離相關的更快的連續成長是否反映了什麼？任何形式的量化都會有所幫助。

Yeah, sure. So the one change we did make, Chris, this quarter, which is slightly different than kind of our typical cadence that we've been going every quarter, was to include the benefit as well that we're going to see from the divestiture of the Apps business, where we have a slight pickup in revenue. So we also incorporated that within the Q3 guide.

是的，當然。因此，克里斯，本季度我們確實做出了一個改變，這與我們每個季度的典型節奏略有不同，那就是也包括了我們將從應用程式業務剝離中看到的好處，我們的收入略有回升。因此我們也將其納入了第三季指南中。

Got it. Thank you.

知道了。謝謝。

Jason Bazinet, Citi.

花旗銀行的 Jason Bazinet。

So I think on the last call, you all said that you're definitely working towards building two AXON models, right, one that's specific to games and one that's more focused on the other e-commerce. You use a new term now. I can't remember what it is, but I'll call it e-commerce. But I know it's more holistic.

所以我認為在上次通話中，你們都說過，你們肯定正在努力建立兩個 AXON 模型，對吧，一個專門針對遊戲，另一個更專注於電子商務。你現在使用了一個新術語。我不記得它是什麼了，但我會稱它為電子商務。但我知道它更全面。

When you think about all the data that you had to ingest and the models that you had to build to make the gaming model as good as it is, what's sort of the right runway that you have in your mind before those two could be at parity in terms of efficacy?

當您考慮到您必須攝取的所有數據以及您必須建立的模型以使遊戲模型盡可能好時，在兩者在功效方面達到同等水平之前，您心中的正確跑道是什麼？

I mean, look, Jason, it's a great question. There's -- one, there's no way to know that, right, because that's future looking, and we're going to go acquire a lot of data as we open up the platform. But in terms of architecture of the models, the two models are distinctly different because on the one hand, a user is going to a website and on the other hand, a user is going to app store and that app store isn't even something that we can measure.

我的意思是，傑森，你看，這是一個很好的問題。首先，我們無法知道這一點，因為這是面向未來的，當我們開放平台時，我們將獲得大量資料。但就模型架構而言，這兩個模型截然不同，因為一方面，使用者造訪的是網站，另一方面，使用者造訪的是應用程式商店，而應用程式商店甚至不是我們可以衡量的東西。

And so you've got two frameworks. And then as you touched on in gaming, when we launched AXON 2.0, we already had pretty good market penetration. Now, we didn't work with the largest customers, so we continue to get better as we broke into the largest customers in the category. But I want to say we had 50%, 60% penetration and now market penetration in gaming, as you all know, we're almost a requirement at this point given how good we are.

所以你有兩個框架。正如您在遊戲方面提到的那樣，當我們推出 AXON 2.0 時，我們已經擁有了相當不錯的市場滲透率。現在，我們還沒有與最大的客戶合作，因此，隨著我們打入該類別中最大的客戶，我們會繼續做得更好。但我想說的是，我們的滲透率已經達到了 50%、60%，現在遊戲市場的滲透率，正如大家所知，鑑於我們的優秀表現，我們幾乎已經成為了必需品。

In e-commerce, when we launched in the pilot, why we were so excited about the result is that the first beauty shop that went live, the model knew nothing about the shopper behavior inside the beauty category. And we talked about this. That was the big concern. Could AppLovin's tech actually go out and figure out predictively how to convert a user to go buy from a beauty store with no knowledge?

在電子商務中，當我們在試點中推出時，我們之所以對結果如此興奮，是因為上線的第一個美容店，模型對美容類別內的購物者行為一無所知。我們談論了這一點。這是最令人擔憂的問題。AppLovin 的技術是否真的能夠預測性地找出如何讓用戶在不具備任何知識的情況下去美容店購買產品？

Now we've talked about the one disclosure we gave hundreds of e-commerce businesses live on the platform. That's a pathetic amount versus the grand scope of how big these categories are. I mean, Shopify alone has millions of shops, right? So you have so much potential for data accumulation. And then just remember, the data in our platform is not unique to one or the other part of the business. So we've always felt that games will benefit a lot from our ability to break into these other categories.

現在我們已經討論了在平台上向數百家電子商務企業揭露的資訊。與這些類別的宏大範圍相比，這個數字實在是太少了。我的意思是，光是 Shopify 就有數百萬家商店，對嗎？所以你在資料累積方面有很大的潛力。然後請記住，我們平台中的資料並不是業務某個部分所獨有的。因此，我們始終認為，如果我們能夠打入其他類別，遊戲將會受益匪淺。

You can imagine if someone buys a $4,000 handbag from a store, that person is probably a whale for a match-three game. Now I can tell you that, but I can tell you the technology is going to come up with a lot of cross-correlations and conclusions that are a lot more powerful than that.

你可以想像一下，如果有人從商店購買了一個價值 4,000 美元的手提包，那麼這個人很可能就是三消遊戲的狂熱愛好者。現在我可以告訴你這一點，但我可以告訴你，這項技術將會產生許多比這更強大的交叉關聯和結論。

So in terms of opportunity for us, not only does opening up the platform get us more demand, which is going to be massively accretive and incremental to our business. It gets us more data. And so every single quarter, you're going to have that flywheel effect that that then paired with our engineers' ability to take added data and improve the technology and its interpretation of that data creates a real strong foundation for growth for a long time to come.

因此，就我們的機會而言，開放平台不僅能為我們帶來更多需求，而且對我們的業務也將產生巨大的增值和增量作用。它為我們提供了更多數據。因此，每個季度，你都會感受到飛輪效應，再加上我們的工程師獲取額外數據、改進技術以及對數據的解讀的能力，為未來很長一段時間的增長奠定了真正堅實的基礎。

Rob Sanderson, Loop Capital.

桑德森（Rob Sanderson），Loop Capital。

Yeah, good afternoon, gentlemen. Thanks for taking my question. My question is on just improving performance for your web-based advertisers. Again, as we speak to pilot customers, obviously, we hear you're highly effective at driving conversions, but really lacks a lot of basic targeting functions like small things like just excluding existing customers, et cetera.

是的，先生們，下午好。感謝您回答我的問題。我的問題只是關於如何提升你們網路廣告商的表現。再次，當我們與試點客戶交談時，顯然，我們聽說您在推動轉換方面非常有效，但確實缺乏許多基本的定位功能，例如排除現有客戶等小事。

They see -- this seems pretty simple to solve for, but maybe not when you're multi-app inventory. I don't know. I mean the question is really targeting and other improvements that web-based customers are asking for.

他們看到了——這個問題似乎很容易解決，但當你有多應用程式庫存時可能就不是這樣了。我不知道。我的意思是，問題實際上在於網路客戶所要求的目標和其他改進。

Are these largely things that you've already seen and solved for gaming customers and it just takes time to collect feedback and work optimizations into the product, et cetera? Or are you really solving for like a whole new set of challenges, maybe some of both?

這些主要是您已經為遊戲客戶看到和解決的問題，只是需要時間來收集回饋並對產品進行最佳化等等？或者您真的在解決一系列全新的挑戰，或者兩者兼而有之？

Yeah, it's great question. So look, when we got into the space, it surprised us how much of the market Meta advertising was in the D2C space. And it was the majority. And when you have one platform that's that big, everyone wants every other platform to give them the same exact tools.

是的，這是個很好的問題。所以，當我們進入這個領域時，我們驚訝地發現元廣告在 D2C 領域佔據了多大的市場。而且這是大多數。當你擁有一個如此大的平台時，每個人都希望其他平台也能為他們提供完全相同的工具。

Now we don't have an e-mail address or that persistent identifier that matches up with their audience data. So technically doing an exclude is not going to be as accurate as what Meta can deliver to these advertisers. There are other nuances and differences.

現在我們沒有電子郵件地址或與他們的受眾資料相符的持久識別碼。因此，從技術上講，排除並不像 Meta 向這些廣告商提供的資訊那麼準確。還有其他細微差別和差異。

We serve a full-screen advertisement, and then we can pair the video with a dynamic product ad. That creates more intent for the shoppers. So the vast majority of all the transactions we drive happen within 1 hour or 2, and the vast majority of the transactions Meta drive, they take attribution for a much longer time frame. So are these differences between platforms, but advertisers like comfort.

我們提供全螢幕廣告，然後我們可以將影片與動態產品廣告配對。這為購物者創造了更多的意圖。因此，我們推動的所有交易中的絕大多數都發生在 1 到 2 小時內，而 Meta 推動的絕大多數交易都需要更長的時間範圍進行歸因。平台之間也存在這些差異，但廣告商喜歡舒適。

They like what they know. They want to stick in the same processes. What we believe, because we can see it, our ads drive a lot of intent and drive two conversions very, very quickly, and we've been able to extract a lot of value out of the space.

他們喜歡他們所了解的事物。他們希望堅持相同的流程。我們相信，因為我們可以看到，我們的廣告能夠激發很多意圖，並且非常非常快地推動兩次轉化，並且我們已經能夠從這個空間中提取出很多價值。

I mean if you take the disclosure that I gave you in -- I think it was Q1, hundreds of advertisers, I believe it was in the 600s, and $1 billion run rate. The market penetration in terms of market of scale of advertisers to total market is probably 1% or less in that number.

我的意思是，如果你接受我給你的披露——我認為是第一季度，有數百個廣告商，我相信數量在 600 左右，運行率為 10 億美元。以廣告商的市場規模佔整個市場的滲透率大概只有1%甚至更低。

And to get to $1 billion run rate, clearly, even if they're complaining about feature missing, the money is showing up at a level that you wouldn't expect for that little market penetration. So then as you go extrapolate to getting new customers on, we think the products that we deliver work. We think their mindsets will change over time because they'll realize platforms are different, and that's good for their business. It's not bad for their business.

而為了達到 10 億美元的運行率，顯然，即使他們抱怨缺乏功能，但對於如此小的市場滲透率來說，這筆錢的流入水平是意想不到的。因此，當您推斷出獲得新客戶時，我們認為我們提供的產品是有效的。我們認為他們的心態會隨著時間的推移而改變，因為他們會意識到平台是不同的，這對他們的業務有利。這對他們的生意來說並不壞。

Now the last thing I'll leave you with is we do believe in automation entirely throughout the funnel. We don't allow gaming companies to use any sort of manual targeting in our platform. The platform allows them to input a goal, put it in a budget and get that result.

現在我要告訴你們的最後一件事是，我們確實相信整個漏斗的自動化。我們不允許遊戲公司在我們的平台上使用任何形式的手動定位。該平台允許他們輸入目標、納入預算並獲得結果。

And that's what they pay us for, the technology being extremely precise and removing that human mind out of the equation. We bring the same view to this category. All these technologies in advertising are going to move to AI, automating most of this funnel in the future. We've already done that in our largest part of our business, and we are committed to doing that in this category and not allowing people to override what is a smarter system than the human being.

他們付錢給我們就是為了實現這個目標，因為這項技術極為精確，而且不需要人類的思維。我們對這類別抱持同樣的看法。廣告中的所有這些技術都將轉向人工智慧，未來將實現大部分流程的自動化。我們已經在最大的業務領域做到了這一點，我們致力於在這一類別中繼續這樣做，不允許人們凌駕於比人類更聰明的系統之上。

If I could ask a follow-up, you've said in previous calls that you're going to be patient. And we want great experience for all customers before you open up. And you said, I think, 80% are having good outcomes, but maybe 20% aren't. And maybe the product performing around a B, things like this in the past.

如果我可以問後續問題，您在之前的通話中說過您會耐心等待。我們希望在您開業之前為所有顧客提供良好的體驗。您說，我認為 80% 的人取得了良好的結果，但可能有 20% 的人沒有。也許產品的表現在 B 左右，過去的情況就是這樣。

So October 1, obviously, you're still referral-based, so you're not completely open, but it's definitely a big step forward. So can we infer that it's largely improving performance that's informing that decision to go bigger? And if so, kind of how would you grade that? Or how would -- are you doing better than -- or any commentary on performance?

因此，10 月 1 日，顯然您仍然是基於推薦的，因此您還沒有完全開放，但這絕對是一個很大的進步。那麼，我們是否可以推斷，很大程度是為了提高效能才做出擴大規模的決定？如果是的話，您會如何評價呢？或者說，您的表現如何——比——有什麼評論嗎？

Look, last call, I laid out a list that I thought would take quarters, and our team knocked out most of that list in a single quarter. We got the Shopify app integration live. We don't announce these things, but you can find the app in the Shopify App Store. It creates one click integration.

瞧，最後一次，我列出了一份我認為需要幾個季度才能完成的清單，而我們的團隊在一個季度內就完成了清單上的大部分任務。我們已上線 Shopify 應用程式整合。我們沒有公佈這些事情，但您可以在 Shopify App Store 中找到該應用程式。它創建一鍵式整合。

We got dynamic product ads live that made a material uplift to the ability for the advertiser to run a video plus inventory ad to then transact the user in a very short amount of time. We did more in-depth integrations with the attribution companies, which we talked about last time. It was really important so that we can report the same way that the advertiser looks at data and then let the model interpret data that way.

我們推出了動態產品廣告，這大大提升了廣告商投放影片和庫存廣告的能力，從而可以在很短的時間內與用戶進行交易。我們與歸因公司進行了更深入的整合，我們上次談到了這一點。這非常重要，這樣我們就可以以與廣告商查看資料相同的方式進行報告，然後讓模型以這種方式解釋資料。

We're continuously improving the core underlying model. So as we looked at the advertisers that we have, the growth rate that they're seeing is substantial enough and the feedback that we're getting is strong enough that we have no reason to open up our platform as soon as we can now.

我們正在不斷改進核心底層模型。因此，當我們審視我們擁有的廣告商時，他們看到的成長率已經足夠可觀，我們得到的回饋也足夠強烈，因此我們沒有理由現在盡快開放我們的平台。

Clark Lampin, BTIG.

克拉克·蘭平（Clark Lampin），BTIG。

Thanks, good evening. I have -- obviously, I guess, some follow-ups on the self-serve launch. But I wanted to see, I guess, as you guys are thinking about taking on more of the customers that are in backlog, onboarding new sort of cohorts and verticals next year.

謝謝，晚上好。顯然，我想，我對自助服務推出做了一些後續工作。但我想看看，當你們考慮接納更多積壓的客戶時，明年是否會接納新的群體和垂直產業。

In the near term, is there any significant difference, I guess, if we think about the customers that are coming on from a size or product vertical standpoint and then from -- as we think about, I guess, the referral dynamic. Does that have any impact on expenses or sort of margins? Will those customers be given credits or any sort of compensation for bringing on, I guess, peers?

我想，在短期內，如果我們從規模或產品垂直的角度來考慮客戶，然後再考慮推薦動態，是否有顯著的差異。這對費用或利潤有影響嗎？我猜，這些客戶會因為帶來同行而獲得信用或任何形式的補償嗎？

Yeah. Second one first, Clark. I mean, obviously, the incremental value of a dollar in our business is worth a ton. So if we decide to pay a referral, you know that we like to make money, so I wouldn't expect you'll even see it in the margin number. It's just the business is way too big.

是的。第一個是第二個，克拉克。我的意思是，顯然，我們業務中一美元的增量價值是巨大的。因此，如果我們決定支付推薦費，您知道我們喜歡賺錢，所以我不會指望您會在保證金數字中看到它。只是生意太大了。

But it's very likely that what we look at is what I mentioned earlier. The chance for our clients to tell their peers to come on our platform while we've been exclusive for this long is a perk in itself. So we're not going to see much of a need to go pay for that client acquisition. And then on the first question, if you don't mind, quick recap.

但很有可能我們看到的就是我之前提到的那些。我們的客戶有機會告訴他們的同行加入我們的平台，而且我們已經獨家合作了這麼久，這本身就是一個福利。因此，我們認為沒有必要為獲取客戶而付出太多的代價。然後關於第一個問題，如果您不介意的話，請快速回顧一下。

Customer profile, yes, it will vary because we're going to open up to much more count. So, so far, we've gone through a period where, at first, I think we were constrained to $25 million, $30 million of GMV. We dropped it down for a little bit. We raised it up this quarter to $100 million-plus to really constrain what's coming on the platform.

客戶資料，是的，它會有所不同，因為我們將開放更多的數量。所以，到目前為止，我們經歷了這樣一個時期，起初，我認為我們的 GMV 被限制在 2500 萬美元、3000 萬美元。我們把它放低了一點。本季我們將這筆金額提高到 1 億美元以上，以真正限制平台上的內容。

As we open up, the goal that we have is to see that any small business of any type can market on our platform. If we clear that goal, we know we're going to be a product that can be very important to economies and job creation around the world. That's our goal. Now if we don't see that, we're going to work on that. But we think the product is ready for that.

隨著我們開放，我們的目標是看到任何類型的小型企業都可以在我們的平台上進行行銷。如果我們明確了這個目標，我們知道我們的產品將對世界各地的經濟和創造就業機會非常重要。這就是我們的目標。現在，如果我們沒有看到這一點，我們就會努力解決。但我們認為該產品已經做好準備。

We've seen cases of regardless of size, customers just plug in and it works. They can find their customers. They can measure the way they need, and they get the result that they would expect to grow their business. And so if we see that in this referral state, then we're going to be quickly on our way to opening up the business and really pushing it hard.

我們見過這樣的情況，無論大小，客戶只需插入電源即可工作。他們可以找到他們的客戶。他們可以按照自己需要的方式進行衡量，並獲得他們期望的業務成長結果。因此，如果我們看到這種推薦狀態，那麼我們將迅速開展業務並大力推動它。

Okay. And then not that this is something that's, I guess, a huge priority right now. But in terms of, I guess, sort of long-run supply expansion, I think you guys, as you have more model improvements, will unlock supply across the existing gaming footprint. But when you think, I guess, maybe years down the road, where would it make sense? I guess is there a way to sort of expand MAX sort of logically outside of the gaming ecosystem?

好的。但我想，這並不是現在的首要任務。但就長期供應擴張而言，我認為隨著模型的更多改進，你們將在現有遊戲領域解鎖供應。但我想，當你思考，也許幾年後，它還會有什麼意義呢？我猜想，有沒有辦法在遊戲生態系統之外以某種邏輯方式擴展 MAX？

Yeah. I don't think it's going to be years down the road if we're able to go get the demand that is broad reaching and drive up the client count substantially over the coming quarters. Pretty quickly, we're going to look at new supply sources as well. There's absolutely no reason why we wouldn't want to plug into other properties even if they're large social networks, music apps, news apps, sports apps, websites. The audience itself, again, this gamer audience is a human being that's doing a whole bunch of other things, and games are probably somewhere in the neighborhood of 10% to 15% of their time spent on mobile.

是的。我認為，如果我們能夠滿足廣泛的需求並在未來幾季大幅增加客戶數量，那麼這不會花費數年時間。很快，我們也將尋找新的供應來源。我們完全沒有理由不想接入其他資產，即使它們是大型社交網路、音樂應用程式、新聞應用程式、體育應用程式或網站。再次強調，就玩家本身而言，這些遊戲玩家也是在做很多其他事情的人，而遊戲大概佔他們在行動裝置上花費時間的 10% 到 15%。

And if we can access them outside of the largest-walled gardens that won't let us in everywhere else, if we know them and we have data on them, we want to show that ad to them every chance we get and we think that will be a very lucrative transaction moment for us and the advertiser, we're going to go after that. And so I wouldn't expect that that's a years out into the future initiative.

如果我們能夠在最大的圍牆花園之外接觸到他們，如果我們了解他們並且擁有他們的數據，我們就會抓住一切機會向他們展示廣告，我們認為這對我們和廣告商來說都是一個非常有利可圖的交易時刻，我們會追求這一點。因此，我不認為這是一個需要多年時間才能實現的計劃。

James Heaney, Jefferies.

詹姆斯希尼，傑富瑞集團。

Great, thank you guys. Adam, have you seen any changes in overall user acquisition spend from gaming companies post the Apple versus Epic lawsuit? I mean, conceptually, it should be a great tailwind for your business but interested if you're seeing any benefits playing out currently.

太好了，謝謝大家。亞當，在蘋果與 Epic 的訴訟之後，您是否發現遊戲公司的整體用戶獲取支出有任何變化？我的意思是，從概念上講，它應該對您的業務產生很大的推動作用，但如果您目前看到任何好處，我會很感興趣。

Not yet. I think I mentioned this on the last call, but we sort of expect this one to be -- take longer than people expect. Certainly, some apps are bypassing the App Store now to cut that rate down. But the biggest gaming companies tend to move really slowly and tend to operate in fear of the big platforms.

還沒有。我想我在上次通話中提到過這一點，但我們預計這次會比人們預期的花費更長的時間。當然，一些應用程式現在正在繞過 App Store 來降低這一比率。但最大的遊戲公司往往行動非常緩慢，並且往往因懼怕大型平台而運作。

And so in order to really do it, I think it's going to take a few quarters for them to optimize the user experiences and go bank it. And then from there, you'll start seeing it compound pretty quickly in terms of benefit to us as an ad platform because once the very large leaders start doing it, you'll start seeing the smaller to mid-sized ones really pick it up quickly.

因此，為了真正做到這一點，我認為他們需要幾個季度的時間來優化用戶體驗並將其付諸實踐。然後從那時起，你就會開始看到它作為廣告平台為我們帶來的利益迅速增加，因為一旦大型領導者開始這樣做，你就會開始看到中小型公司很快跟上。

So no impact yet. And I would guess it will probably take two to four quarters from some impact. And by four to eight quarters, you're going to get pretty material impact in pricing on our platform.

所以目前還沒有影響。我估計可能需要兩到四個季度才會產生影響。到四到八個季度後，我們平台的定價就會受到相當大的影響。

Great. And then maybe just another one on capital allocation. I mean, looks like without the apps business you're generating a 60% plus free cash flow margin so just interested to hear how you're thinking about use of the cash.

偉大的。然後也許還有另一個關於資本配置的問題。我的意思是，看起來即使沒有應用程式業務，您也能產生 60% 以上的自由現金流利潤率，所以我很想知道您是如何考慮使用現金的。

Yeah. I mean the approach that we're going to take, James, going forward is very consistent with what we've done in the past. I mean, obviously, first, continuing to allocate capital towards organic initiatives, continuing to hire on very high-quality engineering and business development talent, right, to help continue to grow the organic business and then after that, then to continue to return capital to shareholders via share buybacks. So that's the approach going forward as well.

是的。詹姆斯，我的意思是，我們今後採取的方法與我們過去的做法非常一致。我的意思是，顯然，首先，繼續將資金分配給有機計劃，繼續聘請高素質的工程和業務開發人才，以幫助繼續發展有機業務，然後繼續透過股票回購向股東返還資本。這也是我們今後採取的方法。

Martin Yang with Opp Co.

Opp Co. 的 Martin Yang

Hi, thank you for taking my question. One question on international expansion. Can you give us a sense of how do you view the size of the US market versus international when it comes to your ideal target customer in the near term? And then I have a follow-up.

你好，謝謝你回答我的問題。關於國際擴張的一個問題。您能否告訴我們，就近期的理想目標客戶而言，您如何看待美國市場與國際市場的規模？然後我有一個後續問題。

Yeah. I mean, one, let's divide it into two things. One, there's international businesses like local shop in Japan versus domestic US businesses. Second, there's the traffic.

是的。我的意思是，首先，讓我們把它分成兩件事。一方面，有國際業務，例如日本的本地商店，也有美國國內業務。其次，是交通問題。

And for us, historically, because we've grown word of mouth, getting customers that were local to markets was always tricky. So it was much easier for us to get companies that were international buyers. And the revenue was driven by where is the audience and how monetizable is the audience.

對我們來說，從歷史上看，由於我們依靠口碑傳播，因此獲得本地市場的客戶一直很棘手。因此，我們更容易找到國際買家公司。而收入則取決於觀眾在哪裡以及觀眾的貨幣化程度。

And so to date, I want to say this off the top of my head, and Matt can correct if I'm wrong, but it's roughly half-half on domestic US versus rest of world. And remember, we don't tend to operate inside China. And so you start with that.

到目前為止，我想脫口而出地說一下，如果我錯了，馬特可以糾正，但美國國內和世界其他地區的比例大致各佔一半。請記住，我們通常不在中國境內開展業務。所以你就從那裡開始吧。

You go, okay, well, we have -- and the only disclosure we've given you on the e-commerce vertical, that $1 billion run rate, I look at that and go, well, if we open up all these other countries for the most part, is $1 billion going to become $2 billion overnight. Probably not because most of these western companies can't go out and buy Japan and Korea and things that are really localized in language. But there's a lot of markets that look and feel similar to the US.

你說，好吧，我們有——我們在電子商務垂直領域給你的唯一披露是 10 億美元的運行率，我看著這個，然後想，如果我們在很大程度上開放所有這些其他國家，10 億美元將在一夜之間變成 20 億美元。可能不是，因為大多數西方公司無法出去購買日本和韓國的產品以及真正語言在地化的產品。但有許多市場看起來和感覺起來與美國相似。

So once we open it up, one is going to become much more than one. We just don't know where it's going to land. And then as we open up the platform, that local Japanese company should be able to come into our platform.

因此，一旦我們打開它，一個就會變得不只一個。我們只是不知道它會降落到哪裡。然後，隨著我們開放平台，日本當地公司應該能夠進入我們的平台。

And like we said, if we don't see them organically coming in, we're going to market to them and ensure they're coming in. The one in Korea is going to come on the platform, and we'll start getting penetration in every one of the markets that our users are in.

就像我們所說的那樣，如果我們沒有看到他們自然而然地加入，我們就會向他們推銷並確保他們加入。韓國市場即將登陸該平台，我們將開始滲透到我們用戶所在的每個市場。

And that incremental advertiser in these categories, as I touched on, the revenue from that is worth so much to us that we have the type of financial model that will allow us to go get coverage all over the world.

正如我所提到的，這些類別中的增量廣告商為我們帶來了巨大的收入，因此我們擁有一個財務模型，可以讓我們覆蓋全球。

Got it. My follow-up question on that is on the pace of onboarding international customers, do you see a pent-up demand among those? So when it comes to pacing, should we see a sudden uptick similar to 4Q last year in the US or international onboarding will be more gradual in line with your overall rollout?

知道了。我的後續問題是，關於國際客戶的接納速度，您是否看到其中有被壓抑的需求？那麼，就節奏而言，我們是否應該看到與去年美國第四季類似的突然上升，或者國際入職是否會隨著您的整體推廣而更加循序漸進？

Yeah. I can't say I know. I mean, I haven't looked at the queue of domains. You'd have to infer where the company is based, and we don't ask like country of origin. But the reality is, as we open up the platform, the customers that we have inside gaming are global.

是的。我不能說我知道。我的意思是，我還沒有看過網域隊列。您必須推斷公司所在地，而我們不會詢問原產國。但事實是，隨著我們開放平台，我們的遊戲客戶將遍布全球。

So as they invite their peers, their peers are going to be wherever their headquarters are, right? So there's coverage all over the world. The companies that are live inside of e-commerce are predominantly US. So probably their peers are going to be in western markets. And so you'll have a mix of referrals going out. And I don't think it's going to be like everyone's focused on the user only in the US to invite. It's going to be broad reaching, and there's no constraints that we're going to be setting as we start opening the platform up.

因此，當他們邀請同行時，他們的同行就會來到他們的總部所在地，對嗎？因此其覆蓋範圍遍佈世界各地。活躍於電子商務領域的公司主要來自美國。因此他們的同行可能會在西方市場。這樣你就會有各種各樣的推薦。而且我不認為每個人都會只專注在邀請美國用戶。它的影響範圍將會很廣泛，而且當我們開始開放平台時我們不會設定任何限制。

Jim Callahan, Piper Sandler.

吉姆·卡拉漢、派珀·桑德勒。

Hi, thanks guys for taking the question. On e-commerce, you're working with a lot of different types of advertisers with different bidding goals and purchase windows. Kind of what have you learned through that process so far if that's informing the self-serve kind of tool set?

嗨，感謝大家回答這個問題。在電子商務中，您要與許多不同類型的廣告商合作，他們有不同的競價目標和購買窗口。如果這是為自助式工具集提供的訊息，那麼到目前為止，您透過這個過程學到了什麼？

I mean, it is much more fragmented when it comes to attribution and integration than the mobile app ecosystem. The mobile app ecosystem has two major MMPs, mobile measurement partners. We own one of them. So integration is pretty easy across the advertiser base, and everyone has the same attribution model. Everyone looks at things on a last click basis, and so everything is standardized.

我的意思是，在歸因和整合方面，它比行動應用生態系統更加分散。行動應用生態系統有兩個主要的 MMP，即行動測量合作夥伴。我們擁有其中之一。因此，跨廣告客戶群的整合非常容易，每個人都有相同的歸因模型。每個人都根據最後點擊來看待事物，因此一切都是標準化的。

When we got into web, not only did we have to contend with the fact that most of the media buying was happening on Meta. So everyone wanted things the way they looked at things on Meta. Secondarily, we had to contend with the fact that the space was completely fragmented.

當我們進入網路領域時，我們不僅要面對大多數媒體購買都發生在 Meta 上的事實。因此每個人都希望以他們在 Meta 上看待事物的方式看待事物。其次，我們必須面對空間完全碎片化的現實。

So this was one of the bigger lifts that we had to go accomplish over the last quarter or two, was do all those integrations that I laid out on the last call to get us ready so that we can go out and really open up the platform. We wouldn't be able to if every shop by shop came on to our platform and couldn't integrate one click. That's how they integrate everywhere else.

因此，這是我們在過去一兩個季度必須完成的較大任務之一，即完成我在上次電話會議上提出的所有整合工作，讓我們做好準備，以便我們能夠真正開放這個平台。如果每家商店都進入我們的平台並且無法一鍵整合，我們就無法做到這一點。這就是他們在其他地方進行整合的方式。

So we needed to have that deal struck and have that app integrated in their store and already approved and functional. We needed to ensure that we were integrated with the major attribution companies, so numbers line up. And the advertisers can see things the way they need, and also the model gets the data that it needs to be able to go optimize on their behalf.

因此，我們需要達成交易並將該應用程式整合到他們的商店中，並獲得批准並投入使用。我們需要確保與主要的歸因公司進行整合，以便數字保持一致。廣告商可以按照他們需要的方式看到事物，並且模型也可以獲得所需的數據，以便能夠代表他們進行最佳化。

So those things were the harder points. And we've gotten through it. We've seen obviously positive reception, which gives us the confidence to get going on our journey to open up the platform.

所以這些事情是比較困難的。我們已經渡過了難關。我們顯然看到了正面的反響，這給了我們信心，讓我們繼續開放平台的旅程。

Great. That's helpful. And just on core gaming in the quarter, anything to call out in terms of like model enhancements, tweaks, performance improvements?

偉大的。這很有幫助。就本季的核心遊戲而言，在模型增強、調整和效能改進方面有什麼值得關注的嗎？

Yeah. I mean nothing that's like double-digit step function. But look, the numbers are getting bigger now. So we're now in a place where we continue to put up big numbers every single quarter. Q2 isn't particularly strong seasonally in the category, but we still grew at a really healthy rate.

是的。我的意思並不是像兩位數階躍函數那樣的東西。但你看，現在數字越來越大了。因此，我們現在每個季度都會繼續取得巨大的業績成長。從季節角度來看，第二季該類別的表現並不是特別強勁，但我們仍然以非常健康的速度成長。

We continue to have iterative lifts, and we continue to deliver the advertisers a lot of value. So every quarter, they come back to reinvest more as they continue to see strong return on ad spend on our platform.

我們不斷進行迭代提升，並持續為廣告主提供大量價值。因此，每個季度，他們都會回來進行更多投資，因為他們繼續看到我們平台上的廣告支出獲得強勁回報。

So what gives us confidence is as our numbers are getting bigger and without something that's what we would call the next big model enhancement, we're still growing at really healthy rates. So everything else that we're talking about just creates this upside opportunity on the business, not only the incremental data flywheel from the new category, the new demand from the new category but also the fact that inside gaming, there's more improvement to our technology to come. And some of that improvement at times will be those model enhancement, substantial lifts that no one expects.

因此，讓我們充滿信心的是，隨著我們的數字越來越大，並且沒有我們所謂的下一個重大模型增強，我們仍然以非常健康的速度成長。因此，我們正在談論的其他一切都為業務創造了上行機會，不僅是來自新類別的增量資料飛輪、來自新類別的新需求，而且在遊戲內部，我們的技術還將得到更多改進。有時，這些改進將是那些模型增強，是沒有人預料到的實質提升。

Bernie McTernan, Needham & Company.

麥克特南 (Bernie McTernan)，Needham & Company。

Great. Thanks for taking questions. Incremental margins of the business are obviously pretty incredible. Should there be any changes that we expect as e-commerce expands whether from marketing to acquire advertisers or the economics if you're looking to plug into other supply sources, like this 80%, 90%, 100% incremental margins? Is that still the right way to think about the business?

偉大的。感謝您的提問。該業務的增量利潤顯然相當驚人。隨著電子商務的擴張，我們是否應該期待任何變化，無論是從行銷到獲取廣告商，還是經濟方面（如果您希望接觸其他供應源），例如 80%、90%、100% 的增量利潤率？這仍然是思考業務的正確方式嗎？

Well, the latter when we plug in will be net revenue reported, right? Because still, we're going to report revenue rec the same way. The former will be a cost line item. But as you've seen, we're really responsible with the dollars that we spend. One, you'll see the sales and marketing line, so we'll talk about it as we go and test it. Two, we'll test it.

好吧，當我們插入時，後者將是報告的淨收入，對嗎？因為我們仍然會以同樣的方式報告收入。前者將作為成本項目。但如您所見，我們對所花的錢非常負責。首先，您會看到銷售和行銷線，因此我們會在進行和測試時討論它。二，我們將對其進行測試。

And three, we're only going to spend the dollars on performance marketing if we're printing cash on the other side, and it's creating this growth upside that we expect. So it's going to be a cost line item that wouldn't have existed before that will and one that I think everyone who's a shareholder is going to love to see growing because it will imply we're going to make a huge spread on the LTV to cost of user acquisition on the other side.

第三，只有當我們在賺錢並且能夠創造我們預期的成長潛力時，我們才會在績效行銷上投入資金。因此，這將成為一項以前不存在的成本項目，我認為每個股東都會樂意看到它成長，因為這意味著我們將在 LTV 和用戶獲取成本之間實現巨大的利差。

Arsenije Matovic, Wolfe.

阿塞尼耶·馬托維奇，沃爾夫。

Thanks for taking my question. Adam, with Meta reintroducing the advanced mobile measurement and just supporting it, do you see that maybe as an opportunity to showcase like apps engines and incremental performance for e-commerce campaigns versus Meta?

感謝您回答我的問題。亞當，Meta 重新引入了先進的行動測量並對其進行支持，您是否認為這可能是一個機會，可以展示應用引擎和電子商務活動相對於 Meta 的增量性能？

And could your advertisers that are opting in drive stronger traction with e-commerce advertisers, especially with like the referral-based rollout in October? Is that timing of that rollout also still keeping that 10% of that revenue being e-commerce intact for the year? And then, Matt, just a quick follow-up after.

選擇加入的廣告主能否對電子商務廣告商產生更大的吸引力，尤其是在 10 月推出基於推薦的廣告之後？這個推出時機是否也能保證全年電子商務營收的 10% 維持不變？然後，馬特，接下來是簡短的跟進。

You blitzed those questions. They sort of ran together for me. So let's start with advanced mobile measurement, and then let's jump into the next one. So advanced mobile measurement, Meta -- I mean, we don't comment on other companies' measurement, integrations, but they've always been integrated as a self-attributing network with Adjust. Adjust is no different than AppsFlyer.

你一下子回答完了這些問題。對我來說，他們就像是一起奔跑的。因此，讓我們從高級移動測量開始，然後進入下一個。因此，先進的行動測量、Meta——我的意思是，我們不評論其他公司的測量、集成，但它們一直作為自我歸因網絡與 Adjust 進行集成。Adjust 與 AppsFlyer 沒有什麼不同。

We run that business completely separately. So they do support attribution models across all these companies. I don't think there's any overlap with our business on these integrations, at least not that we've ever seen in the past. And therefore, that's something that happens on a team that we don't closely manage and really runs their business to maximize the SaaS business that they have.

我們完全獨立地經營這項業務。因此他們確實支持所有這些公司的歸因模型。我認為這些整合與我們的業務沒有任何重疊，至少我們過去從未見過。因此，這種情況發生在我們未密切管理的團隊中，他們真正經營自己的業務以最大化其擁有的 SaaS 業務。

I think your second question was percentage of e-com revenue in the future, correct? Like we're starting at 10. Now we've said -- I think on the last earnings call, I think I said 0.5% market penetration. It's probably -- I mean we're certainly sub-1% of Meta reports over 10 million advertisers, and we're going after all businesses of all sizes. We don't have a lot of advertisers.

我認為您的第二個問題是未來電子商務收入的百分比，對嗎？就像我們從 10 點開始一樣。現在我們已經說過—我想在上次財報電話會議上，我說過市場滲透率為 0.5%。有可能——我的意思是，在 Meta 報告的超過 1000 萬廣告客戶中，我們肯定只佔不到 1%，而且我們針對的是各種規模的企業。我們沒有很多廣告商。

So the opportunity in that category, both in terms of advertiser count and in terms of scale of TAM is much greater than gaming. So as we go forward, if we're doing our job right and performance is as we would expect based on the data that we have so far, it's going to quickly become more and more of our business.

因此，無論從廣告主數量或 TAM 規模來看，該類別的機會都比遊戲大得多。因此，隨著我們不斷前進，如果我們正確地完成工作，並且根據目前掌握的數據，業績符合我們的預期，那麼它將很快成為我們的業務重點。

And will one day our business become 90% web-based advertising, 10% mobile gaming? I don't know. But what I think, because of the data flywheel and the system that benefits both sides and the improvements that we continue to have on gaming is that gaming is going to get bigger and that other category is going to get really big.

未來有一天我們的業務會變成 90% 網路為基礎的廣告，10% 是基於手機遊戲嗎？我不知道。但我認為，由於數據飛輪和使雙方受益的系統以及我們在遊戲方面不斷取得的改進，遊戲將會變得更大，其他類別也會變得非常大。

Got it. And then, Matt, just on the actual guide, so I think 5% sequential growth. Conservatism last quarter was 4%. Do we just say, hey, the incremental point that's kind of the UA spend tied from the studios being off book and getting some of that revenue? Is that a fair way to think?

知道了。然後，馬特，僅根據實際指南，我認為連續成長 5%。上個季度的保守主義率為 4%。我們是否只能說，嘿，增量點就是 UA 支出，它與工作室的帳外收入有關，並且會從中獲得部分收入？這是一種公平的想法嗎？

Yeah, I mean historically, what we were doing was trying to guide towards the component of growth, right, that we feel very confident in. That's very predictable, which is when you look at kind of the components of overall growth in our technology, right.

是的，從歷史上看，我們所做的就是試圖引導我們非常有信心的成長部分。這是非常可預測的，當你觀察我們技術整體成長的組成部分時就會發現。

It's the directed model enhancements that Adam mentioned, changes that the engineers are making. And then the ongoing reinforcement learning within the model and that ongoing reinforcement learning has kind of generally trended towards the 3% to 5% per quarter. So the difference between that and then the guide this quarter is the incremental uplift from the additional revenue that we're going to get from the Apps divestiture.

這就是亞當提到的定向模型增強，是工程師正在進行的改變。然後，模型中持續的強化學習以及持續的強化學習總體上趨勢是每季 3% 到 5%。因此，這與本季度指南之間的差異在於，我們將從應用程式剝離中獲得額外收入的增量提升。

Got it. Thanks.

知道了。謝謝。

Alec Brondolo, Wells Fargo.

富國銀行的亞歷克·布朗多洛。

Perfect. Thanks so much for the question. I think what everybody is trying to figure out is when self-service goes live on October 1 on a referral basis or perhaps broader general availability next year, there's some notable amount of latent demand that's waiting to get into the program, right, advertisers that have asked you to join but maybe meet the GMV thresholds or perhaps you didn't have the capacity to onboard in the past.

完美的。非常感謝您的提問。我認為每個人都想弄清楚的是，當自助服務於 10 月 1 日以推薦方式上線或明年可能更廣泛地普及時，會有大量潛在需求等待加入該計劃，對吧，那些邀請你加入但可能滿足 GMV 門檻或你過去可能沒有能力加入的廣告商。

Can you maybe help us understand how big that list of advertisers is? You mentioned a queue in response to a previous question. So it seems like there's some amount of demand that you've already identified and waiting to come into the program.

您能否幫助我們了解廣告商名單有多大？您在回答上一個問題時提到了隊列。因此看起來您已經確定了一定數量的需求，而這些需求正在等待進入該計劃。

Yeah. Look, the referral program is to make sure that our advertisers who find success on our platform thus far are effectively curating the next set of advertisers. And prior to that, it was our team doing that, and we were very, very restrictive. So we think our advertisers are going to cause an onboarding moment that will be multiples bigger than what we were manually curating.

是的。你看，推薦計畫是為了確保迄今為止在我們的平台上取得成功的廣告商能夠有效地策劃下一組廣告商。在此之前，我們的團隊也在這麼做，而且我們非常非常嚴格。因此，我們認為我們的廣告商將帶來比我們手動策劃的更大的入職時刻。

Now it's not necessarily true that we're going to take our queue that's built over the last year and just say, everyone you're in. They're still going to have to get invited to get into the platform. So it will be still curated onboarding.

現在，我們不一定要按照去年建立的隊列，然後只說「你們都在」。他們仍然需要受到邀請才能進入該平台。因此它仍將受到精心策劃。

The reality is like Q4 is going to end up being a fun quarter. You've got the advertiser cohort that we didn't have last Q4 that was growing in the quarter to the point where we reported huge numbers and then had huge numbers in Q1. But we're going to have those advertisers primed and ready to go for the full Q4.

事實是，第四季度最終將是一個有趣的季度。我們擁有一個廣告客戶群體，而我們在上個季度沒有這個群體，但這個群體在本季度不斷增長，以至於我們報告了巨大的數字，然後在第一季度也取得了巨大的數字。但我們會讓這些廣告商做好準備，迎接整個第四季。

We're going to have advertisers inviting their friends onto our platform in Q4, and we're going to be opening up international all at the same time. So there's going to be a lot of fun moment -- moments for us and our customers in this e-commerce or web-based category that will set sort of a new baseline for that business. And then obviously, then we will go through hopefully another inflection when we really truly open up the platform and try to get into a state where we're more stable long term.

我們將在第四季度邀請廣告商邀請他們的朋友加入我們的平台，同時我們將開放國際市場。因此，對於我們和我們的客戶來說，在這個電子商務或基於網路的類別中，將會有許多有趣的時刻，為該業務設定新的基準。然後顯然，當我們真正開放平台並努力進入長期更穩定的狀態時，我們將有望經歷另一個轉折點。

Yeah. And then maybe just one more if I could. You've kind of articulated a vision for the self-service ads manager where there's a lot of agentic capabilities built in, in the campaign workflow. Is that live on October 1? Or is October 1 a version of the product that has credit card, that has measurement, but then we add kind of the AI stuff later maybe?

是的。如果可以的話，也許再多一個。您已經闡明了自助廣告管理器的願景，其中在廣告活動工作流程中內建了許多代理功能。10 月 1 日有直播嗎？或者 10 月 1 日推出的產品版本具有信用卡功能，具有測量功能，但我們稍後可能會添加一些 AI 功能？

Yeah, parts. I mean, look, there are different levels of agents. There's an agent that can respond and answer questions and give you responses to those questions and help you along. That's sort of an onboarding widget that whether we want to call this AI or not, now AI has made it a lot more capable, is something that's probably going to be in present early in the release of this product.

是的，零件。我的意思是，你看，代理商有不同等級。有一位代理商可以回應和回答問題並為您提供這些問題的答案並幫助您。這是一種入門小部件，無論我們是否願意稱之為 AI，現在 AI 已經使其功能更加強大，它可能會在這個產品發布的早期出現。

The more complex level is an agent that does everything for you. You're an advertiser and you wonder how your campaign is doing, analyze the performance, talk to me about the ads I'm running, that's a complex agent task. But analysis task at the advertiser level is not trivial.

更複雜的級別是代理為您做所有事情。您是廣告商，您想知道您的廣告活動的效果如何，分析其效果，並與我討論我正在投放的廣告，這是一項複雜的代理任務。但廣告商層面的分析任務並不簡單。

That will be on the way, too. And then you've got a third form, and I wouldn't call this an agent. I call it tools, but I keep referencing this generative AI-based ad creative tools. We're going to give our customers the ability to automatically create advertisements, whether video or end card. What that will do is create way more diversity of advertisement on our platform for the model to go use to personalize ad to the other side and hopefully much higher response rate from the consumer.

這也將會實現。然後你就有了第三種形式，我不會稱之為代理。我稱之為工具，但我一直引用這種基於生成 AI 的廣告創意工具。我們將為客戶提供自動製作廣告的能力，無論是影片還是片尾廣告。這樣做的目的是在我們的平台上創造更多樣化的廣告，以便該模型能夠為對方提供個人化廣告，並希望獲得更高的消費者回應率。

In particular, small businesses don't have the resources to go build ads that adapt for our platform, given the ads that adopt for our platform tend to engage the user for 30 to 60 seconds instead of 3 to 6 seconds like they do on social. And so we want to give them those tools so that they can have the same capacity the larger brands do so that they can buy at the maximum capability that their business justifies on our platform.

特別是，小型企業沒有資源製作適合我們平台的廣告，因為適合我們平台的廣告往往會吸引用戶 30 到 60 秒，而不是像社群媒體那樣吸引用戶 3 到 6 秒。因此，我們希望為他們提供這些工具，以便他們能夠擁有與大品牌相同的容量，從而能夠在我們的平台上以符合其業務的最大容量進行購買。

All of these things will come and then we'll talk about the next tools behind it because no product in our world stops being iterative. We launch. We then add tools. We try to make the advertiser's life seamless and fully automated, and we will keep giving them tools that we think will benefit them, which in turn benefits consumers and us.

所有這些事情都會發生，然後我們再討論背後的下一個工具，因為在我們的世界中沒有任何產品會停止迭代。我們發射。然後我們添加工具。我們努力使廣告商的生活變得無縫且完全自動化，我們將繼續向他們提供我們認為對他們有益的工具，這反過來又使消費者和我們受益。

Thank you so much.

太感謝了。

And that concludes the question-and-answer session for this quarter. We thank you all for joining us today. Have a good afternoon.

本季的問答環節到此結束。我們感謝大家今天的參加。祝您下午愉快。