AppLovin Corporation (APP) 2026 Q2 法說會逐字稿

Welcome to AppLovin's earnings call for the second quarter ended June 30, 2026. I'm David Hsiao, Head of Investor Relations. Joining me today to discuss our results are Adam Foroughi, our Co-Founder and CEO; and Matt Stumpf, our CFO.

歡迎參加 AppLovin 截至 2026 年 6 月 30 日止第二季的財報電話會議。我是投資人關係主管 David Hsiao。今天與我一同討論公司業績的有共同創辦人暨執行長 Adam Foroughi，以及財務長 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 investor.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, our expected growth opportunities, the expected future financial performance of the company, and other future events. These statements are based on our current assumptions and beliefs, and we assume no obligation to update them except as required by law.

請注意，我們迄今為止的 SEC 申報文件，以及討論第二季表現的財務更新與新聞稿，均可於 investor.applovin.com 查閱。在今天的電話會議中，我們將作出前瞻性陳述，包括但不限於：我們平台未來的發展與觸及範圍、我們預期的成長機會、公司預期的未來財務表現，以及其他未來事件。這些陳述係基於我們目前的假設與信念，除法律要求外，我們不承擔更新之義務。

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 fiscal quarter ended March 31, 2026. Additional information may also be found in our quarterly report on Form 10-Q for the fiscal quarter ended June 30, 2026, which will be filed today.

我們的實際結果可能與預測結果存在重大差異。我們鼓勵各位查閱我們最近提交的 Form 10-Q（截至 2026 年 3 月 31 日止財政季度）的風險因素。更多資訊亦可見於我們截至 2026 年 6 月 30 日止財政季度的 Form 10-Q 季度報告，該報告將於今日提交。

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 reconciliations of our GAAP and non-GAAP financial measures in our earnings release and financial update available on our Investor Relations site. This conference call is being recorded, and a replay and transcript will be available for a period of time on our IR website.

我們也將討論非 GAAP 財務衡量指標。這些非 GAAP 指標並非意在優於或取代我們的 GAAP 結果。請務必查閱我們投資人關係網站上之財報新聞稿與財務更新中所提供的 GAAP 結果，以及 GAAP 與非 GAAP 財務衡量指標的調節表。本次電話會議將被錄音，並將於一段時間內在我們的 IR 網站提供重播與逐字稿。

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.

接下來我將把時間交給 Adam 與 Matt 做開場致詞，之後由主持人帶領進行問答。

Thanks, everyone, for joining us today. I'm going to get right to it. This quarter, we delivered almost $2 billion in revenue, which was just below the midpoint of our guidance range, and our adjusted EBITDA was just below the range. We've always managed this business with the goal of outperforming our own expectations. And this quarter, we fell short of that standard.

感謝各位今天加入我們。我就直接切入重點。本季我們實現了接近 20 億美元的營收，略低於我們指引區間的中點，而調整後 EBITDA 也略低於該區間。我們一向以超越自身預期為目標來經營這項業務。而本季，我們未能達到這個標準。

What matters is that we know what happened, and it's already been addressed. Q3 is off to a strong start, and the business is back on the trajectory we expect. Let me explain. Gaming is still the majority of our revenue, and the single biggest driver of its growth is model performance. When our models improve, advertisers can profitably deploy more spend at their target return on ad spend goals, and budgets naturally step up.

重要的是，我們知道發生了什麼，而且已經處理完畢。第三季開局強勁，業務已回到我們預期的軌道上。我來說明一下。遊戲仍占我們營收的大宗，而其成長的最大單一驅動因素是模型表現。當我們的模型改善時，廣告主就能在其目標廣告投資報酬率（ROAS）下更有利可圖地投入更多預算，預算也會自然上調。

This quarter came down to timing. Our pace of meaningful model improvement was lighter than normal during the quarter, and the next step-up in model performance landed just after quarter end. Importantly, nothing we saw suggested weakening advertiser demand or a change in the competitive environment. In fact, MAX publisher earnings grew double digits quarter over quarter, and our share of publisher waterfalls remained consistent. With those improvements now live and heading into what is a seasonally stronger part of the year, the business is reaccelerating.

本季主要是時間點的問題。本季期間，我們在具實質意義的模型改進方面的節奏比平常更輕，而下一次模型表現的躍升發生在季末之後不久。重要的是，我們所觀察到的情況並未顯示廣告主需求轉弱或競爭環境出現變化。事實上，MAX 發布商收益較前一季呈雙位數成長，而我們在發布商瀑布流（waterfalls）中的占比也維持一致。隨著這些改進現已上線，且進入一年中季節性更強的時段，業務正在重新加速。

Now let me talk about consumer, which had an outstanding quarter. Advertiser spend set another record, finishing 28% above Q4 2025 levels. And remember, Q4 is the seasonal peak for these advertisers. Growing well past peak season levels in a seasonally slow quarter tells you how steep this curve is. Consumer isn't yet large enough to fully smooth a quarter like this, but that will change as we continue to ramp up our consumer business every quarter.

接著談談消費者（consumer）業務，本季表現非常出色。廣告主投放金額再創新高，較 2025 年第四季水準高出 28%。請記得，第四季是這些廣告主的季節性高峰。在季節性偏淡的一季仍能大幅超越旺季水準，足以說明這條成長曲線有多陡峭。消費者業務目前規模尚不足以完全平滑像本季這樣的波動，但隨著我們每一季持續擴大消費者業務，這點將會改變。

Stepping back, I want to put our long-term growth in context, how we think about the next decade. We built gaming into a far larger business far faster than we believe possible, and gaming keeps improving. What consumer adds is runway. We run one auction across multiple advertiser categories, and every category we bring in extends the opportunity in front of us. Over the longer term, as we continue improving gaming and expanding consumer, we believe this business can compound at roughly 30% annually.

退一步來看，我想把我們的長期成長放在脈絡中，談談我們如何看待未來十年。我們把遊戲業務打造得更大、速度也更快，超出我們原本認為可能的程度，而遊戲業務仍在持續改善。消費者業務帶來的是更長的成長跑道。我們在多個廣告主類別上運行同一個拍賣機制，而每引入一個新類別，都會延伸我們面前的機會。從更長期來看，隨著我們持續改善遊戲並擴張消費者業務，我們相信這項業務可望以約 30% 的年複合成長率成長。

Now on EBITDA. Adjusted EBITDA grew to a bit over $1.6 billion, up more than 50% year over year. While this result lands just below our guidance range, the incremental investments were exactly where we believe they should be, in our technology.

接著談 EBITDA。調整後 EBITDA 成長至略高於 16 億美元，年增超過 50%。雖然該結果略低於我們的指引區間，但新增投資完全投入在我們認為應該投入的地方——我們的技術。

We've been investing in architectural changes that let us build more complex models, models that benefit far more from additional training compute. That investment includes additional compute spend on the model improvements now live in Q3. And every dollar of it is dependent on return.

我們一直在投資架構層面的變更，使我們能建立更複雜的模型，而這些模型能從額外的訓練算力中獲得更大的效益。這項投資也包含在第三季已上線的模型改進所需的額外算力支出。而其中每一美元的投入都以回報為前提。

When additional compute produces substantially more revenue through better model performance, that's a trade we'll make every day. These higher training and inference costs are built into our guidance for next quarter.

當額外算力能透過更好的模型表現帶來顯著更多的營收時，這樣的取捨我們每天都願意做。這些較高的訓練與推論成本已納入我們對下一季的指引之中。

During the quarter, we also opened up our platform to the public under its original name, AppLovin Ads Manager. As we said last quarter, we never expected the public launch to change the business overnight. We're sequencing deliberately mid-market advertisers first, where the platform performs best today, with the long tail unlocking as our data compounds, the same way gaming developed. We'll execute on this through partnerships, and you'll see us continue to invest there.

本季期間，我們也以其原始名稱 AppLovin Ads Manager 將平台向大眾開放。如同我們上季所說，我們從未預期公開上線會在一夜之間改變業務。我們會有節奏地先導入中型市場（mid-market）廣告主，因為平台目前在該族群的表現最佳；隨著資料累積、長尾（long tail）將逐步解鎖，方式與遊戲業務的發展相同。我們將透過合作夥伴關係來推進這項工作，而你們也會看到我們持續在此投入。

Before I close, here's what we're focused on as a team: first, improving our core models, which is the primary driver of our near-term growth; second, advancing the architectural work that lets us benefit more from scaling compute, which we believe unlocks meaningfully larger gains over time; third, improving our creative tools and ad formats so advertisers can optimize their use of our platform to achieve even better outcomes; fourth, bringing more high-quality advertisers onto the platform through strategic partnerships.

在我結束之前，以下是我們團隊的重點：第一，改善核心模型，這是我們短期成長的主要驅動因素；第二，推進架構工作，使我們能更充分受益於算力擴張，我們相信這將隨時間推移解鎖顯著更大的增益；第三，改善我們的創意工具與廣告格式，讓廣告主能最佳化其平台使用方式，以取得更好的成效；第四，透過策略性合作夥伴關係，將更多高品質廣告主帶上平台。

Let me close with this. We've spent years building an advertising platform whose economics improve as our models improve. Advertiser demand is healthy. Our models continue to improve. Consumer is scaling rapidly. Nothing we saw this quarter changed our conviction in the long-term opportunity ahead.

最後我想用這段話作結。我們花了多年打造一個廣告平台，其經濟效益會隨著模型改善而提升。廣告主需求依然健康。我們的模型持續改善。消費者業務正在快速擴張。本季我們所看到的一切，都沒有改變我們對長期機會的信念。

With that, I'll turn it over to Matt to walk through the financials.

接下來我把時間交給 Matt，請他帶大家回顧財務表現。

Thanks, Adam, and thanks to everyone for joining us today. This quarter is a good example of the underlying strength of our business. As Adam mentioned, we didn't get the same level of model uplift we've seen in recent quarters, but we still delivered second quarter revenue of $1.92 billion, growing 53% year over year and 4% sequentially, driven by the core gaming business and continued scaling in the consumer vertical.

謝謝你，Adam，也感謝各位今天加入我們。本季是一個很好的例子，展現我們業務的內在韌性與實力。如 Adam 所提到的，我們沒有獲得近幾季所見到的同等幅度模型提升，但我們仍繳出第二季 19.2 億美元的營收，年增 53%、季增 4%，主要由核心遊戲業務帶動，並受惠於消費者垂直領域的持續擴張。

Adjusted EBITDA was $1.61 billion, up 58% year over year, with margins expanding approximately 300 basis points from the same period last year. Quarter-over-quarter flow-through to adjusted EBITDA was 70%. The primary driver of the sequential increase in costs was higher compute associated with training our existing models and with new model development. And that higher compute run rate is reflected in our outlook. We manage this business to EBITDA dollars and free cash flow rather than to a margin percentage, and we'll continue to deploy dollars when we see an opportunity to produce more revenue.

調整後 EBITDA 為 16.1 億美元，年增 58%，利潤率較去年同期擴張約 300 個基點。調整後 EBITDA 的季對季流轉率為 70%。成本環比上升的主要驅動因素，是與訓練我們既有模型及開發新模型相關的更高算力成本。而這個更高的算力運行水準也反映在我們的展望中。我們管理這項業務時更看重 EBITDA 金額與自由現金流，而非利潤率百分比；當我們看到有機會帶來更多營收時，將持續投入資金。

Free cash flow for the quarter was $863 million. As I previewed on the last quarter's call, conversion was below our normal cadence in the second quarter due to the timing of international cash tax and interest payments. This is a timing dynamic, not a change in the earnings power of our business. We expect free cash flow conversion to improve in the third quarter and to normalize to roughly 75% of adjusted EBITDA for the full year. We ended the quarter with $3.05 billion of cash against $3.7 billion of total debt, which puts net leverage at approximately 0.1 times trailing 12-month adjusted EBITDA, well within the approximately 1 times where we'd expect to operate over the long term.

本季自由現金流為 8.63 億美元。如我在上季電話會議中預告，第二季因國際現金稅與利息支付時點的影響，轉換率低於我們的正常節奏。這是時點因素，而非我們業務獲利能力的改變。我們預期第三季自由現金流轉換率將改善，並在全年回歸至約為調整後 EBITDA 的 75%。本季末我們持有 30.5 億美元現金，對應 37 億美元總債務，使得淨槓桿約為過去 12 個月調整後 EBITDA 的 0.1 倍，遠低於我們長期預期約 1 倍的營運水準。

During the second quarter, we repurchased and withheld a total of approximately 1.14 million shares for $551 million and ended the quarter with 335 million shares outstanding and approximately $1.8 billion remaining under our share repurchase authorization. Our choice to moderate the pace of our buybacks this quarter relative to the roughly $1 billion we have deployed in the first quarter reflects consideration of our lower free cash flow during the quarter and does not indicate a change in conviction or in how we intend to use the authorization going forward.

第二季期間，我們回購並扣留合計約 114 萬股，金額為 5.51 億美元；季末流通在外股數為 3.35 億股，且在股票回購授權額度下尚餘約 18 億美元。本季我們選擇相較第一季已投入約 10 億美元而放緩回購步伐，反映的是對本季較低自由現金流的考量，並不代表我們的信心或未來使用該授權的方式有所改變。

One other item before I turn to our outlook, we continue to get questions on the reported SEC inquiry. So let me close the loop. It was a voluntary request, which we never deemed material. The SEC has recently advised us that it concluded its inquiry with no recommended action. We're pleased to have it resolved.

在我轉到展望之前，還有一點：我們仍持續收到關於已報導之 SEC 調查的提問。所以讓我把這件事交代清楚。那是一項自願性資訊請求，我們從未認定其具重大性。SEC 近期已告知我們，其已結束調查且未建議採取任何行動。我們很高興此事已獲解決。

Turning to our outlook for the third quarter of 2026. We expect revenue between $2.055 billion and $2.085 billion, representing 46% to 48% year-over-year growth or 7% to 8% sequentially. We expect adjusted EBITDA between $1.71 billion and $1.74 billion, representing 48% to 50% year-over-year growth with an adjusted EBITDA margin of approximately 83%. That outlook reflects the model improvements that are already live and performing, continued scaling in our consumer vertical, normal seasonality, and the higher training and compute costs I mentioned. It does not assume additional model releases that have not yet been deployed.

接著談我們對 2026 年第三季的展望。我們預期營收介於 20.55 億至 20.85 億美元，代表年增 46% 至 48%，或季增 7% 至 8%。我們預期調整後 EBITDA 介於 17.1 億至 17.4 億美元，代表年增 48% 至 50%，調整後 EBITDA 利潤率約為 83%。該展望反映了已上線並表現良好的模型改進、我們在消費者垂直領域的持續擴張、正常季節性，以及我提到的較高訓練與算力成本。其中不包含尚未部署的額外模型發布。

To close, this was not the quarter we hold ourselves to, and we've been direct with you about why. What hasn't changed is the shape or strength of this business, nearly $2 billion of quarterly revenue, growing better than 50% year over year, margins above 80%, and extremely strong cash generation. And with the next step-up in model performance already live, our outlook reflects the business getting back to the trajectory we expect.

最後總結，這一季並非我們對自己要求的水準，我們也已直接向各位說明原因。不變的是這門業務的形態與強度：單季接近 20 億美元營收、年增超過 50%、利潤率高於 80%，以及極為強勁的現金創造能力。而隨著下一階段的模型效能提升已經上線，我們的展望反映業務正回到我們預期的成長軌跡。

With that, let's move to Q&A.

接下來，我們進入問答環節。

(Operator Instructions) Jason Bazinet, Citi.

(接線員指示) Jason Bazinet，花旗。

Maybe I've missed it, but I think this is the first time I've heard you talk about partners to bring in more customers. Can you just expand on the partnership opportunity that you see?

也許我漏聽了，但我想這是我第一次聽到你們談到透過合作夥伴帶來更多客戶。你能否再多談談你們看到的合作機會？

Yeah. Thanks, Jason. We have done a couple of deals so far with third-party companies, one of the larger analytics companies in the market in e-commerce. And we found that if we go to the source that works with these companies on the other side, the advertisers that we want, it's a more targeted way to get the right kinds of advertisers onto our platform. So rather than just start by buying ads and bringing in long tail, which is harder to make work with where the model evolution is today, we're going through these partnerships to try to get very targeted customers into the platform.

是的。謝謝你，Jason。到目前為止，我們已與第三方公司做了幾筆合作，其中一家是電商領域市場上較大型的分析公司。我們發現，如果我們從與這些公司合作的源頭切入，也就是我們想要的廣告主所在的那一端，這是一種更精準的方式，能把合適類型的廣告主帶到我們的平台上。因此，相較於一開始就買廣告、引入長尾客戶（以目前模型演進的狀態更難奏效），我們透過這些合作夥伴關係，嘗試把非常精準的客戶導入平台。

So this is like Triple Whale?

所以這像是 Triple Whale 那樣？

You sort of nailed it with what you're saying.

你說得很到位。

James Heaney, Jefferies.

James Heaney，Jefferies。

Great. I just wanted to get under the hood of the gaming advertising business. And maybe it's difficult to explain it in a simple response, but just curious exactly what the sort of model breakthroughs that you were looking for that didn't happen. Just trying to understand like the timing of it and kind of what happened in the quarter, just to --

很好。我想更深入了解遊戲廣告業務的細節。也許很難用簡短回答說清楚，但我想知道你們原本在尋找、卻沒有發生的那些模型突破究竟是什麼。我只是想理解它的時點，以及本季到底發生了什麼，來--

Yeah. Thanks, James. If you go back the last 12 quarters, every quarter, we've had really good growth, except for Q2 in single digits. I think every other quarter since Axon 2 has been double digits. Now Q2 starts as the seasonally weakest quarter.

好的。謝謝你，James。如果你回顧過去 12 個季度，每一季我們都有非常好的成長，只有第二季是個位數。我想自 Axon 2 以來，其他每一季都是雙位數。而第二季本來就是季節性最弱的季度。

But every quarter that we've had these outsized growth rates, we've had improvements to our model. So our team isn't just sitting there doing nothing for the last 12 quarters. It's constant improvement. We just don't really break it out.

但我們每一季能有這些超常的成長率，背後都伴隨著模型的改進。所以我們的團隊過去 12 個季度並不是坐在那裡什麼都不做。而是持續不斷地改進。只是我們沒有特別把它拆分出來說明。

So if you recall, our gaming growth model we had given you all back in the day included one sizable improvement per year, but constant improvement to the models as well as on an ongoing basis. In this case, in Q2, we didn't have the same amount of uplift that we normally have in any other prior quarter. That came right after the quarter. So that's why Q3 has started really well, and that's why we've guided strong going forward.

如果你還記得，我們過去給各位的遊戲成長模型，包含每年一次較大的改進，同時也有持續性的模型優化。在這次的第二季，我們沒有像以往任何季度那樣獲得同等幅度的提升。那個提升是在季度結束後才出現的。所以第三季一開始表現非常好，也因此我們對後續給出了較強的指引。

Great. And then just one more question, if I can. I think last quarter, you called out reaching sort of record revenue in the consumer vertical in April. Curious if you can provide an update on the progress that you saw in the second quarter? How much of that growth was just built on existing customers versus maybe what you saw from new advertisers?

了解。如果可以的話再問一題。我記得上季你提到消費者垂直領域在 4 月達到某種程度的創紀錄營收。想請你更新一下第二季你們看到的進展？這些成長有多少是建立在既有客戶之上， versus 你們從新廣告主那邊看到的貢獻？

Yeah. I mean, new advertisers are beneficial, but the base of the business is already pretty good. And so new customers aren't going to go live and really drive impact to that growth rate. So when we say double -- I think I said 26% growth in the talk track over Q4, that's a substantial uptick in this category. Usually, e-commerce, Q1 dips a lot from Q4. Q2 then has to trail back up, and then Q3 and Q4 really build on the year with half the dollar spent in Q4.

是的。我的意思是，新廣告主確實有幫助，但這門業務的基礎本來就相當不錯。因此，新客戶不會一上線就真正對那個成長率帶來很大影響。所以當我們說雙位數——我想我在講稿裡提到相較 Q4 成長 26%，那在這個類別中是相當顯著的上升。通常電商在 Q1 會比 Q4 下滑很多。Q2 接著必須回升，然後 Q3、Q4 才會在全年基礎上繼續累積，而 Q4 的花費金額約占全年一半。

And so having that kind of growth in Q2 implies that the customers that we have on the platform are seeing a lot of success.

因此，第二季出現那樣的成長，意味著我們平台上的客戶正獲得很大的成功。

28%.

28%。

28% to be correct.

正確來說是28%。

Stephen Ju, UBS.

Stephen Ju，瑞銀（UBS）。

So I think the feedback that we're getting from advertisers in the e-commerce segment seems to be that as they spend money on the platform, they're nowhere near hitting that sort of efficient frontier of ROAS ceilings. So I'm just wondering if we're kind of hearing from a positive biased group of folks or if that's what you're hearing overall from a broader group of the advertisers that you're talking to?

所以我認為，我們從電商領域的廣告主那裡得到的回饋似乎是：當他們在平台上花錢時，他們離觸及那種ROAS上限的效率前緣還差得很遠。所以我只是想知道，我們是不是主要聽到的是一群偏正面的受訪者，還是說這就是你們從更廣泛、你們正在接觸的廣告主整體所聽到的情況？

And secondarily, I think, Adam, you talked about starting with the midsized merchants and going long tail. And I think in the past, you had indicated some hesitation on working with some of the larger advertisers. I'm just wondering if that perspective has changed in the last three to six months or so as it seems like some of the mobile game advertisers are also pretty large.

其次，我想，Adam，你談到先從中型商家開始，然後延伸到長尾。而我記得過去你曾表示，對於與一些較大型廣告主合作有些猶豫。我想請問，過去三到六個月左右，這個看法是否有所改變？因為看起來一些行動遊戲廣告主也相當大。

Yeah. I mean, I'll go over the second one first, Stephen. So in gaming, our model is pretty mature, and we've got a lot of data. And so the architecture is just more developed. If you have a new game in the market today that hasn't even launched, and you go on to our platform, the return goals are hit really, really quickly. So there's almost no learning budget. The model is really sophisticated at nailing it for anything.

是的。我的意思是，Stephen，我先回答第二個問題。在遊戲領域，我們的模型相當成熟，而且我們有大量資料。因此整體架構也更完善。如果你今天有一款全新的遊戲，甚至還沒上市，就上到我們的平台，回報目標會非常、非常快就達成。所以幾乎不需要學習預算。這個模型非常精密，幾乎什麼都能精準做到。

In e-commerce or consumer, we're still early in evolution. We've got less data penetration in the market by a lot. We've got less sophisticated model at this point.

但在電商或消費者領域，我們仍處於演進的早期階段。我們在市場上的資料滲透率還低很多。目前模型也沒那麼成熟。

And so if there's a small shop that goes live, it's less likely that with very few dollars spent, it's going to hit their goals and they're going to scale. Some of them do, but in a lot of the smaller ones, they're not able to achieve it.

因此，如果是一家小店剛上線，花很少的錢就能達成目標並擴大規模的機率比較低。有些確實可以，但很多較小的商家做不到。

Now if there's a mid-market brand, they know that there's learning cost to marketing campaigns. So they'll invest more, get to their goals, and scale. So what we found is that mid-market is a sweet spot right now. As data builds, we have no reason to believe that tail is not going to be covered, too, given in gaming, we're able to support any type of game at any level.

但如果是中型市場（mid-market）的品牌，他們知道行銷活動會有學習成本。所以他們會投入更多，達成目標，然後擴大規模。因此我們發現，目前中型市場是甜蜜點。隨著資料累積，我們沒有理由相信長尾不會被涵蓋；因為在遊戲領域，我們能支援任何類型、任何規模層級的遊戲。

On your first question, if I understood it correctly, you're basically saying advertisers haven't reached their maximum amount of spend for the return that they're getting. And we're new -- yeah, we're a new platform, right? Like these companies, first of all, they like manage their budgets really slowly. We launched this product, I think it's now been probably 18 months or so.

至於你的第一個問題，如果我理解正確，你基本上是在說：廣告主在他們目前獲得的回報下，還沒有花到他們的最大可投入金額。而我們是新的——對，我們是一個新平台，對吧？像這些公司，首先，他們管理預算的節奏通常很慢。我們推出這個產品，我想現在大概已經18個月左右了。

And so you've got these companies that usually manage their budgets over somewhere in the neighborhood of one quarter to four quarters ahead. Most of their budget goes to social and search. We're deemed a new bucket, so a testing category. And to graduate up takes time. So this stuff compounds over time, over quarters and years.

因此，這些公司通常會以大約提前一季到四季的方式來管理預算。他們大部分預算都投向社群與搜尋。我們被視為一個新的預算桶，所以屬於測試類別。要升級到更高層級需要時間。所以這些事情會隨著時間、以季度與年度為單位累積發酵。

An advertiser that is live today, we would expect, for sure, given the results that they would be able to see, in 12 months, they're bigger; in 24 months, they're bigger. But the rate of improvement there shouldn't be organically appearing just like some of the more mature marketing platforms. It should be accelerated because they're probably indexing lower on us as they build a history and get to know the platform, invest more in the platform, get comfort in long-term results, run their incrementality studies, et cetera, that ramp-up should be faster.

對於今天已經上線的廣告主，我們很確定，基於他們能看到的成果，12個月後他們會更大；24個月後他們會更大。但那種改善速度不應該像一些更成熟的行銷平台那樣只是自然出現。它應該會更快，因為他們一開始在我們這裡的投入占比可能偏低；隨著他們建立歷史、熟悉平台、增加投入、對長期結果更有信心、做增量（incrementality）研究等等，這個爬坡速度應該會更快。

Ralph Schackart, William Blair.

Ralph Schackart，William Blair。

Just, Adam, maybe to start on the model improvements. I'm guessing you've kind of seen this before, but just any sense of why you didn't get the model uplift that you would have expected in the quarter? I don't know if there's a way to sort of diagnose that? And then maybe just two more, please. Any more you could share on maybe the rate of improvement that you're seeing post-quarter?

Adam，先從模型改善談起。我猜你以前可能也見過這種情況，但能否談談為什麼本季沒有出現你原本預期的模型提升？不知道是否有辦法去診斷原因？然後再兩個問題。你能否多分享一些季後你們看到的改善速度？

And then, Matt, as compute costs continue to increase in scale, how should we think about the margin profile of the business? So there's three in there.

然後，Matt，隨著算力成本持續擴大規模上升，我們應該如何看待公司的利潤率輪廓？所以這裡面有三個問題。

Okay. So I'll have Matt take two and three. I'll start with one. Look, it's R&D, right? Like there's no guarantee that we're always going to have lifts in every single period of three months. The team is always working on lifts.

好的。那我請Matt回答第二和第三個，我先回答第一個。你看，這是研發（R&D），對吧？並不能保證我們每一個三個月的期間都一定會有提升。團隊一直都在做提升。

And usually, in every single quarter, it's not one lift that we get. We compound multiple small lifts. The impact was just smaller in Q2, followed by a pretty material uplift in Q3 in the early part of Q3. And so that's just the reality of when you're building models. You're building models, you don't have a certainty on the impact of what you're testing, you're testing hypotheses, hoping for a good result.

而且通常每一季我們得到的也不會只是一個提升。我們會把多個小幅提升疊加起來。第二季的影響只是比較小，接著在第三季、第三季初期出現了相當顯著的提升。這就是建模的現實。你在建模型時，對於你正在測試的影響沒有確定性；你是在測試假設，期待得到好的結果。

And so the system is a whole bunch of A/B tests looking for lifts. And there are going to be periods where we don't get material lifts. There's going to be other periods where we have huge lifts that contribute to 12%, 13%, 15% Q over Q type quarters.

因此，系統就是一大堆A/B測試在尋找提升。會有一些期間我們拿不到顯著的提升。也會有其他期間出現很大的提升，進而貢獻像是季對季成長12%、13%、15%那樣的季度表現。

Yeah. In terms of the rate of improvement and what's included within the guide, Ralph. So within this quarter, what we've assumed is similar to kind of our guidance philosophy in the past where we have a high level of confidence. But the difference being for this one quarter, we've guided inclusive of the model improvements that we've launched that we're aware of and then also the compute cost increase that we've seen.

是的。至於改善速度以及指引中包含了什麼，Ralph。在本季，我們的假設與過去的指引哲學類似，也就是我們只納入我們高度有信心的部分。但不同的是，這一次我們的指引已經把我們已知、已上線的模型改善納入，同時也納入我們看到的算力成本上升。

To the margin question, over the longer term, we may see some variability, which is consistent with what we've told analysts and investors in the past as well that we may have some short-term fluctuation in the margin profile of the business, but it should be indicative and investors should look at that as a positive because when we're spending, we're very cautious about how we're spending, and we're only spending if we see that there is incremental revenue behind that.

關於利潤率問題，從較長期來看，我們可能會看到一些波動；這也符合我們過去對分析師與投資人所說的：公司的利潤率輪廓在短期可能會有起伏，但這應該具有指標性，投資人也應該把它視為正面訊號，因為我們在支出上非常謹慎，只有在我們看到背後有增量營收時才會花錢。

So when you see these short-term fluctuations, you should then realize that over time, we're going to continue to improve based on those cost increases. So over the longer term, we still have a very high level of confidence that we'll be within kind of the low 80% EBITDA margin.

所以當你看到這些短期波動時，你應該理解，隨著時間推進，我們會在那些成本上升的基礎上持續改善。因此，從長期來看，我們仍然非常有信心，EBITDA利潤率會落在大約80%出頭（low 80%）的區間。

Omar Dessouky, Bank of America.

Omar Dessouky，美國銀行。

So you opened your platform, I think, to everybody in late June. It's been about, I think, six weeks now since it was open. Can you tell us a little bit about how you would evaluate the performance of your onboarding in terms of the number of advertisers onboarded, whether it kind of met your expectations, exceeded, or was below? And how you expect that trajectory to -- of the number of advertisers to go for the rest of the year and into calendar '27?

所以我想你們在六月下旬把平台開放給所有人。自從開放以來，我想大約已經六週了。你能否跟我們談談，你會如何評估你們的導入（onboarding）表現——就已導入的廣告主數量而言，是否大致符合你們的預期、超出預期，或低於預期？以及你預期今年剩餘時間到 2027 年曆年，廣告主數量的成長軌跡會如何發展？

Specifically, should we be thinking about your consumer business as one where there are tens of thousands of advertisers over time that spend a little or a few advertisers that spend a lot, right? Because you said you're being targeted now. But I want to kind of try to understand how long you expect to be targeted as compared to just advertisers flocking to your platform? And what you need to do to get to that point where advertisers flock to your platform and it can become a very widely used platform like Meta and Google?

更具體地說，我們是否應該把你們的消費者業務視為：長期來看會有數以萬計、每家花費不多的廣告主；還是只有少數、但每家花費很多的廣告主？因為你說你們現在是在做定向開發。但我想試著理解，相較於廣告主自然湧入你們平台，你們預期還要定向開發多久？以及你們需要做什麼，才能走到那個廣告主會湧入你們平台、並且能像 Meta 和 Google 那樣成為非常廣泛使用的平台的階段？

Yes. It'd be the latter on your own words, Omar, that few advertisers contributing more today. So we look at the launches as expected. We said it's not going to be a big marketing push behind the launch. We're targeting the marketing relationships and the dollars that we spend to where we can go get those mid-market brands.

是的。用你的話來說，Omar，會是後者——也就是目前由少數廣告主貢獻較多。所以我們認為這次上線的表現符合預期。我們也說過，這次上線不會有大規模的行銷推廣。我們把行銷關係與我們投入的預算，鎖定在能夠拿下那些中型市場品牌的地方。

We need to execute on the bigger-ticket customers first. So not the very, very big because we'll get to the head of the market later and not the very, very small that would contribute count of advertisers, but be considered SMB or long tail, that's harder for us to make work right now. So that mid-market is what we're focused on. If we're able to execute there, which we have a lot of confidence in, that's what most of the customers in the consumer category spending today are, then you broaden it out. But that gets us more data coverage.

我們需要先把大額客戶執行好。所以不是最、最頂級的大客戶——我們之後會再切入市場頂端；也不是非常小、雖然能增加廣告主數量、但屬於中小企業（SMB）或長尾的客戶，因為那對我們目前而言比較難做出規模化的成效。因此我們聚焦在中型市場。如果我們能在那裡執行到位——我們對此很有信心——而且目前消費者類別中大多數在投放的客戶也正是這類型，那接著就能把範圍擴大。但這也會讓我們獲得更多資料覆蓋。

It gets us a time to get to a place where the model is more sophisticated because with time, we're going to build a better model, and it gives us time for a reputation to seep into the bigger parts of the market. Now remember, these systems take a long time to build out. You talked about Google and Facebook. They've been around two decades and a decade plus, right? It took us 14 years to be fully penetrated into the gaming category in terms of any customer in the world in mobile gaming today would be foolish not to spend dollars on our platform.

這也讓我們有時間把模型推進到更成熟的階段，因為隨著時間，我們會建立更好的模型；同時也讓我們的口碑有時間滲透到市場更大的區塊。要記得，這些系統需要很長時間才能建構完成。你提到 Google 和 Facebook。它們分別已經存在二十年、以及十年以上，對吧？我們花了 14 年才在遊戲類別中做到完全滲透——就全球行動遊戲而言，今天任何客戶若不在我們平台上投放預算，都會是很不明智的。

So it will take time to build this thing up, but we're seeing very quick growth because the customers that are on the platform are seeing a lot of success. And as I already touched on, they're not even at a ceiling of what they can spend given the results that they're seeing. The ad templates are going to improve over time. The model is going to improve over time. With every incremental mid-market customer, the data in the system improves.

所以把這件事做大需要時間，但我們看到非常快速的成長，因為已在平台上的客戶正在取得很大的成功。而且如我先前提到的，基於他們看到的成效，他們甚至還沒有達到可投放金額的上限。廣告模板會隨時間改善。模型也會隨時間改善。每增加一位中型市場客戶，系統中的資料就會變得更好。

And then over time, we're going to be able to broaden it out. We can't tell you when. That's just something that's not predictable.

然後隨著時間，我們就能把範圍擴大。我們無法告訴你會在什麼時候。那不是一件可預測的事。

Got it. So which specific parts of your technology would you say are maybe ahead of schedule and behind schedule? I know generative AI creatives was one that was talked a lot about in the last six months. But what other pieces of the technological development should -- we should try to keep an eye on to gauge your progress?

了解。那你會說你們技術的哪些特定部分可能進度超前、哪些部分落後？我知道過去六個月大家談了很多生成式 AI 創意。但還有哪些技術開發的其他環節，是我們應該關注、用來衡量你們進展的？

Yeah. I mean, look, we're early in a business that is pretty substantial at this point and growing very quickly. The reality with it is that we can't say what's behind or ahead of schedule. The requirement is get more customers, get more data, continue to write more sophisticated models, improve the ad template, and you get lifts for the current customers and every single new customer.

是的。我的意思是，你看，我們在一個到目前為止已相當可觀、而且成長非常快的業務上仍處於早期階段。現實是，我們無法說哪些落後或超前進度。我們的要求是：取得更多客戶、取得更多資料、持續撰寫更精密的模型、改善廣告模板，並為現有客戶以及每一位新客戶帶來成效提升。

Over time, that compounds as we get more success stories on the platform, more narratives out there, more agencies that are media buying agencies knowing that our platform works really well as the third sort of anchor in the marketplace. More customers are just naturally going to come. So that's sort of the formula to the future.

隨著時間推移，當平台上有更多成功案例、更多市場敘事、更多媒體採買代理商知道我們的平台作為市場上的第三個錨點運作得非常好，這些都會產生複利效應。更多客戶就會自然地進來。所以這就是未來的公式。

Rob Sanderson, Loop Capital.

Rob Sanderson，Loop Capital。

I've got two questions, both related to e-commerce. The funnel -- can you talk about the funnel, like you -- for folks coming through self-service, you said earlier in the year, you were getting some -- like 57% of qualified leads were coming through. And I think lack of creative was kind of a primary reason for leakage. But that was before you had tools for ad creation. So anything you could sort of share on how that dynamic might be changing now that you're maybe better equipped with some of that tool set?

我有兩個問題，都與電商相關。關於漏斗——你能談談漏斗嗎？例如走自助式（self-service）流程進來的人，你們今年稍早說過，有一些——像是 57% 的合格線索是從那裡進來的。我記得創意素材不足是流失的主要原因。但那是在你們還沒有廣告製作工具之前。所以你能否分享一下，現在你們在工具組上更完善之後，這個動態是否正在改變？

That's question one. And then a little bit on feedback from commerce advertisers. What's been working well with the self-service platform now that you're GA, what are some things you still need to improve upon or that you'd like to get better? And then are there sort of common requests for features like -- are folks looking for better targeting or better measurement or more campaign automation? Are there any sort of categoric things that are kind of next steps for you to go to work on the product side?

這是第一個問題。第二個是關於來自電商廣告主的回饋。你們自助式平台現在已全面可用（GA）之後，哪些地方運作得很好？還有哪些你們仍需要改善、或希望做得更好的地方？另外，是否有一些常見的功能需求——例如大家想要更好的定向、更好的衡量，或更多的活動自動化？在產品端，有沒有一些類別性的事項是你們接下來要著手的重點？

Yeah. Thanks for all the questions. So both answers are sort of related. The creative is the biggest hurdle in our system. So nothing has really changed there.

是的。謝謝這些問題。所以兩個答案其實有點相關。創意素材是我們系統中最大的門檻。所以在這方面其實沒有太大改變。

Lengthy video plus interactive end card, which is in a common ad unit, we can auto-generate the interactive end card with pretty high efficiency at this point. We're not at the point where we can get a high-quality video for 30 to 60 seconds in the hand of an advertiser out of the box. So that's still work in progress. And we have examples where it works, other examples where it doesn't work. The reality is, once we can get that to happen or we create alternative forms of templates that don't require a video, we'll be able to hand the advertisers a one-click campaign creation and that should resolve any sort of conversion rate issues on the flow.

長影片加上互動式結尾卡（interactive end card）是一種常見的廣告單元；目前我們可以以相當高的效率自動生成互動式結尾卡。但我們還沒到能夠直接「開箱即用」就把 30 到 60 秒的高品質影片交到廣告主手上的程度。所以這仍在進行中。而且我們有一些例子是有效的，也有一些例子是無效的。現實是，一旦我們能做到這件事，或是我們建立不需要影片的替代模板形式，我們就能提供廣告主一鍵式的活動建立，這應該能解決流程中任何轉換率的問題。

So it's just not something that we've gotten to yet, is -- that's the same concern that people have on number two. When they buy on social and search, a lot of the advertisements in the -- when you buy on open web, social and search, a lot of the advertisements are just dynamic catalogs or they're more static, like come get a sale, click here to buy this product, much less so a 30- to 60-second video. So it's not that common that someone has that unless they're a pretty material advertiser that is buying on television.

所以我們還沒做到那一步——這也是大家在第二個問題上同樣的顧慮。當他們在社群與搜尋投放時——或者說在開放網路、社群與搜尋上投放時——很多廣告其實只是動態商品目錄，或是更靜態的形式，例如「來參加特賣，點這裡購買這個商品」，而不是那種 30 到 60 秒的影片。所以除非是相當大型、也會投放電視的廣告主，否則一般並不常見有人會有那樣的素材。

And so if you think about that like mid-market and up, probably has templates that match the creative needs on our platform. When you get into the SMBs that are signing up directly, they probably don't. And so we've got to resolve that as we go into wanting to get in more in the long tail over time.

因此，如果你把它想成是中型市場以上的客戶，可能有符合我們平台創意需求的範本。但當你進入那些直接註冊的中小企業（SMB）時，他們可能沒有。所以，隨著時間推進、我們想在更長尾的市場滲透得更深時，我們必須逐步把這個問題解決。

Alec Brondolo, Wells Fargo.

Alec Brondolo，富國銀行（Wells Fargo）。

I'd love to ask about the health of the mobile game ecosystem. I think a lot of us are having conversations with mobile game publishers and developers. They're speaking to CPI inflation, maybe waning ROAS, there's some market data that suggests that mobile game app downloads are down something like 10% to 15% year over year over the last several months.

我想請教行動遊戲生態系的健康狀況。我想我們很多人都在和行動遊戲發行商與開發者交流。他們提到 CPI 上升、ROAS 可能走弱；也有一些市場數據顯示，過去幾個月行動遊戲 App 下載量年增率下滑大約 10% 到 15%。

And just how do you guys feel about the category? How are your conversations with customers progressing? And then I have a follow-up.

你們對這個類別的看法如何？你們與客戶的對話進展如何？另外我還有一個追問。

Yeah. So it's a good question. The category has to be good for us to perform in gaming, right? That's why we gave you the stat that the MAX marketplace grew double digits Q over Q, which is not a material growth. In fact, that's incredibly outsized growth.

是的。這是個好問題。這個類別必須表現良好，我們才能在遊戲領域有好的表現，對吧？這也是為什麼我們提供那個數據：MAX 市集（marketplace）季對季成長達到雙位數，這並不是「不重要」的成長。事實上，那是非常顯著、遠超常態的成長。

The ad-supported market is growing really quickly. The in-app purchasing market has some misleading data out there because companies that are analytics providers can't actually track the purchases that are going off platform. And as more and more in-app purchasing games go off-platform, the revenue didn't just disappear or shrink. It actually became incremental to them on a net revenue basis. It's just not measurable.

廣告支持（ad-supported）的市場成長非常快。應用內購買（in-app purchasing）市場有一些誤導性的數據，因為提供分析的公司其實無法追蹤那些在平台外完成的購買。而且，隨著越來越多的內購遊戲轉到平台外交易，營收並沒有消失或縮水。實際上，以淨營收（net revenue）口徑來看，對他們而言反而是增量。只是無法被量測到。

So as you look at what companies are reporting as services revenue in the category or what these analytics providers are reporting, you're not going to catch all that incremental revenue that's missed.

所以，當你看公司在這個類別中揭露的服務營收，或是這些分析供應商所報告的數據時，你不會捕捉到那些被漏掉的增量營收。

Now when it comes to installs, the trend over the last few years has been a movement away from very high install, low-quality hyper-casual, much more to casual and deeper games and in-app purchasing. As you get into deeper funnels, you get a much higher CPI and a much higher LTV. So as the cost goes up, it does not mean that the value goes down. Everything is on a return-on-ad-spend basis.

至於安裝量（installs），過去幾年的趨勢是從「安裝量很高但品質較低」的超休閒（hyper-casual），轉向更偏休閒（casual）與更深度的遊戲，以及應用內購買。當你進入更深的漏斗（funnels）時，CPI 會更高，LTV 也會更高。因此，成本上升並不代表價值下降。一切都以廣告投資報酬（return-on-ad-spend）為基礎。

Now the last bullet is, we drive a lot of the market too, especially for Discovery. We're the biggest in the world as far as I know, when it comes to mobile gaming user acquisition. And so when we have a quarter where we don't push lifts, that's not great for the category as well. We're driving growth in the category given the scale that we operate at. Then when we go follow it up and we push a model release early in Q3, if you talk to a lot of those same advertisers, they'll say, install rates went up, CPIs went down, performance has improved.

最後一點是，我們也在很大程度上推動了市場，尤其是在 Discovery 方面。就我所知，在行動遊戲用戶獲取（user acquisition）上，我們是全球最大。因此，當我們某一季沒有去推動提升（lifts）時，對這個類別也不是好事。以我們的營運規模，我們正在推動這個類別的成長。接著當我們在第三季初推出模型版本更新（model release），如果你去問同一批廣告主，他們會說：安裝率上升了、CPI 下降了、成效變好了。

So we are the catalyst in large part in this category at this point if you're talking about user acquisition. And that is something that we take very seriously. We've got to be on top of our game so that these game developers can continue to see the type of success that our tools and platform have enabled for years.

所以，如果談的是用戶獲取，在目前這個類別裡，我們在很大程度上就是催化劑。而這是我們非常認真看待的事情。我們必須保持最佳狀態，讓這些遊戲開發者能持續看到多年來我們的工具與平台所帶來的那種成功。

Yeah. And then maybe one on the web advertising business and advertiser acquisition. You were branded AppLovin, then you changed the brand to Axon when you entered web advertising. You changed it back to AppLovin when you went to general availability. I think the question I have is, do you feel like the AppLovin or Axon brand, the lack of kind of knowledgeability of the brand among consumers is hurting your ability to acquire advertisers? Because it seems to me that in the long tail, the advertisers are choosing their ads platforms based on the platforms they know as consumers.

好的。接著想問一個關於網頁廣告業務與廣告主獲取的問題。你們原本的品牌是 AppLovin，進入網頁廣告時把品牌改成 Axon。在進入全面可用（general availability）後又改回 AppLovin。我的問題是：你們是否覺得 AppLovin 或 Axon 這個品牌，因為一般消費者對品牌的認知度不高，正在傷害你們獲取廣告主的能力？因為在我看來，在長尾市場，廣告主會依據他們作為消費者所熟悉的平台來選擇投放平台。

Obviously, they don't know AppLovin as a platform. And so I wonder if you have the right branding to go out and acquire the next 100,000 longer-tail e-commerce and web advertisers.

顯然，他們不會把 AppLovin 視為一個平台。所以我想知道，你們是否具備正確的品牌定位，去獲取接下來 10 萬個更長尾的電商與網頁廣告主。

Yeah. I mean, look, we can't get rid of the name AppLovin, unfortunately, but we switched to Axon, everyone kept calling us AppLovin, we switched back. The reality is, if you talk to 100 customers that don't use us in e-commerce today, there is going to be an awareness problem, but we're early in this category. A lot of people may actually know of us, but think of us as a mobile gaming platform.

是的。我的意思是，你看，我們不幸地無法把 AppLovin 這個名字拿掉，但我們改成 Axon 後，大家還是一直叫我們 AppLovin，所以我們又改回來。現實是，如果你去跟 100 個目前在電商領域還沒使用我們的客戶聊，確實會有認知度（awareness）問題，但我們在這個類別還很早期。很多人其實可能聽過我們，但把我們當成行動遊戲平台。

Now if you asked us 10 years ago in mobile gaming, do the customers know about us? The answer would have been no. A lot of customers don't know about us. We have some customers, but they're going to Facebook and Google much more so than they're coming to AppLovin. If you ask us today, well, every customer is on our platform. So in some part, brand does matter, but you earn brand loyalty with performance.

如果你在 10 年前問我們，在行動遊戲領域客戶是否知道我們？答案會是否。很多客戶並不知道我們。我們有一些客戶，但他們去 Facebook 和 Google 的程度遠高於他們來 AppLovin。如果你今天問我們，那麼每個客戶都在我們的平台上。所以在某種程度上，品牌確實重要，但品牌忠誠度是靠績效表現贏來的。

If we continue to compound improvements in technology, improvements in the templates, more success stories out there, over time, the customers are going to find out about our platform. And I mean, again, I said this a couple of minutes ago, it took Google a couple of decades to become the de facto standard for customers in anything search related. It took Facebook well over a decade.

如果我們持續累積技術改進、範本改進，並且有更多成功案例流傳，隨著時間推移，客戶就會認識我們的平台。而且我再說一次，我幾分鐘前提過：Google 花了幾十年才成為所有與搜尋相關領域中客戶的事實標準（de facto standard）。Facebook 也花了超過十年。

So these are not things that just build overnight. We're not going to have out of those 100 brands, a high percentage in the next quarter, knowing of us. But if we do our job right over the coming quarters, it's going to compound and get to a point where the brand is going to be recognized as a standard. And these analytics providers do help as well as they talk to the customers on their platform, if they see us performing as the second or third best channel for customers that we have live, it behooves them to push us to their clients. And in some way, it will also benefit them economically because we'll strike a lot of partnership deals.

所以這些都不是一夕之間就能建立的。在接下來一季，我們不可能讓那 100 個品牌中有很高比例的人認識我們。但如果我們在未來幾季把工作做好，它會複利式累積，最後到達一個品牌會被視為標準、被廣泛認可的程度。而這些分析供應商也會有所幫助：當他們與其平台上的客戶溝通時，如果看到我們在已上線的客戶中表現是第二或第三好的渠道，他們就有動機把我們推薦給他們的客戶。而且在某種程度上，這也會在經濟上讓他們受益，因為我們會達成很多合作夥伴協議。

Robert Coolbrith, Evercore.

Robert Coolbrith，Evercore。

I wanted to ask a couple on mobile gaming. Some of the feedback we received from advertisers this quarter is that they're spending a bit more on Android than they had expected entering the year for a variety of reasons, some publisher-specific, some influenced by things going on in the Play ecosystem. And that may be impacting their share of voice or share of wallet with AppLovin, just given the mix dynamics. Just wondering if you can maybe talk about the opportunity within Android to maybe drive further competitive distance versus your peers and higher share of wallet over time?

我想問幾個關於行動遊戲的問題。我們這一季從廣告主那裡收到的一些回饋是：由於各種原因，他們在 Android 上的花費比年初預期的更多；有些是特定發行商因素，有些則受到 Play 生態系內發生的事情影響。考量到組合（mix）動態，這可能正在影響他們在 AppLovin 的聲量份額（share of voice）或錢包份額（share of wallet）。想請你談談：在 Android 之內是否有機會，隨著時間推進，進一步拉開相對同業的競爭距離，並提高錢包份額？

And then second one is just on the World Cup. We've got some feedback that, that may have been a headwind for some of the IAP-focused advertisers. Anything you can maybe call out around those particular dynamics or other cyclical dynamics in Q2?

然後第二個就是關於世界盃。我們收到一些回饋，這可能對某些以 IAP 為重點的廣告主造成了逆風。你能否就這些特定動態或 Q2 其他週期性動態做一些說明？

Yeah. So the World Cup first, World Cup has spikes in spend from World Cup-specific advertisers during games. And it's not that material one in the grand scheme of a whole quarter. I wouldn't say the World Cup has much of any impact on if an in-app purchasing or ad-supported game customer can actually spend on platform and we don't index high to any sort of customer in any category nor do we run any branding. So for us, we still cater to the game advertiser.

是的。先談世界盃，世界盃期間在比賽進行時，會有世界盃特定廣告主的投放支出尖峰。但放在整個季度的大盤來看，影響並不算太大。我不會說世界盃對於應用內購買或廣告支持型遊戲客戶是否能在平台上實際投放有什麼影響；而且我們在任何類別的客戶占比都不算特別高，我們也不做任何品牌廣告。所以對我們而言，我們仍然是服務遊戲廣告主。

On the first question, we are very competitive on both platforms. Now there is a very large company out there that owns the Google Play platform that provides a lot of installs and spend for customers on Google Play. So I wouldn't say that there -- I would say, for sure, like we look at the opportunity to grow equally on both platforms. As the models get better, our ability to scale spend goes up on both platforms. It's just that the Android platform is more competitive because there's a very big competitor out there doing well there.

關於第一個問題，我們在兩個平台上都非常有競爭力。不過市場上有一家非常大的公司擁有 Google Play 平台，能為 Google Play 上的客戶帶來大量安裝量與投放支出。所以我不會說那裡——我會說，我們確實把在兩個平台上同等成長視為機會。隨著模型變得更好，我們在兩個平台上擴大投放規模的能力都會提升。只是 Android 平台競爭更激烈，因為那裡有一個非常大的競爭對手表現得很好。

Jim Callahan, Piper Sandler.

Jim Callahan，Piper Sandler。

One for me on the revenues by geography. It looks like US accelerated Q over Q, but international was closer to flat with 1Q. Anything to sort of call out there on the delta?

我有一個關於按地理區域劃分收入的問題。看起來美國環比加速，但國際市場與 1Q 相比更接近持平。這個差異有什麼需要特別說明的嗎？

No, nothing material. I mean, international has been strengthening for a period of time over the past few quarters, and then it's slowed down a little bit, but nothing specific. And that's based on user location. So it's really the demographics based on region.

沒有，沒有什麼重大因素。我的意思是，過去幾個季度國際市場一直在走強，然後最近放緩了一點，但沒有特定原因。而且這是基於用戶所在地。所以本質上是按區域的人口結構差異。

Now also remember the web consumer business is more concentrated to the West than it is internationally at least today.

另外也要記得，至少在目前，網頁端消費者業務相較於國際市場，更集中在西方地區。

Okay. That all makes sense. And if we think about non-gaming, but maybe excluding e-comm, I think there have been some discussions of verticals like either short dramas or prediction markets that may be leaning into spend kind of on the mobile gaming category. Any commentary you can provide there?

好的。這都說得通。如果我們看非遊戲領域，但可能先排除電商，我想有一些討論提到像短劇或預測市場等垂直類別，可能在投放上有點向行動遊戲類別靠攏。你能對此提供一些評論嗎？

Yeah. I mean, we just touched on the World Cup, prediction markets was a big category there. But those other categories aren't yet a focus of ours. If you split the world up into consumer business, e-commerce on the web and then non-gaming apps. The former is a much bigger category and that's what we're going after first. So we will go after the second later, and it just hasn't been a focus of ours yet in what we model.

可以。我們剛提到世界盃，預測市場在那段期間是一個大類別。但其他那些類別目前還不是我們的重點。如果把世界分成消費者業務、網頁端電商，以及非遊戲 App。第一個類別要大得多，也是我們優先要攻的方向。所以第二個我們之後才會做，而在我們目前的建模中，它還不是重點。

Matt Swanson, RBC.

Matt Swanson，RBC。

Great. You've been able to build this company incredibly lean. And when you were talking about the model improvements and doing A/B tests, but what could the advantages be of maybe going from dozens of iterations to hundreds of iterations? Like is there a thought process around the right level of R&D spend? Or is there a diminishing return from a ramp like that?

很好。你們能把公司打造得非常精實。你剛提到模型改進與做 A/B 測試，但如果把迭代從數十次提升到數百次，可能的優勢會是什麼？對於合適的研發支出水準，有沒有一套思考框架？或者這樣加速會有邊際效益遞減嗎？

No. I mean, like you want to do -- you have to do smart A/B tests, right? Like just a whole bunch of nominal different A/B tests or low-IQ A/B tests aren't going to yield uplifts. So in a way, we want very high talent density, which we have. We still need to hire more research scientists, do more A/B tests, but we need to become even more sophisticated using AI technology to accelerate the rate of those tests, too.

不會。我的意思是，你想做——你必須做聰明的 A/B 測試，對吧？只是做一大堆名義上不同的 A/B 測試，或低智商的 A/B 測試，是不會帶來提升的。所以某種程度上，我們希望有非常高的人才密度，而我們確實有。我們仍需要招聘更多研究科學家、做更多 A/B 測試，但也需要用 AI 技術變得更成熟，以加速這些測試的速度。

So we do a lot of that, where AI is helping aid testing. And the rate of testing is, for sure, gone up in our company over the last couple of years as LLMs have gotten a lot more sophisticated writing code. So our focus won't change in terms of hire very, very smart people and then enable them to drive to more A/B tests powered by large language models helping with a lot of the coding.

所以我們做了很多這類事情，AI 正在協助測試。而且隨著 LLM 在寫程式碼方面變得更成熟，過去幾年我們公司的測試速度確實提升了。因此我們的重點不會改變：招聘非常非常聰明的人，並讓他們在大型語言模型協助大量編碼的情況下，推動更多由 LLM 驅動的 A/B 測試。

Clark Lampen, BTIG.

Clark Lampen，BTIG。

We've asked a bunch of questions on sort of funnel dynamics so far. And I think when you were answering Omar's question, you sort of broke it down in a very simplistic way where you can think about customer growth, data growth, and then that having a downstream impact on sort of performance and model uplifts. It sounds like you're addressing the first portion, at least near-term, with partnerships.

到目前為止我們問了很多關於漏斗動態的問題。我想你在回答 Omar 的問題時，用一個很簡化的方式拆解：可以把它看成客戶成長、數據成長，然後對下游的表現與模型提升產生影響。聽起來你們至少在短期內，是透過合作夥伴來解決第一部分。

Is there anything that you can do, I guess, on that sort of second layer in terms of data collection, whether it would be something like extending ad credits or maybe building out more of a managed service presence for an advertiser to help them sort of scale volume? Maybe that doesn't make sense, but it sounds like they're sort of interrelated, and I'm just curious if sort of pushing on that second lever might help the first.

那在第二層，也就是數據收集方面，你們能做些什麼嗎？例如延長廣告額度，或為廣告主建立更多代管服務（managed service）的能力，幫助他們擴大量級？也許這不一定合理，但聽起來它們彼此相關；我只是好奇，推動第二個槓桿是否能幫助第一個。

And the second question, if I may, the lead gen business isn't something that's sort of come up thus far on the call, unless I missed it. And I'm curious if you could provide just a very brief update on where things stand right now, whether it's ramping or just sort of directional trends?

如果可以的話，第二個問題：名單開發（lead gen）業務在這通電話中似乎還沒被提到，除非我漏聽了。我想請你非常簡短地更新一下目前進展如何，是在加速爬坡，還是一些方向性的趨勢？

Yeah. Second one first, and thanks, Clark. Lead gen business is still work in progress. So we're still in testing with customers. There's nothing new to report there.

好的。先回答第二個，謝謝你，Clark。名單開發業務仍在進行中。所以我們仍在與客戶進行測試。目前沒有新的內容可以報告。

On the first one, every new customer we get live, their spend doesn't determine the type of data that we are able to access. They have to share their data with us to tap into the deep learning model that we have to get output, good advertising results out of our system. So the goal is getting more mid-market customers that can give our model visibility into more of the users' transactional behavior. If we're able to do that over time, which is inevitable, it's just getting more customers on the platform that are substantial, we will have more data that then we can model against, write a more complex model and get a better output. So that's sort of the function over time.

關於第一個，每一個新上線的客戶，他們的投放支出並不會決定我們能取得哪種類型的數據。他們必須與我們分享他們的數據，才能接入我們的深度學習模型，從系統中產出良好的廣告成效。所以目標是獲得更多中型市場客戶，讓我們的模型能更看得到用戶更多的交易行為。如果我們能隨時間做到這點——這是必然的，只是要讓更多具規模的客戶上平台——我們就會有更多數據可用來建模，寫出更複雜的模型並得到更好的輸出。這就是隨時間推進的運作方式。

And it compounds because the more customers we get, the better our performance will be. And then you'll get to a place where, hopefully, we reach a tipping point and the rate of customer acquisition goes up and then it's sort of just automatically off to the races.

而且它會產生複利效應，因為我們獲得的客戶越多，我們的表現就會越好。然後你會希望我們能到達一個臨界點，客戶獲取速度上升，接著就會像自動加速一樣一路狂奔。

Aaron Lee, Macquarie.

Aaron Lee，Macquarie。

I wanted to take another crack at the margin question from earlier. How should we be thinking about the incremental investment in tech and compute going forward? Obviously, we have the 3Q guide, so we can kind of back into that. But just thinking beyond 3Q, should we expect an elevated runway for investment? And just how much visibility into those training and compute costs going forward do you have?

我想再回到先前的毛利問題再追問一次。我們應該如何看待未來在技術與運算方面的增量投資？顯然我們已經有第三季的指引，所以大致可以從中推算。但如果把視野放到第三季之後，我們是否應該預期投資期間會維持在較高水準？另外，對於未來訓練與運算成本，你們有多大的能見度？

Yeah. I mean, we don't expect any departure from the higher-level guidance that we've given to analysts and investors in the past that of the incremental dollar in revenue, we're spending about $0.10 on compute. We're at that level within the guide. And we're at that level if you look at the data center disclosure, you'll get in the 10-Q as well. We're still tracking at that level.

是的。我的意思是，我們不預期會偏離過去提供給分析師與投資人的高層指引：每增加 1 美元的營收，我們大約會在運算上花 0.10 美元。在本次指引中，我們也處於這個水準。而且如果你看我們在 10-Q 裡的資料中心揭露，也會落在約 10% 的水準。我們仍然維持在這個水準上。

So we don't expect any difference from here, but we may see variability over the longer term, and we'll communicate that and the reasons behind why we've increased compute if there is such an increase. But today, we don't expect any change from that kind of guide.

所以我們不預期從現在起會有任何不同，但長期來看可能會有波動；如果運算支出確實增加，我們會對外溝通並說明增加的原因。但就目前而言，我們不預期會改變這類指引。

Now we will say, look, engineers are doing R&D, right? So if someone figures out a way to write a more complex model, way more parameters, train a bigger model, and run it, and there's a material revenue uplift, we're not going to hold it back. So if we go into a quarterly earnings call and four weeks later, we get that word, we're pushing forward. And you saw as we talked about the guide and reaccelerating, we have a model uplift that releases, more complex model creates revenue reacceleration, that's always a good thing in our business.

不過我們也要說，工程師是在做研發，對吧？所以如果有人想出方法能寫出更複雜的模型、更多參數、訓練更大的模型並且能運行，而且能帶來顯著的營收提升，我們不會阻止。因此如果我們在某次季度財報電話會議後四週就得到這樣的訊息，我們會推進。而且你也看到，當我們談到指引與再加速時，我們有模型提升的釋出；更複雜的模型帶來營收再加速，這在我們的業務裡永遠是好事。

Got you. That's helpful. And then on the figure you've given out previously for $70,000 of gross spend in the first year for new customers. Is that still the right bogey we should be thinking about? And are model breakthroughs the key driver to getting that number higher? Or are there any other drivers you would highlight?

了解。這很有幫助。另外，關於你們先前提到新客戶第一年約 70,000 美元的總支出（gross spend）這個數字，現在仍然是我們應該參考的基準嗎？這仍然是正確的目標值嗎？而模型突破是否是推動這個數字提高的關鍵驅動因素？或者還有其他你會特別強調的驅動因素？

It will likely go up as we do partnership deals and target mid-market. So that's just an indicator of are the customers coming to the platform more just signing up on the website, pushed by marketing, where you're going to get a lot more SMBs and low-GMV shops. Or are the customers that are signing up less in quantity and higher GMV that we're getting through strategic partnerships. The latter being the focus, we'd expect that number to go up.

隨著我們做更多合作夥伴交易並鎖定中型市場，這個數字很可能會上升。所以這只是用來指標化：客戶是更多透過網站自行註冊、由行銷推動而來——那樣你會得到更多中小企業（SMB）與低 GMV 的商家；還是註冊的客戶數量較少、但 GMV 較高，且是透過策略合作夥伴關係帶來的。若我們聚焦於後者，我們預期那個數字會上升。

Martin Yang, OpCo.

Martin Yang，OpCo。

Can you maybe talk a bit more about the nature of your partnership. Is it going through those partner -- attribution partners giving the customers additional credit, better data integration? What are the customers of these partners benefiting from having AppLovin?

你能否再多談一些你們合作夥伴關係的性質？是透過那些合作夥伴——歸因（attribution）合作夥伴——為客戶提供額外的信用、或更好的資料整合嗎？這些合作夥伴的客戶，從使用 AppLovin 中獲得了什麼好處？

Yeah. I mean, it's a good question. We're still working through how to do it the optimal way. But if you're an attribution company, and we're willing to pay for leads and they benefit from their customers having more complex attribution, another big vendor in the mix, there's a big win. If the customer on the other side gets another big vendor in the mix, it's a big win for them, too. So you've got a construct that's just win-win-win across the board.

是的。這是個好問題。我們仍在摸索如何以最優方式來做。但如果你是一家歸因公司，而我們願意為線索（leads）付費，他們也會因為其客戶能做更複雜的歸因、並且在組合中加入另一個大型供應商而受益，這就是很大的雙贏。而另一端的客戶能在組合中加入另一個大型供應商，對他們也是很大的好處。所以你會得到一個在各方面都是 win-win-win 的架構。

But we've got to get past the awareness problem and the implementation problem. So Triple Whale was mentioned earlier on the call. That is a good example of one that we're iterating on actively to find the right way to let a large, large set of customers working with them understand that our platform is beneficial. As a partner, they're highly motivated, one, because there's obviously revenue potential, but two, because it is very beneficial for their business model to add us into the mix for all their clients.

但我們必須先跨過認知度問題與導入（implementation）問題。剛才電話會議中提到 Triple Whale。那就是一個很好的例子：我們正積極迭代，找出正確方式，讓與他們合作的一大群客戶理解我們的平台是有益的。作為合作夥伴，他們的動機非常強：第一，顯然有營收潛力；第二，將我們納入其所有客戶的組合，對他們的商業模式非常有利。

Got it. And then a longer-term question, as you move beyond mid-markets, when you think about working with Fortune 500 brands, for instance, sometimes you talk about mobile ads in general of having more quality control issues. So when you go to those Tier 1 brands, are you -- or would you think about expanding supplies to non-gaming?

了解。再問一個更長期的問題：當你們走出中型市場、開始思考與《財富》500 強品牌合作時，例如你們有時會談到行動廣告整體可能有較多品質控管問題。所以當你們面向那些一線（Tier 1）品牌時，你們是否——或會不會考慮——把供給擴展到非遊戲領域？

Yeah. Look, supply will expand over time. But I would say, actually, it's a little bit crazy to think that there's quality controls in mobile gaming because there's no UGC. The apps are in the App Store, approved. These are only adult apps, and there's no UGC. So it's a very, very controlled framework.

會的。你看，供給會隨時間擴張。但我會說，其實認為行動遊戲存在品質控管問題有點不可思議，因為那裡沒有 UGC（使用者生成內容）。這些 App 都在 App Store 上，且已獲核准。而且這些都只面向成人的 App，並且沒有 UGC。所以那是一個非常、非常受控的框架。

An ad inside Candy Crush is materially safer than an ad almost anywhere else because that inventory is so controlled. And then obviously, the attention given to that ad in mobile gaming is much more than anywhere, including television, given our ads run longer than 30 seconds to the user. So it sets us up to be a really good framework.

在 Candy Crush 裡的廣告，在實質上比幾乎任何其他地方的廣告都更安全，因為那個庫存（inventory）非常受控。而且很明顯地，在行動遊戲中使用者對該廣告的注意力也遠高於其他任何地方，包括電視，因為我們的廣告對使用者播放時間超過 30 秒。因此這讓我們成為一個非常好的框架。

Now for us to get to the head of the market, we need sales. That requires actual people selling. And we need a reputation that suggests we've made the middle of the market work. We're more focused on the middle of the market because there, you've got hungrier companies, leaner marketing teams, probably not agencies, at least big holding companies in the middle, so things can go quicker. But if we're able to build a really solid reputation here, we're definitely going to go to the head of the market. I don't think inventory quality will be the concern inside gaming.

現在，若要打到市場頂端，我們需要銷售。那需要真正的人去賣。我們也需要一個聲譽，能證明我們已經把中型市場做起來。我們更聚焦在中型市場，因為那裡的公司更有企圖心、行銷團隊更精實，可能也不透過代理商——至少在中間市場不太會有大型控股公司旗下的代理商——所以事情可以更快推進。但如果我們能在這裡建立非常扎實的口碑，我們一定會往市場頂端走。我不認為遊戲內的庫存品質會是疑慮。

Now that does not mean that supply expansion is, for sure, not one of the levers that we'll pull on over time as a growth lever.

當然，這並不表示供給擴張在未來作為成長槓桿之一就一定不會被我們採用。

Vasily Karasyov, Cannonball.

Vasily Karasyov，Cannonball。

I have a long-term question, Adam, and it's about run-time data integration with Unity Vector. It's not a competitive impact question. So obviously, AppLovin has a very strong position because you capture pretty much all the advertising signal there is to be caught -- captured, right, and you utilize it very well. But that said, if I understand correctly, that run-time data gives a very specific, very unique engine-level set of signals that, as I understand, you don't have native access to, right? And if that's true, then it seems like there is an argument can be made that both channels can become complementary because they're using different sets of signals, right?

Adam，我有一個長期問題，關於與 Unity Vector 的執行期（run-time）資料整合。這不是一個競爭影響的問題。很明顯，AppLovin 的地位非常強，因為你們幾乎捕捉到了所有能捕捉的廣告訊號——都被捕捉到了，對吧？而且你們也把它運用得很好。但即便如此，如果我理解正確，執行期資料會提供一組非常特定、非常獨特、引擎層級的訊號；據我了解，你們並沒有原生（native）存取權，對吧？如果這是真的，那似乎可以主張兩個渠道可能會互補，因為它們使用的是不同的訊號集合，對吧？

So I was wondering if you could tell me if I'm thinking about it the wrong way and what are the implications of this new -- if I understand correct, new set of signals coming to the market, what does it mean for the industry? Does the overall spend grow as a result?

所以我在想，您能否告訴我是不是我用錯了角度在看這件事，以及這個新的——如果我理解正確的話——一組新的訊號進入市場，對產業意味著什麼？整體支出會因此而成長嗎？

Yeah. I mean, look, this ties to what we tell customers that come to us and we're their largest channel of spend. Sometimes we get asked, should they only spend with us? Or are they competing with themselves that they spend on other platforms. Everyone's data and model is different.

是的。我的意思是，你看，這和我們對來找我們、而且我們是他們最大支出渠道的客戶所說的內容有關。有時候我們會被問，他們是否應該只在我們這裡投放？或者他們在其他平台投放是否等於在跟自己競爭。每家的數據與模型都不一樣。

And in this space, I mean, this has been -- I think when we first went public, this was part of the narrative that was negative on us is that we're in a zero-sum sector. I think we've proven over the last five years that this is not a zero-sum sector.

而在這個領域，我的意思是，這一直——我想當我們剛上市時，對我們不利的敘事之一就是我們身處一個零和的產業。我認為過去五年我們已經證明，這不是一個零和產業。

As marketing companies improve across the board, scale of user acquisition goes up, growth goes up, the P&L of the end client, the gaming customer improves, then they can reinvest more dollars into marketing platforms. We're obviously the biggest. We want the market to improve. In that quote I gave you with the MAX marketplace improving double digits quarter over quarter, that's a huge amount of growth for a very, very large market at this point in terms of our penetration into the ad-supported market and the ecosystem.

隨著各家行銷公司整體能力提升，用戶獲取的規模會上升、成長會上升，終端客戶的損益表（P&L）——也就是遊戲客戶——會改善，然後他們就能把更多資金再投入到行銷平台。我們顯然是最大的。我們希望市場變得更好。我剛才提到 MAX 市集（marketplace）季度環比以雙位數改善，對於目前我們在廣告支撐市場與生態系中的滲透率而言，在一個已經非常、非常大的市場上，這是非常可觀的成長幅度。

So we've proven the space is not zero-sum, and that's because these models, when trained with differentiated data, create a higher ceiling for the gaming customer, which is beneficial.

所以我們已經證明這個領域不是零和的，而原因在於：這些模型在以差異化數據訓練後，能為遊戲客戶創造更高的上限，這是有利的。

Tim Nollen, SSR.

Tim Nollen，SSR。

Adam, I think I caught a comment from you regarding Wurl recently. This is something which has not been addressed on this call, I don't think yet. So my question is that now that the consumer business is launched, GA, are you turning your attention a bit more to Wurl? You had said that you would -- you'd kind of wait on that until you got this consumer business out there. Now that it's out there, I'm curious if Wurl becomes more of a topic of interest for you?

Adam，我想我最近聽到你提到 Wurl 的一段評論。我想這件事在這通電話裡還沒有被談到。所以我的問題是，現在消費者業務已經推出、進入 GA（正式可用），你是否會把注意力多放一些到 Wurl 上？你之前說你會——你會先等到把這個消費者業務推出去之後再說。現在既然已經推出了，我很好奇 Wurl 是否會成為你更感興趣的主題？

And maybe how does your consumer experience lend itself to Wurl? So from going from in-game ads to CTV ads, presumably now with lots of new relationships on the advertiser side as well. Any comment on Wurl would be great.

另外，你們的消費者端經驗如何有助於 Wurl？也就是從遊戲內廣告走向 CTV 廣告，推測現在在廣告主端也會有很多新的合作關係。如果能談談 Wurl 就太好了。

Yeah. Thanks, Tim. Still getting used to your new firm. But on the Wurl and CTV question, look, in consumer, we're still budgetarily constrained. We're still early in the category. And if we were to go expand supply, the most natural is going to non-gaming apps and other open web type placements, but the consumer vertical allows us to actually go out and expand supply.

是的。謝謝你，Tim。我還在適應你新的公司。不過關於 Wurl 和 CTV 的問題，你看，在消費者端，我們在預算上仍然受限。我們在這個類別也還很早期。如果我們要去擴大供給，最自然的會是走向非遊戲 App，以及其他開放網路（open web）類型的版位；但消費者垂直領域其實讓我們能夠走出去並擴大供給。

Gaming is much more niche. And really, if you look at the world of the open web, social, you almost never see gaming ads. Of course, you see e-commerce ads, right? So we think the first path to supply expansion will be on device, and the second path will be connected TV.

遊戲要小眾得多。而且如果你看開放網路、社群的世界，你幾乎從來看不到遊戲廣告。當然，你會看到電商廣告，對吧？所以我們認為供給擴張的第一條路徑會是在裝置端，第二條路徑會是連網電視。

It's definitely an area of interest for us because that same lipstick selling ad on the mobile device that's full screen should port really, really well to television. And we know that television drive shopper behavior, everyone knows that at the scale that TV operates at. So it's an opportunity that's sitting there. We're just not at the point yet where we can go execute on it.

這絕對是我們感興趣的領域，因為同一支在行動裝置上全螢幕播放的口紅銷售廣告，應該能非常、非常順利地移植到電視上。而且我們知道電視會驅動消費者的購物行為，大家都知道在 TV 所運作的那種規模下會有這樣的效果。所以這是一個擺在那裡的機會。只是我們還沒到能去執行它的階段。

Yeah, like I said, it just seems like a big market opportunity that itself is shifting to more performance-based advertising, and you've done such a great job on that. I guess the issue is it's less shifting to CTV apps and more to shifting to the interface and the audience and the advertiser base, to understand those better?

是的，就像我說的，這看起來是一個很大的市場機會，而且它本身也正在轉向更以成效為導向的廣告，而你們在這方面做得非常出色。我想問題在於：與其說是轉向 CTV App，不如說是轉向介面、受眾以及廣告主基礎，去更好地理解這些？

Yeah. It's more -- think of our consumer advertisers, if they're spending, the earlier question of -- they're seeing really good ROAS, but they're not even spending at their ceiling yet, we're not at a point where we have excess budgets to take out. So if we just launched CTV and actually made it work and shifted dollars, we're weakening our position in mobile to go launch into CTV, and we wouldn't actually expand our revenue. To go into new supply, we need to believe there's going to be more dollars flowing.

是的。更像是——想想我們的消費者類廣告主，如果他們在投放，回到前面那個問題——他們看到非常好的 ROAS（廣告投資報酬率），但他們甚至還沒花到自己的上限，我們目前也還沒到有多餘預算可以挪出的程度。所以如果我們現在就推出 CTV、而且真的把它做起來並把資金移過去，我們是在削弱我們在行動端的地位去啟動 CTV，而我們其實不會擴大營收。要進入新的供給，我們需要相信會有更多資金流入。

So step one would be the obvious, it's non-gaming apps. And gaming apps that don't currently run games, which would run non-gaming ads inside those apps. And then step two would be the open web. Step three would be connected TV. All of that's going to come for us because it's just -- those are just obvious levers of growth to pull when we have the budget to go out and get it.

所以第一步會是很明顯的：非遊戲 App。以及目前不投放遊戲廣告的遊戲 App——在那些 App 裡投放非遊戲廣告。然後第二步會是開放網路。第三步會是連網電視。這些對我們來說都會到來，因為——當我們有預算可以出去拿的時候，這些就是很明顯可以拉動成長的槓桿。

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

以上為本季問答環節的結束。感謝各位今天的參與。祝各位下午愉快。