Pattern Group Inc. (PTRN) 2026 Q2 法說會逐字稿

Good day, and thank you for standing by. Welcome to the Pattern Q2 2026 earnings conference call. (Operator Instructions) Please be advised that today's conference is being recorded.

各位好，感謝您耐心等候。歡迎參加 Pattern 2026 年第二季（Q2）財報電話會議。（接線員指示）請注意，今日會議將被錄音。

I would now like to hand the conference over to your first speaker today, Hamish Chung, VP of Finance. Please go ahead.

現在我想把會議交給今天的第一位講者——財務副總裁 Hamish Chung。請開始。

Thank you, operator. Good afternoon and thank you for joining Pattern's earnings call for the Q2 2026.

謝謝，接線員。各位下午好，感謝您參加 Pattern 2026 年第二季（Q2）財報電話會議。

Before we begin, I'd like to remind everyone that today's discussion may contain forward-looking statements based on our current expectations, assumptions, and forecasts about future events. These forward-looking statements are subject to risks and uncertainties that could cause actual results to differ materially from those projected. Please refer to our latest filings with the Securities and Exchange Commission for more information on these risks and uncertainties.

在開始之前，我想提醒各位，今天的討論可能包含以前瞻性陳述為主的內容，係基於我們目前對未來事件的預期、假設與預測。這些前瞻性陳述存在風險與不確定性，可能導致實際結果與所預測者出現重大差異。關於這些風險與不確定性的更多資訊，請參閱我們最近向美國證券交易委員會（SEC）提交的文件。

We may also refer to certain non-GAAP financial measures. A reconciliation of these non-GAAP measures to the most directly comparable GAAP measures can be found in our earnings release. We'll focus our remarks today on the key highlights and drivers. Additional detail is available in the earnings release.

我們也可能提及某些非 GAAP 財務衡量指標。這些非 GAAP 指標與最直接可比的 GAAP 指標之調節表，可在我們的財報新聞稿中找到。今天我們的重點將放在主要亮點與驅動因素。更多細節請參閱財報新聞稿。

Joining us today are Dave Wright, our Co-Founder and Chief Executive Officer, and Jason Beesley, our Chief Financial Officer. Today's earnings is being webcast, and a replay will be available on our investor relations website following the call. Following our prepared remarks, we will open the call to questions.

今天與會的還有我們的共同創辦人兼執行長 Dave Wright，以及我們的財務長 Jason Beesley。今天的財報電話會議將以網路直播方式進行，會後可於我們的投資人關係網站收看重播。在我們的準備發言結束後，將開放提問。

I'll now turn the call over to our CEO, David Wright. Dave, please go ahead.

接下來我把電話交給我們的執行長 David Wright。Dave，請開始。

Thank you, Hamish, and good afternoon, everyone. We delivered another record quarter. In Q2, revenue grew 47% year-over-year to $877 million. Adjusted EBITDA grew faster still, up 54% to $54 million. That is the fourth consecutive quarter of Adjusted EBITDA outpacing revenue. Before Jason takes you through the details, let me start with net revenue retention, then I will discuss the mix of that revenue. NRR is how we measure how well the machine is working. In Q2, NRR reached another record at 129%, up from 127% last quarter and 118% a year ago. Against a long-term target of 115%, those are pretty extraordinary numbers. We hold ourselves to NRR because it measures what matters most, the outcomes we deliver for our brands. When our brands win, we win.

謝謝你，Hamish，各位下午好。我們再度交出創紀錄的一季。第二季營收年增 47%，達 8.77 億美元。調整後 EBITDA 成長更快，年增 54%，達 5,400 萬美元。這是連續第四季調整後 EBITDA 的成長幅度超過營收。在 Jason 帶各位看細節之前，我先從淨營收留存率（NRR）談起，接著我會討論該營收的組成。NRR 是我們衡量這部機器運作得有多好的方式。第二季 NRR 再創新高，達 129%，高於上季的 127%，也高於一年前的 118%。相較於我們長期目標 115%，這些數字相當驚人。我們以 NRR 自我要求，因為它衡量的是最重要的事——我們為品牌帶來的成果。當我們的品牌夥伴成功，我們就成功。

They stay with us, they expand with us, that record is the most persuasive thing our teams bring to the next brand considering Pattern. Strong partner results create a reinforcing cycle. Rising revenue across new marketplaces and geographies generates both data and logistics scale. The data gives us higher signal density. The scale gives us lower cost and faster transit times across the network. All of it starts and ends with being obsessed with our brands' outcomes. Inside the 47% revenue growth, three strategic highlights are worth calling out. First, international. International revenue grew 87% year-over-year to $110 million, our Q1 ever above $100 million. Second, non-Amazon. Non-Amazon revenue grew 93% year-over-year, with strength across Tmall, TikTok Shop, Walmart, and Coupang. Third, SaaS logistics and other. That line grew 123% year-over-year to $17 million.

他們會與我們長期合作、並與我們一起擴張；這項紀錄是我們團隊在下一個正在考慮 Pattern 的品牌面前，最具說服力的證明。強勁的合作夥伴成果會形成強化循環。在新市場平台與新地理區域的營收成長，同時帶來數據規模與物流規模。數據讓我們擁有更高的訊號密度。規模讓我們在整個網路中具備更低成本與更快的運輸時間。一切的起點與終點，都是對品牌成果的極致投入。在 47% 的營收成長之中，有三項策略亮點值得特別提出。第一，國際業務。國際營收年增 87%，達 1.10 億美元，這是我們首次單季突破 1 億美元。第二，非亞馬遜（Non-Amazon）。非亞馬遜營收年增 93%，在天貓、TikTok Shop、Walmart 與 Coupang 等平台皆表現強勁。第三，SaaS 物流與其他。該項目營收年增 123%，達 1,700 萬美元。

It is still a small share of revenue. It deepens what we do for each brand and gives us optionality as a business. In short, our model is working. We delivered another quarter of record results, we are again raising our outlook for the full year. Jason will walk you through the specifics. Since day one, our objective has been the same, achieve exceptional brand outcomes by optimizing the four levers that drive commerce: traffic, conversion, price, and availability. What makes that repeatable is an ontology, and we believe ours is one of the most robust in e-commerce. The AI models will keep improving. The ontology keeps compounding regardless. Our ontology has three layers and 44 patents issued or pending across them. Number one, the data layer. 91 trillion data points accumulated across 13 years of execution in hundreds of brands, geographies, and marketplace.

它在營收中仍占比不大。但它加深了我們為每個品牌所做的事，也讓我們的業務具備更多選擇性。簡言之，我們的模式正在奏效。我們再度交出創紀錄的季度成果，也再次上調全年展望。Jason 將帶各位了解具體內容。自第一天起，我們的目標始終如一：透過最佳化驅動商務的四大槓桿——流量、轉換、價格與供應可得性——來達成卓越的品牌成果。讓這一切可重複、可規模化的關鍵在於本體（ontology），而我們相信我們的本體是電商領域中最強健之一。AI 模型會持續進步。而本體不論如何都會持續累積與複利成長。我們的本體有三層，並在各層合計已有 44 項專利核發或申請中。第一，數據層：在 13 年的執行中，跨數百個品牌、地理區域與市場平台累積了 91 兆個數據點。

Second, the semantic layer, the entities and the map relationships between them. Price against inventory, competitive position against conversion. This is what makes the data reasoning ready rather than merely stored. Third, the execution layer. Pattern Intelligence, or Pi, which we launched in May. Pi runs a sensor actor framework across those relationships and writes governed actions back to the marketplace, millions a day, on behalf of our brand partners. With Pi's release, brand partners also have interactive visibility into that execution. They can review, approve, and modify inputs. Measurement runs in the execution layer. As part of our advancement in that layer, in Q2, we were awarded a US patent covering True ROAS, our true return on ad sales methodology. True ROAS isolates what an ad actually generated, net of organic conditions, competitive dynamics, and long-term incrementality, so actions can be graded on incrementality, not just attribution.

第二，語意層：各種實體，以及它們之間的關係映射。例如價格相對於庫存、競爭態勢相對於轉換。這使得數據具備可推理性，而不只是被儲存。第三，執行層。也就是我們在 5 月推出的 Pattern Intelligence（Pi）。Pi 在這些關係之上運行感測器—執行器（sensor-actor）框架，並代表我們的品牌夥伴，將受治理的行動寫回到市場平台——每天數以百萬計。隨著 Pi 的發布，品牌夥伴也能以互動方式清楚掌握這些執行情況。他們可以檢視、核准並修改輸入。衡量（measurement）是在執行層中進行。作為我們在該層的進展之一，第二季我們獲得一項美國專利，涵蓋 True ROAS——我們的「真實廣告銷售報酬率」方法論。True ROAS 能隔離出廣告實際帶來的增量效果，扣除自然流量條件、競爭動態與長期增量影響，使行動可以依增量而非僅依歸因來評分。

True ROAS pairs with Destiny, our patented ad tech platform. Measurement feeds allocation drives durable organic ranking. In April, we were named TikTok Shop's Strategic Partner of the Year. More than 100 of our brand partners now sell on TikTok Shop, that number is growing every quarter. Social commerce has become a meaningful channel for new brand partner acquisition, particularly in beauty and fashion. As of last week, through ROI Hunter, which we acquired last December, our brand partners can advertise in ChatGPT. From a single platform, they can reach consumers across Meta, Google, Snap, TikTok, and now ChatGPT. I'd like to give you a few examples of brand successes. We accelerated a US-based prestige skincare brand from $5 million to $15 million in revenue over three years.

True ROAS 與 Destiny（我們已取得專利的廣告科技平台）相互搭配。衡量結果回饋到資源配置，進而驅動更持久的自然排名。4 月，我們獲選為 TikTok Shop 的年度策略合作夥伴。目前已有超過 100 家品牌夥伴在 TikTok Shop 上銷售，而且這個數字每一季都在成長。社群電商已成為取得新品牌夥伴的重要管道，尤其是在美妝與時尚領域。截至上週，透過我們去年 12 月收購的 ROI Hunter，我們的品牌夥伴已可在 ChatGPT 投放廣告。透過單一平台，他們可以觸及 Meta、Google、Snap、TikTok，以及現在的 ChatGPT 上的消費者。我想舉幾個品牌成功案例。我們在三年內，將一家美國高端護膚品牌的營收從 500 萬美元加速提升至 1,500 萬美元。

These results were driven by a combination of improvements, including increasing conversion from 9% to 13%, a 36% lift, improving in-stock from 91%-99%, and subscribe-and-save revenue doubled. One more example. A UK-based sports nutrition brand started with us on a single marketplace in Australia. Today, we manage their e-commerce business across 13 countries, including their flagship market in the UK. That's a pattern we see consistently. Brands start with us in one market and expand globally as their confidence in Pattern grows. Zooming out to our long-term strategic positioning, we are tracking the shift from discovery to transaction within LLMs closely. We are making two long-term investments to position Pattern to win in both. The first is commerce infrastructure as a service. Every agentic transaction has to be fulfilled with real-time inventory, forward and reverse logistics, and customer interactions.

這些成果是由多項改善共同推動，包括轉換率從 9% 提升至 13%（提升 36%）、現貨供應率從 91% 改善至 99%，以及訂閱省（subscribe-and-save）營收翻倍。再舉一個例子。一家以英國為基地的運動營養品牌，最初在澳洲的單一電商平台與我們合作。如今，我們在 13 個國家管理其電商業務，包括其在英國的旗艦市場。這是我們一再看到的模式。品牌先在一個市場與我們合作，隨著對 Pattern 的信心提升而擴展至全球。從更長期的策略定位來看，我們正密切追蹤在 LLM 中從「探索」走向「交易」的轉變。我們正在進行兩項長期投資，使 Pattern 能在兩者中都取得勝利。第一項是「商務基礎設施即服務」（commerce infrastructure as a service）。每一筆代理式（agentic）交易都必須以即時庫存、正向與逆向物流，以及客戶互動來完成履約。

We operate that layer today. We are extending it to agentic shopping. The second is our plan to continue expanding our brand agentic commerce acceleration capabilities, which optimize brands for LLM surfaces and carries that same infrastructure underneath. Pattern is building for both the intelligence to win on LLM surfaces and the infrastructure to meet customer expectations.

我們今天就已在營運這一層。我們正將其延伸到代理式購物（agentic shopping）。第二項是我們計畫持續擴大品牌的代理式商務加速能力，該能力可讓品牌在 LLM 介面（surfaces）上最佳化，並在底層承載同樣的基礎設施。Pattern 正同時打造兩者：在 LLM 介面上取勝所需的智慧能力，以及滿足客戶期待所需的基礎設施。

Before I hand it over to Jason, I'll close with the point I care about most. E-commerce is a team sport. As a matter of fact, all businesses are a team sport. Everything you heard today came from an exceptional team at Pattern. Culture and execution are the same thing. In the last few months, US News & World Report named Pattern one of the best companies to work for in 2026. We also ranked number nine on America's Top 100 Most Loved Workplaces of 2026, our second year in the top 100. I am proud of what we are building and even more proud of the team building it.

在我把時間交給 Jason 之前，我想用我最在意的一點作結。電商是一項團隊運動。事實上，所有企業都是團隊運動。你們今天聽到的一切，都來自 Pattern 一支卓越的團隊。文化與執行是一回事。在過去幾個月，《US News & World Report》將 Pattern 評為 2026 年最佳雇主之一。我們也在《America's Top 100 Most Loved Workplaces of 2026》中排名第九，且連續第二年進入前 100 名。我為我們正在打造的成果感到自豪，更為打造這一切的團隊感到自豪。

Jason, over to you.

Jason，交給你。

Thanks, Dave. Good afternoon, everyone. Q2 was another record quarter for Pattern on many fronts. We continue to see broad-based strength across brand partners, geographies, and marketplaces. We delivered $877 million of revenue, up 47% year-over-year. Adjusted EBITDA grew 54%, outpacing revenue growth for the fourth consecutive quarter. Our performance gives us confidence to increase our full-year outlook for both revenue and adjusted EBITDA. Regarding Q2 growth, I'll start with our biggest revenue driver, existing brand partner revenue. We're excited to report that we delivered another record NRR of 129% for our brand partners, up from 127% in Q1 and 118% a year ago. We have three distinct drivers of that growth. Technology-driven optimization remains the foundation of our growth formula and primary driver. Our unified AI-native intelligence layer monitors and acts across the marketplaces we operate in, driving stronger conversion, traffic, and availability.

謝謝你，Dave。各位下午好。第二季在多個面向上都是 Pattern 的另一個創紀錄季度。我們持續看到品牌合作夥伴、地理區域與各大電商平台的全面性強勁表現。我們實現 8.77 億美元營收，年增 47%。調整後 EBITDA 成長 54%，連續第四個季度增速超越營收成長。我們的表現讓我們有信心上調全年營收與調整後 EBITDA 的展望。談到第二季成長，我先從我們最大的營收驅動因素——既有品牌合作夥伴營收——開始。我們很高興報告，品牌合作夥伴的 NRR（淨收入留存率）再創新高，達 129%，高於第一季的 127%，也高於一年前的 118%。這項成長有三個明確的驅動因素。以科技驅動的最佳化仍是我們成長公式的基礎與主要驅動力。我們統一的 AI 原生智慧層會在我們營運的各個平台上監控並採取行動，帶動更高的轉換、流量與供應可得性。

Because it operates across multiple variables simultaneously, the impact compounds. We also grow by expanding marketplaces and geographies. Embedded in our international revenue growth of 87% to $110 million in the Q2 is a milestone worth noting. This is our Q1 with international revenue above $100 million. One highlight across our international regions is Asia. We entered our first Asian marketplace in 2019, and over the past seven years, we've grown and now operate in 20 marketplaces across China, Hong Kong, Korea, Malaysia, Japan, and Singapore. Not only are we building on our success in existing markets, we are expanding into new markets and continuing to accelerate our growth. On top of our financial success in the region, we also established ourselves as a key partner for domestic marketplaces.

由於它能同時跨多個變數運作，其影響會產生複利效應。我們也透過擴展平台與地理版圖來成長。第二季國際營收成長 87% 至 1.10 億美元，其中包含一個值得注意的里程碑。這是我們首次單季國際營收突破 1 億美元。在我們的國際區域中，一個亮點是亞洲。我們於 2019 年進入第一個亞洲電商平台；在過去七年中，我們持續成長，現已在中國、香港、韓國、馬來西亞、日本與新加坡等地的 20 個平台上營運。我們不僅在既有市場延續成功，也正擴展到新市場並持續加速成長。除了在該區域的財務成功之外，我們也已成為本土電商平台的重要合作夥伴。

For example, Pattern was the only non-China-based company named a Gold Star service provider for the health category by Tmall. Finally, expanding product selection from our brand partners. Introducing more product lines and new products on existing marketplaces is another growth driver. These opportunities come every year and can vary in timing across quarters. We are pleased with revenue growth related to new brand partners across many categories, which tracked at a similar pace to last year. We also grew SaaS, logistics, and other monetization revenue up 123% in Q2 to $17 million. Turning to operating expenses and profitability, adjusted EBITDA was $54 million in Q2, up 54% year-over-year. Of note, we realized costs in the quarter related to Accelerate, our annual global e-commerce summit, startup costs related to our East Coast facility, and increased R&D investment.

例如，Pattern 是天貓在健康品類中唯一被評為金星服務商的非中國本土公司。最後一項是擴大品牌合作夥伴的產品選擇。在既有平台上導入更多產品線與新品，也是另一個成長驅動因素。這些機會每年都會出現，但在各季度之間的發生時點可能不同。我們對多個品類中新品牌合作夥伴帶來的營收成長感到滿意，其增速與去年大致相當。我們的 SaaS、物流及其他變現營收在第二季也成長 123% 至 1,700 萬美元。轉向營運費用與獲利能力，第二季調整後 EBITDA 為 5,400 萬美元，年增 54%。值得注意的是，本季我們認列了與 Accelerate（我們每年舉辦的全球電商高峰會）相關的成本、與東岸設施相關的啟動成本，以及增加的研發投資。

Our East Coast fulfillment facility is now operational, and early throughput is in line with our goals. Excluding stock-based compensation and related taxes, R&D expense was $12 million, up 89% year-over-year. We continue to invest ahead of revenue in our data infrastructure, Pi expansion, and AI capabilities while improving cost leverage in other areas. Variable cost components, cost of goods sold, marketplace commissions, and fulfillment grew slightly slower than revenue, consistent with Q1. This was primarily driven by revenue mix across various products and other monetization strategies. Turning to cash flow. For the trailing 12 months ended June 30th, we generated $136 million of operating cash flow, up 76% year-over-year, and $106 million of free cash flow, up 92% year-over-year. This was driven by our operating results, improved inventory turns, and tax-related benefits from the stock-based compensation expenses recognized at last year's IPO.

我們的東岸履約中心現已投入營運，初期處理量符合我們的目標。扣除股份基礎給付及相關稅負後，研發費用為 1,200 萬美元，年增 89%。我們持續在資料基礎設施、Pi 擴張與 AI 能力上超前於營收進行投資，同時在其他領域改善成本槓桿。變動成本項目——銷貨成本、平台佣金與履約成本——的增幅略低於營收，與第一季一致。這主要是由各類產品的營收組合，以及其他變現策略所驅動。接著談現金流。截至 6 月 30 日的過去 12 個月，我們產生 1.36 億美元的營運現金流，年增 76%；自由現金流為 1.06 億美元，年增 92%。這主要由我們的營運成果、庫存周轉改善，以及去年 IPO 所認列的股份基礎給付費用帶來的稅務相關利益所驅動。

We ended Q2 with $346 million in cash and cash equivalents, no outstanding debt, and $150 million of borrowing capacity. Our balance sheet continues to be a strategic asset. Turning to our outlook. The outperformance in Q2 was broad-based. Existing brand partner revenue acceleration, new brand partner revenue growth, strong non-Amazon international results, and overall healthy execution across the platform. Our recent performance and the momentum we are carrying into the back half gives us confidence to raise our full year outlook. We now expect full year revenue in the range of $3.4 billion to $3.5 billion, representing approximately 37% to 38% growth year-over-year. As I mentioned previously, our year-over-year comps get harder in the H2.

第二季結束時，我們持有 3.46 億美元的現金及約當現金，無任何未償債務，並有 1.50 億美元的借款額度。我們的資產負債表仍是重要的策略性資產。接著談我們的展望。第二季的超預期表現是全面性的。包括既有品牌合作夥伴營收加速、新品牌合作夥伴營收成長、非亞馬遜的國際業務強勁表現，以及整體平台的健康執行力。我們近期的表現，以及我們帶入下半年後段的動能，讓我們有信心上調全年展望。我們目前預期全年營收將落在 34 億至 35 億美元區間，約代表年增 37% 至 38%。如我先前提到的，我們在下半年（H2）的年對年比較基期會更具挑戰。

We will lap the record growth rates from last year in Q3 and Q4. We expect year-over-year revenue growth to moderate to the 30%-plus range, which is reflected in our outlook. We are also raising our full year adjusted EBITDA outlook to approximately $211 million to $213 million, representing approximately 38% to 40% growth year-over-year. We are continuing to grow the company in balance. We expect full year adjusted EBITDA margin accretion even as we continue to accelerate our R&D investment. We are extremely pleased with our NRR performance of 129%, and our updated outlook implies that the ending point of NRR this year will be approximately 123% to 124%, above our long-term target of 115%. We continue to expect NRR to slowly moderate over the next few quarters based on the tougher comparables I already mentioned.

我們將在第三季與第四季面臨去年創紀錄的高成長率基期。我們預期年對年營收成長將放緩至30%以上區間，這也反映在我們的展望之中。我們也將全年調整後EBITDA展望上調至約2.11億至2.13億美元，約代表年對年成長38%至40%。我們持續以均衡的方式推動公司成長。即使我們持續加速研發投資，我們仍預期全年調整後EBITDA利潤率將提升。我們對129%的NRR表現感到非常滿意，而我們更新後的展望意味著今年NRR的期末水準約為123%至124%，高於我們115%的長期目標。基於我先前提到的較嚴峻可比基期，我們仍預期未來幾季NRR將緩步回落。

When looking at Q3, it is important to note that Q2 benefited from large marketplace promotional events such as Amazon Prime Day, Walmart Deals, and Target Circle moving from the Q3 into the Q2 this year. This represented approximately four points of growth shift from Q3-Q2, affecting both revenue and adjusted EBITDA. For the Q3, we expect revenue in the range of $840 million to $860 million, representing approximately 31% to 34% growth year-over-year. We expect Q3 adjusted EBITDA in the range of $51 million to $53 million, growing 25% to 29% year-over-year. In closing, this is our Q4 reporting earnings as a public company. In that time, we've delivered four consecutive quarters of 40%-plus revenue growth with 50%-plus adjusted EBITDA growth. Over that same last 12-month time period, our free cash flow has grown 92%.

在看第三季時，重要的是要注意第二季受惠於大型電商平台促銷活動，例如Amazon Prime Day、Walmart Deals，以及Target Circle今年從第三季移到第二季。這大約造成從第三季到第二季約四個百分點的成長移轉，影響營收與調整後EBITDA。就第三季而言，我們預期營收介於8.40億至8.60億美元，約代表年對年成長31%至34%。我們預期第三季調整後EBITDA介於5,100萬至5,300萬美元，年對年成長25%至29%。最後，這是我們作為上市公司以來的第四次財報電話會議。在這段期間，我們連續四季達成40%以上的營收成長，且調整後EBITDA成長超過50%。在同樣的過去12個月期間，我們的自由現金流成長了92%。

We are delivering significant revenue growth outpaced by adjusted EBITDA and free cash flow growth in a market with significant runway remaining. We believe this is a formula for long-term value creation and puts us in a unique group of companies that grow sustainably at scale. At the end of the day, what matters most is that we are delivering growth for our brand partners. NRR at 129% reflects that.

我們在仍具顯著成長空間的市場中，交出可觀的營收成長，且調整後EBITDA與自由現金流的成長更為強勁。我們相信這是一個創造長期價值的公式，並使我們躋身於一個能在規模化下仍可持續成長的獨特公司群體。歸根結底，最重要的是我們正在為品牌合作夥伴帶來成長。129%的NRR正反映了這一點。

With that, I'll turn it back to Dave before we open up the call for questions.

接下來，在我們開放提問之前，我把時間交回給Dave。

Thanks, Jason. Q2 was our fourth consecutive quarter of 40%-plus revenue growth. Also in Q2, NRR hit a record of 129%. International revenue was above $100 million for the first time. Pi is running at scale. Our brands can now reach consumers on ChatGPT. We enter Q3 with a platform and pipeline we feel great about. The surface area of e-commerce keeps expanding. Pattern will continue to complement brands in their execution and management of these vast surface areas. Pattern is built for both the intelligence to win in that environment and the infrastructure to fulfill what it generates. We remain focused on optimizing the e-commerce equation on behalf of brands, removing friction for brands, and delivering measurable outcomes at scale. Thank you for your continued support.

謝謝你，Jason。第二季是我們連續第四季達成40%以上的營收成長。同時在第二季，NRR創下129%的新高。國際營收也首次突破1億美元。Pi正在以規模化運行。我們的品牌現在可以在ChatGPT上觸及消費者。我們帶著一個讓我們非常有信心的平台與管線進入第三季。電商的觸及面持續擴大。Pattern將持續在這些廣大觸及面的執行與管理上，為品牌提供補強。Pattern同時具備在該環境中致勝所需的智慧能力，以及履行其所帶來成果的基礎設施。我們仍專注於代表品牌最佳化電商方程式、為品牌移除摩擦，並以規模化交付可衡量的成果。感謝各位持續的支持。

We'll now open the call for questions.

我們現在開放提問。

(Operator Instructions) Brian Pitz, BMO Financial Group.

(接線員指示) Brian Pitz，BMO Financial Group。

Maybe with P&G's announced acquisition of Thorne, you can discuss whether a change of control creates any considerations for your relationship with Thorne. Maybe more broadly, how you think about customer retention when brands are acquired by larger strategic owners with more established distribution capabilities. Then I have a follow-up.

或許就P&G宣布收購Thorne一事，你們可以談談控制權變更是否會對你們與Thorne的合作關係帶來任何考量。更廣泛地說，當品牌被擁有更成熟通路能力的大型策略性買家收購時，你們如何看待客戶留存。我還有一個追問。

Appreciate your question. First off, I want to start with a congrats to the Thorne team and the P&G team. Regarding the sale that was announced yesterday, we found out about that at the same time as the market and have no comments on the specifics of that transaction. However, it's important to note that we already work with other brands in the P&G portfolio, and of course, we look forward to working with them in the future. We discussed this yesterday with Thorne management, and we both reiterated the importance of our partnership and our confidence in the future. What it means now is business as usual with Thorne. To your specific question, Brian, there is no change of control provision in our agreement, and we believe we've got runway to demonstrate our value to the partnership going forward.

感謝你的提問。首先，我想先恭喜Thorne團隊以及P&G團隊。關於昨天宣布的出售案，我們與市場同時得知，對該交易的具體內容不予置評。不過，重要的是我們已經與P&G產品組合中的其他品牌合作，當然也期待未來能與他們合作。我們昨天已與Thorne管理團隊討論此事，雙方再次強調我們合作夥伴關係的重要性，以及對未來的信心。目前而言，與Thorne的合作一切照常。針對你的具體問題，Brian，我們的協議中沒有控制權變更條款，我們相信未來仍有空間持續向合作夥伴展現我們的價值。

I'll turn it over to Dave for maybe the second part of your first question and probably a broader context.

我把時間交給Dave，來回應你第一個問題的第二部分，並提供更廣泛的背景。

Yeah. It's a great outcome. If you step back a little, we do what we call a joint business plan with brands at the beginning of all significant time periods. And of course, we have discussions with the brands to say, what are your objectives? What are you hoping for? And I think all brands would agree as they watch Thorne, this has been a successful outcome. And I believe we played a small part in that, and the team has done a tremendous job. One of the things, over the years, we've worked with lots of brands who've had this type of success and have been acquired by larger CPG companies. Some of our best and longest partnerships are inside those CPG conglomerates. Some of them started there, and some of them have landed there via acquisition, successful brand outcomes, I believe.

是的。這是一個很好的結果。如果稍微拉遠來看，我們會在每個重要期間開始時，與品牌一起制定我們所稱的聯合商業計畫。當然，我們會與品牌討論：你們的目標是什麼？你們希望達成什麼？我想所有品牌在看到Thorne的案例時都會同意，這是一個成功的結果。我相信我們在其中扮演了一小部分角色，而團隊做得非常出色。多年來，我們與許多取得這類成功、並被更大型CPG公司收購的品牌合作過。我們一些最優秀、合作最久的夥伴關係就在那些CPG集團之內。其中有些一開始就在那裡，有些則是透過收購、成功的品牌成果而進入那裡，我相信是如此。

I guess one of the things we're excited about is when this happens, generally you have significant knowledge, resources, investment that go into the asset, into the teams. We expect that will come from this partnership. Thorne is just a tremendous brand, high quality. I expect that they'll continue to differentiate themselves with any brands that are just obsessed about their consumers. I think at the end of the day, this will just be a great outcome for the consumers of Thorne. We're excited about the future, and we see it as business as usual.

我想我們感到興奮的一點是，當這種情況發生時，通常會有大量的知識、資源與投資投入到該資產與團隊之中。我們預期這段合作關係也會帶來這些。Thorne是一個非常卓越、品質很高的品牌。我預期他們會持續以對消費者的高度投入，與其他品牌做出差異化。我認為歸根結底，這將會是Thorne消費者的一個很好的結果。我們對未來感到興奮，也認為一切將照常進行。

You mentioned a second question, Brian.

你提到還有第二個問題，Brian。

Yeah, just a real quick one. As you look across your customer base, how are brands balancing either marketplace participation or DTC investment, and what does that imply for the long-term risk of client attrition or channel shift? Any insights there? Because there's obviously a lot of options for some of the brands that are out there.

是的，很快一個。從你們的客戶基礎來看，品牌如何在參與電商平台（marketplace）與投入DTC之間取得平衡？這對長期客戶流失或通路轉移的風險意味著什麼？有任何洞見可以分享嗎？因為顯然市面上一些品牌有很多選項。

Yeah. There is some, of course, conversations you always have where you say, hey, will one cannibalize the other? If a brand is successful on a direct consumer, their site, will it cannibalize marketplace revenue and so forth? We haven't seen that as yet. We're generally always encouraging that collaboration. We believe that when they're successful in any of their channels, it tends to raise marketplace awareness and tends to do a great job there. I don't know if that answers your question. I think our goal is just what's best for a brand in terms of their outcomes, and anything that we can do to support that is where we go. It's usually best for Pattern as well.

是的。當然，總是會有一些對話，你會說，嘿，會不會彼此侵蝕？如果一個品牌在直面消費者（DTC）的自家網站上很成功，會不會侵蝕市集（marketplace）的營收等等？我們目前還沒有看到這種情況。我們通常一直都在鼓勵那種協作。我們相信，當他們在任何一個渠道成功時，往往會提升市集的知名度，並且在那裡也會有很好的表現。我不知道這是否回答了你的問題。我想我們的目標就是：就品牌的成果而言，什麼對品牌最好；而我們能做的任何支持，我們就會往那個方向去做。這通常對 Pattern 也最好。

Thanks for the color.

謝謝你補充的細節。

Eric Sheridan, Goldman Sachs.

Eric Sheridan，高盛。

I know this is a topic we've talked about before. Maybe just to pull on the thread a little bit, given the results you keep putting up. The net revenue retention numbers just keep kind of moving up. Kind of a two-parter. One, what is the signal you're getting about any ceiling that might exist in the business in terms of NRR among the existing and older cohorts that we should be thinking about? Conversely, as the business becomes more diversified over time away from Amazon and into more international markets, how do we think about a countervailing factor of faster growth in newer cohorts or newer verticals as diluting NRR, but also contributing wider to revenue growth for the platform?

我知道這是我們之前談過的主題。也許再順著這條線多問一點，考量到你們持續交出的成績。淨營收留存（NRR）的數字似乎一直在往上走。我有兩個部分的問題。第一，關於既有與較早期客群（cohorts）的 NRR，你們收到的訊號是：這個業務是否存在某個我們應該思考的天花板？反過來，隨著業務隨時間推進而更加多元、逐步降低對 Amazon 的依賴並進入更多國際市場，我們該如何看待一個相反的因素：較新客群或較新垂直領域更快的成長，可能會稀釋 NRR，但同時也會更廣泛地推動平台的營收成長？

The NRR numbers are sort of astounding if you take a look back and you say, okay, total digital growth is, say, 7% to 9%, and we're rolling at 129%. We have some help. Our ability to execute on a technology roadmap, we believe has approximately doubled our expectation from, say, 12 months ago. That is essentially Software Factory, our ability to leverage AI in that process. We can go into some details there if you're interested. Overall, that allows us to hit more marketplaces. If you think of the problem we're solving, we need to solve for a brand, how do we optimize revenue, which will be traffic conversion, availability, price, of course, globally. Hundreds of countries, hundreds of marketplaces. The further we can get down a very complex technology roadmap, the better for brands, better for Pattern.

如果回頭看，NRR 的數字確實有點驚人：整體數位成長可能只有 7% 到 9%，而我們卻跑在 129%。我們有一些助力。我們在技術路線圖上的執行能力，我們相信大約把我們相較於例如 12 個月前的預期翻了一倍。那基本上就是 Software Factory，以及我們在該流程中運用 AI 的能力。如果你有興趣，我們可以再深入一些細節。整體而言，這讓我們能觸及更多市集。如果你想想我們在解決的問題：我們需要替品牌解決的是，如何最佳化營收——也就是流量轉換、供應可得性、價格，當然還包括全球層面。數百個國家、數百個市集。我們在一條非常複雜的技術路線圖上走得越深，對品牌越好，對 Pattern 也越好。

I think you'll just continue to see acceleration there. We just have some tailwinds and some inflection that most companies are probably experiencing right now with AI. Luckily, we're positioned to sort of naturally be good there from day one. We had patents that we had submitted around what you would refer to as AI, back then, just classic machine learning, before it even became interesting. As those technologies have progressed, we've naturally progressed rapidly with it, and we're just seeing that in the results.

我認為你們會持續看到那裡的加速。我們確實有一些順風，以及一個拐點——大多數公司現在可能都正在經歷 AI 帶來的這些變化。幸運的是，我們從第一天起就處在一個自然就能做得好的位置。我們在大家所稱的 AI（當時其實就是傳統機器學習）還沒變得有趣之前，就已經提交過相關專利。隨著這些技術的進展，我們也自然地快速跟進，而我們正從結果中看到這一點。

Maybe, Eric, I'll add just a little bit more on kind of the cohort question of existing and older cohorts and new cohorts. We believe that one of the most powerful things about Pattern is that even brands who have been with us for many, many years can still outgrow the market in a significant way. That is a testament to what Dave talked about on technology optimizations, as well as there are still so many levers to help brands grow by expanding in marketplaces and geographies. To your question about kind of newer cohorts diluting NRR, that's always a possibility, but we're not seeing anything meaningful there. Again, we're just feeling like we're just getting started in a large opportunity set. We're going to keep going at it as fast as we can, expanding and bringing on more brands and more marketplaces to keep that going.

Eric，我再補充一點，關於你提到的既有與較早期客群，以及新客群的 cohort 問題。我們相信 Pattern 最強大的一點之一是：即使是與我們合作了很多很多年的品牌，仍然可以以顯著幅度跑贏市場成長。這一方面印證了 Dave 談到的技術最佳化，另一方面也因為仍有非常多槓桿可以透過擴展到更多市集與地理區域來幫助品牌成長。至於你問到較新客群是否會稀釋 NRR，這永遠有可能，但我們沒有看到任何有意義的影響。再說一次，我們的感覺是：在一個龐大的機會集合裡，我們才剛開始。我們會盡可能快地持續推進，擴張並導入更多品牌與更多市集，讓這個動能延續下去。

Mark Mahaney, Evercore.

Mark Mahaney，Evercore。

Hey, guys. This is Austin Riddick on for Mark Mahaney. Thanks for taking the question. I think non-Amazon revenue of $82 million still implies over 90% of total revenue still runs through Amazon. I just wanted to get your thoughts on where do you see that non-Amazon share going in one to two years? Any color there would be helpful.

嗨，各位。我是 Austin Riddick，代替 Mark Mahaney 提問。謝謝讓我提問。我認為非 Amazon 營收 8,200 萬美元仍然意味著你們超過 90% 的總營收仍然是透過 Amazon 產生。我想聽聽你們的看法：在未來一到兩年，你們預期非 Amazon 的占比會走向哪裡？任何補充說明都會很有幫助。

Okay. Mark, I'll start with that one. It was a little hard to hear you, so I'm going to do my best to kind of fill in the gaps there. Sounded like you were asking about non-Amazon revenue concentration. Where do we see it going? Generally speaking, we love operating on Amazon. It's a great platform, very innovative, great for consumers, great for sellers, things like that. There obviously is a GMV split around the world that is lower in terms of Amazon's percentage of GMV than what we're currently at. As we grow and go where the consumer is at, it's a natural thing for us to diversify away from Amazon. I think some of the stats we shared in the prepared remarks is really strong non-Amazon growth, really strong international growth. Those stats continue to be very strong.

好的。Mark，我先回答這題。剛剛有點聽不清楚，所以我會盡力把缺的部分補上。聽起來你是在問非 Amazon 營收的集中度。我們認為它會往哪裡走？一般來說，我們很喜歡在 Amazon 上營運。它是一個很棒的平台，非常創新，對消費者很好，對賣家也很好，諸如此類。當然，從全球來看，GMV 的分布中，Amazon 的 GMV 占比其實低於我們目前的占比。隨著我們成長並前往消費者所在之處，對我們而言，自然就會逐步分散對 Amazon 的依賴。我想我們在事先準備的講稿裡分享的一些數據顯示：非 Amazon 成長非常強勁，國際成長也非常強勁。這些數據仍然持續非常強。

High double digits, sometimes triple digits on good numbers. I think when you take a step back and even just look at a year ago, our non-Amazon revenue was 7% of our total business, and now it's 9% of our total business. That's in just one year. We've got growth rates in everything excluding Amazon at double or sometimes triple what the growth on Amazon is. We're not providing a specific projection on where that will go, but we like the trend primarily because it's where the customers are trending, and we want to help brands find customers wherever they are.

高雙位數，有時在基期好的情況下甚至是三位數。如果你退一步看，甚至只看一年前，我們的非 Amazon 營收占總業務的 7%，現在是 9%。這只是一年的變化。我們在所有不含 Amazon 的項目上，成長率是 Amazon 成長的兩倍、甚至三倍。我們不提供具體的預測說它會到哪裡，但我們喜歡這個趨勢，主要是因為客戶的趨勢就在那裡，而我們希望幫助品牌在客戶所在的任何地方找到客戶。

Yeah. Maybe just two points of color I'll throw out there. The biggest single line item of the non-Amazon growth, if we break it down by individual marketplace and so forth, is quite simply our SaaS logistics and other bucket. That continues at a pace that's tremendous, gives us a lot of optionality. We're very excited about that piece. One thing that is just somewhat remarkable is to watch a Walmart. Walmart is making tremendous progress. From two years ago, our business on Walmart is up 3.4 times what it was two years ago. That's a combination, of course, our ability to execute there, but Walmart continues to do amazing things. I guess it's just fun to watch the ecosystem evolve. I think we'll be talking more and more about LLMs and agentic surfaces in the years to come.

是的。我再補充兩點說明。如果我們按單一市集等維度拆解非 Amazon 的成長，最大的一個單項其實很簡單：就是我們的 SaaS 物流以及其他（other）那一桶。那一塊持續以非常驚人的速度成長，給了我們很多選擇權。我們對那一塊非常興奮。另外一件相當值得注意的事，是看著 Walmart 的發展。Walmart 正在取得巨大的進展。相較於兩年前，我們在 Walmart 上的業務規模已經成長到兩年前的 3.4 倍。這當然一部分來自我們在那裡的執行能力，但 Walmart 本身也持續做出很了不起的事情。我想看著整個生態系演進，確實很有意思。我認為未來幾年我們會越來越常談到 LLM，以及 agentic surfaces（代理式介面/觸點）。

I'm sure there will be winners that we're talking about and excited to partner with. Again, one of the key strategic points on Pattern is we are agnostic to the channel. We're primarily focused on a brand. As the channels shift and folks do a better job with the consumer, then they will grow, and we will grow with them. As that shifts, we expect to shift. Jason's point on GMV is fantastic, because if you look at GMV around the world, you'll continue to see our diversification quite simply because that's where We started in the US on Amazon largely, and as we continue to get larger and more scale and just more geographic reach, you'll just continue to see those numbers diversify.

我相信一定會有一些贏家是我們會持續關注、並且很期待與之合作的。再說一次，Pattern 的其中一個關鍵策略重點是：我們對通路保持中立。我們主要聚焦在品牌本身。隨著通路轉移、以及各方在消費者經營上做得更好，他們就會成長，而我們也會跟著一起成長。當這些轉變發生時，我們也預期會隨之調整。Jason 關於 GMV 的觀點非常棒，因為如果你看全球的 GMV，你會持續看到我們的多元化；很簡單，因為我們一開始主要是在美國、以 Amazon 為主起步，而隨著我們規模變大、擴張得更有規模、地理覆蓋更廣，你就會持續看到這些數字變得更加分散多元。

Ralph Schackart, William Blair.

Ralph Schackart，William Blair。

Just on the overall growth profile of the business that's really been exceeding expectations since you've gone public. Maybe if you could walk through where you're seeing this really strong outperformance, through the categories that you outlined between technology, I think geography and marketplaces and selection. I think it'd be helpful to understand what's driving outperformance. Maybe a second question related to that is, this business continues to scale rapidly. Can you just kind of walk us through the infrastructure needs that you will need and currently have to support the continued really strong growth.

就整體業務的成長輪廓而言，自從你們上市以來確實一直超出預期。也許你可以帶我們梳理一下，你看到這種強勁超預期表現主要來自哪些地方——在你提到的幾個面向之間，包括技術、我想還有地理、平台與選品。我覺得了解是什麼在驅動超預期表現會很有幫助。第二個相關問題是，這項業務仍在快速擴張。你能否大致帶我們了解一下，為了支撐這種持續且非常強勁的成長，你們需要、以及目前已具備哪些基礎設施？

Great question. Largely, the bulk of our performance always comes from the technology stack. There's just simply not a way to outperform a machine. Especially when you're talking about millions of actions taken a day. It's just impossible to do. It's not just about, hey, how much can we automate to reduce cost? But any automation is often better execution, and you see that in the results. Now, in terms of infrastructure, we have a phenomenal team there, one of the best in the business, I believe. They're continuing to look forward on what is needed. You saw that we launched Bethlehem, Pennsylvania, this year. We have some both hardware and software innovations there that are increasing our throughput, what we believe is currently at 2x, which is sort of astounding given where we already are.

好問題。整體而言，我們表現的大宗來源一直都是技術堆疊（technology stack）。因為根本沒有辦法在表現上超越機器。尤其當你談的是每天數以百萬計的動作。那就是不可能做到的。這不只是關於：我們能自動化多少來降低成本？而是任何自動化往往代表更好的執行，而你會在結果中看到。至於基礎設施，我們那邊有一支非常出色的團隊，我認為是業界最頂尖之一。他們持續前瞻性地評估未來需要什麼。你也看到我們今年在賓夕法尼亞州伯利恆（Bethlehem）啟用了據點。我們在那裡有一些硬體與軟體創新，正在提升我們的吞吐量；我們認為目前大約是 2 倍，考量到我們原本的水準，這相當驚人。

Those teams just continue to ideate and build technology around moving that forward. We have pretty, I would say, you might look at it and call it aggressive, but the team makes sure that they're measured growth in those areas. I don't think you've ever seen us, what you maybe refer to as an overbuild. I think that's just a tribute to that team. They do a phenomenal job there. I think we're in a great spot on infrastructure. Now, we're very bullish on the movement of boxes and this becoming a much more complex problem in the future. The more surface area you get, the more people who win there. Think of forward logistics as one thing, but the number I have, I don't believe it's been independently verified, is 19.3% of all goods globally are returned.

這些團隊持續發想並打造技術來推動這件事往前走。我們有相當——我會說，你可能會覺得偏積極——但團隊確保在這些領域的成長是有節制、可衡量的。我想你從未看過我們出現你可能稱之為「過度建置」（overbuild）的情況。我認為這就是對那支團隊的肯定。他們在那方面做得非常出色。我認為我們在基礎設施上處於很好的位置。另外，我們對於箱件移動（movement of boxes）以及未來這會變成更複雜的問題非常看多。你擁有的觸及面（surface area）越大，在那裡勝出的人就越多。把正向物流（forward logistics）當作一件事，但我手上的數字——我不認為已被獨立驗證——是全球所有商品中有 19.3% 會被退回。

That's a complex problem that a lot of marketplaces have solved, and some are still working to solve it. As we move into agentic surfaces, that problem has to be resolved, and it's a very interesting problem, very fun problem to solve. I think that our commerce infrastructure as a service, we're laying the foundation to do that.

這是一個複雜的問題，很多平台（marketplaces）已經解決了，也有一些仍在努力解決。當我們走向代理式介面（agentic surfaces）時，這個問題必須被解決，而這是一個非常有趣、也非常好玩的問題值得去解。我認為我們的「商務基礎設施即服務」（commerce infrastructure as a service）正在為此打下基礎。

That's helpful.

這很有幫助。

Doug Anmuth, JPMorgan.

Doug Anmuth，JPMorgan。

Hi, this is Maggie on for Doug. Thanks for taking the question. Just following up on that, any chance you could provide some more color on the investments you're making across commerce infrastructure as a service, and then also those agentic commerce acceleration capabilities?

嗨，我是 Maggie，代 Doug 提問。謝謝讓我提問。延續剛才的話題，你們是否能更具體說明一下，你們在「商務基礎設施即服務」方面的投資，以及那些代理式商務加速（agentic commerce acceleration）能力方面的投資？

Sure. Yeah, thanks for the question. On the infrastructure piece, we're very excited about it, but it's sort of a natural problem that we have to solve. It's not particularly interesting, no one's thought of. If you go to buy a good, if you think of what might surround that. Where is my inventory? What do my inventory pools look like? How many units do I have that are close to a consumer? How fast can I get it there? Okay, check that box. Once you have it there, maybe a consumer, maybe they need to redirect it, so there's a process that would be involved there. They might decide, hey, I bought the wrong size, and that would just be an exchange, or maybe it's a full return. There's an entire process around that. Customer experience is very important.

當然可以。是的，謝謝你的問題。在基礎設施這一塊，我們非常興奮，但這其實是我們必須解決的一個自然問題。它並不是什麼特別新奇、沒有人想到的事。如果你要去買一件商品，想想看圍繞著它可能有哪些事情。我的庫存在哪裡？我的庫存池長什麼樣？我有多少件商品離消費者很近？我能多快把它送到那裡？好，這一項打勾。送到之後，消費者可能需要改送到別的地方，因此會有一套流程。他們也可能覺得：我買錯尺寸了，那可能就是換貨，或是完整退貨。這背後有一整套流程。客戶體驗非常重要。

That is infrastructure that we think will be more broadly used. If you think of some of the cost reductions that may come in the future, this is a bit down the line, but we like to think about them anyway. Autonomous driving, robotics. We believe there will be better customer experiences in the future, if you can imagine it. There's already great experience that exists today, but we're investing to be prepared to service that infrastructure for people who aren't quite as advanced as some of the most advanced marketplaces in the world. That's commerce infrastructure. On the agentic commerce enablement and acceleration side, of course, we have some phenomenal advantages. If you think about that problem, More about semantic intent. What you won't find in the SEO world, what you can piece it together if you start concatenating SEO.

我們認為這套基礎設施未來會被更廣泛地使用。如果你去想一些未來可能出現的成本下降——這可能還要一段時間，但我們仍然喜歡提前思考。自動駕駛、機器人。我們相信未來會有更好的客戶體驗，如果你能想像得到的話。今天其實已經有很棒的體驗存在，但我們正在投資，為了能準備好替那些還沒有像全球最先進平台那樣成熟的客戶，提供並服務這套基礎設施。這就是商務基礎設施。至於代理式商務的賦能與加速（agentic commerce enablement and acceleration）這一端，當然，我們有一些非常突出的優勢。如果你思考那個問題，它更關乎語意意圖（semantic intent）。在 SEO 的世界裡你找不到的東西——但如果你開始把 SEO 串接、拼接起來，你就能把它組合出來。

If someone, say, is searching, I use the example of a blanket. They're like, hey, I want to buy a blanket that fits underneath the seat of an airplane. When you build the product descriptions and your product catalogs for that blanket, if you don't include dimensions and size, and someone infers that as their intent to buy the blanket, you will get skipped in an LLM world. The mapping of that intent, which you can map, if you start really thinking deeply about our mode of data around SEO, it is incredibly useful to understand what that mapping might look like. You can probably understand where I'm going here with this. We're positioned very well to help brands accelerate on those surfaces just from a data perspective. It's just execution. We're very excited about the future there.

如果有人在搜尋——我用毯子（blanket）當例子。他們會說：嘿，我想買一條可以放在飛機座位下方的毯子。當你為那條毯子建立產品描述與產品目錄時，如果你沒有包含尺寸與大小，而有人把這個需求推斷為他購買毯子的意圖，那在 LLM 的世界裡你就會被略過。這種意圖的映射（mapping of that intent），如果你開始深入思考我們在 SEO 上的資料模式（mode of data），它對理解這種映射可能長什麼樣子非常有用。你大概可以理解我想表達的方向。僅從資料的角度來看，我們在協助品牌在那些介面上加速方面具備非常好的位置。接下來就是執行。我們對那裡的未來非常興奮。

Bernie McTernan, Needham & Company.

Bernie McTernan，Needham & Company。

Just wanted to ask about margin trends. You mentioned four consecutive quarters of margin expansion with adjusted EBITDA growing faster than revenue. I think the 3.2 guide and the implied 4.2 guide given the annual guidance implies year-over-year margin contraction. Just wanted to see if there's any specific drivers of those trends.

我想請教一下毛利率趨勢。你提到已連續四個季度毛利率擴張，且調整後 EBITDA 的成長快於營收。我認為 3.2 的指引，以及在全年指引下所隱含的 4.2 指引，意味著毛利率將年對年收縮。想確認一下，這些趨勢背後是否有任何特定的驅動因素。

Thanks for the question, Bernie. Appreciate it. I think it's pretty important to just double-click on what I mentioned in the prepared comments about the calendar shift of marketplace events. That causes some noise between Q2 and Q3 on both the revenue and the EBITDA side on a pretty flow through margin basis. When you control for that, both the growth rate in Q3 and the margin rate in Q3 make a lot more sense. Also, when you look at the growth rate in Q3 with that extra four points of growth compared to the growth rate in Q3 last year of 46%, it all starts to line up with what we've been saying all year about tougher comps in the H2. Specifically, in the Q4, there's always some seasonal pressure on EBITDA margin percentage. We see that every year.

謝謝你的提問，Bernie。感謝。我認為很重要的是，針對我在事先準備的講稿中提到的「市集活動的日曆檔期移動」再深入看一下。這會在 Q2 與 Q3 之間，無論在營收端或 EBITDA 端，以相當「隨營收流入」的毛利率基礎造成一些雜訊。當你把這個因素控制住後，Q3 的成長率與 Q3 的毛利率水準就合理多了。另外，當你看 Q3 的成長率，若把那額外的 4 個百分點成長納入考量，並與去年 Q3 46% 的成長率相比，就會更符合我們整年一直在說的：下半年（H2）比較基期更具挑戰。特別是在 Q4，EBITDA 毛利率百分比一向會有一些季節性壓力。我們每年都會看到這種情況。

It's just a bit more expensive to do business in the holiday period, whether that's storage or moving logistics or promotional funding, things like that. That's expected when you look at it from a Q3-Q4 perspective. The only real drag year-over-year is what we've been saying throughout the year, is increased investment in R&D faster than revenue. Overall, you take a step back, you look at the full year and you're still looking at margin accretion on a full year basis, with adjusted EBITDA growing faster than revenue. Interestingly, revenue growing almost at the same rate of what we grew revenue last year on a much smaller base. Feeling pretty good about the overall picture of it, and those are some explanations on your specific questions.

假期期間做生意就是會稍微更貴一些，不管是倉儲、搬運物流，或促銷補貼等這類事情。從 Q3 到 Q4 的角度來看，這是預期之內的。年對年唯一真正的拖累，就是我們整年一直在說的：研發（R&D）投資增速高於營收。整體而言，退一步看全年，你仍然會看到以全年口徑來看毛利率是增加的，且調整後 EBITDA 的成長快於營收。有趣的是，營收成長幾乎與我們去年在更小基期時的營收成長率相當。我們對整體狀況感覺相當不錯，以上也算是對你具體問題的一些說明。

Awesome. Thanks, Jason.

太棒了。謝謝，Jason。

(Operator Instructions) Justin Patterson, KeyBanc.

（接線員指示）Justin Patterson，KeyBanc。

Good afternoon. Dave, it looks like the number of data points on your platform has increased about 38% year-to-date to 91 trillion. Could you talk about how that drives the pace of A/B tests and feeds into product philosophy across the business? It seems like that could be one of the key variables behind just the compounding of the business and the NRR. For Jason, I was hoping you'd talk a little bit more about vertical performance. How's the health and wellness category comparing versus the other verticals you're in?

午安。Dave，看起來你們平台上的資料點數量今年迄今增加了約 38%，達到 91 兆。你能談談這如何推動 A/B 測試的節奏，並如何融入全公司層面的產品理念嗎？這似乎可能是推動業務複利成長與 NRR 的關鍵變數之一。另外想請 Jason 多談一點垂直領域的表現。健康與保健（health and wellness）類別相較於你們其他垂直領域的表現如何？

Of course, we run A/B tests. What we might be best in the world at in e-com, possibly, is just building a framework for measurement. I don't think anyone knows exactly what will work and will not in aggregate for a consumer. You might target a given persona and say, hey, I think I know what will resonate best. But if you have 5 to 10 personas in aggregate, what is the messaging that will drive the most dollars? It's a very difficult problem to solve. Of course, the data and the data moat that we have is invaluable at this point and is continuing to grow. Of course, it just keeps compounding because as our brands have success with us, they expand with us, and then other brands see that success, join the Pattern Pi platform, and that provides us more data.

當然，我們會做 A/B 測試。我們在電商領域可能最世界一流的地方，是建立一套衡量（measurement）的框架。我不認為有人能確切知道，對整體消費者而言，什麼會有效、什麼不會。你可能會鎖定某個特定 persona，然後說：嘿，我覺得我知道什麼最能引起共鳴。但如果你把 5 到 10 個 persona 合在一起看，什麼樣的訊息能帶來最多的營收？這是一個非常難解的問題。當然，我們目前擁有的資料以及資料護城河非常寶貴，而且還在持續擴大。而且它會不斷複利成長，因為當我們的品牌在我們這裡取得成功，他們就會與我們擴大合作；接著其他品牌看到這個成功，就加入 Pattern Pi 平台，這又為我們帶來更多資料。

It's almost this virtuous cycle where we just get better and better, can provide better outcomes for brands, which strengthens the data moat. I think everyone understands that. Of course, there's exceptional nuance in the data, and we get better and better at running any A/B tests. But you almost have to think about them as aggregated A/B tests. I think it just is showing the results.

這幾乎是一個良性循環：我們變得越來越好，能為品牌提供更好的成果，進而強化資料護城河。我想大家都理解這點。當然，資料中有非常細緻的差異，我們在執行各種 A/B 測試上也會越做越好。但你幾乎得把它們視為「彙總後的 A/B 測試」。我認為結果已經反映出來了。

Great. Then just on your question for me, Justin, as it relates to verticals or categories, maybe just as a general reminder, our focus is for every brand to just maximize the outcome that they're trying to get in whatever category they're in. We're in no way managing the category mix from the standpoint of brand results. Even in the world of health and wellness, we still consider ourselves very small when you consider the total GMV, and we love that space. It's great for e-commerce, and obviously you can see, based on transaction yesterday, that there's a lot of interest in that premium space. I would say, there's always great highlights about diversification. One of the ones I'd call out is beauty and TikTok. We mentioned that we were named Strategic Partner of the Year.

很好。那就我這邊回答你的問題，Justin，關於垂直領域或品類，我先提醒一下：我們的重點是讓每個品牌在其所屬的任何品類中，把他們想要達成的成果最大化。從品牌成果的角度，我們完全不會去管理品類組合。即便在健康與保健領域，若以總 GMV 來看，我們仍然認為自己規模很小，但我們很喜歡這個領域。它非常適合電商，而且顯然你也能從昨天的交易看出來，市場對高端（premium）領域有很高的興趣。我會說，多元化總是有很多亮點。其中一個我想特別提的是美妝與 TikTok。我們提到我們被評為年度策略合作夥伴（Strategic Partner of the Year）。

We mentioned that a lot of our beauty inbound is coming through the TikTok channel. The fun part about that is when we do very well for them on TikTok, then we can bring them into the other marketplaces that we represent around the world on that side. As just a data point, beauty grew in the quarter 85%. Pet supplies continues to be fun at over 100% growth. Then in the new business side, and again, I'll just reiterate, this is happening naturally similar to our marketplace diversification, because we're going after so much GMV. We have a target opportunity list of $505 billion that we're attacking with many categories. But that new business, if you look at health and wellness as a percent of the new business, that's lower than the overall business.

我們也提到，我們許多美妝的主動洽詢（inbound）是透過 TikTok 渠道進來的。有趣的是，當我們在 TikTok 上為他們做得非常好之後，就能把他們帶到我們在全球所代表的其他市集中。作為一個數據點，美妝在本季度成長了 85%。寵物用品持續表現亮眼，成長超過 100%。至於新業務方面，我再重申一次：這是自然發生的，類似於我們的市集多元化，因為我們在追逐非常大量的 GMV。我們有一份目標機會清單，規模達 5,050 億美元，我們正以多個品類去攻佔。但若看新業務中健康與保健所占比重，這個比重低於整體業務的占比。

It feels like all of the theses that we had, which is you get into a category, you establish yourself with a track record of performance. That yields a reputation within the category. That can take years, as you have that flywheel of reputation, you sign more brands, you get better and better data and expertise in the category, and it keeps going. We feel like those same green shoots and growth we had in health, and wellness is happening across many other categories, and we like how the business is performing in that space.

感覺我們原先的所有論點都正在發生：你進入一個品類，先用績效表現建立自己的地位。這會在該品類內形成口碑。這可能需要數年時間；當你有了這個口碑飛輪，你就能簽下更多品牌，並在該品類累積越來越好的資料與專業能力，然後持續循環。我們覺得，我們在健康與保健領域看到的那些新芽與成長，也正在許多其他品類同步發生，而我們也喜歡公司在該領域的業務表現。

Colin Sebastian, Baird.

Colin Sebastian，Baird。

I guess two questions for me. First, Dave, on Amazon's call, they talked pretty positively about the performance of their first-party AI interface in terms of conversion rates and overall engagement. Curious, just given some of your efforts with AEO or GEO, if that's an area on the marketplace you're able to take an advantage of. As a second question on Pattern Intelligence, what's the near-term roadmap there to drive more engagement with brands? Is this something that we could see showing up as a measurable growth and margin lever in the relatively near term?

我這邊有兩個問題。第一個，Dave，在亞馬遜的電話會議上，他們相當正面地談到其第一方 AI 介面的表現，就轉換率與整體互動而言。我很好奇，考量你們在 AEO 或 GEO 上的一些努力，這是否是你們能在市集上加以利用的領域。第二個問題關於 Pattern Intelligence，短期內的路線圖是什麼，以推動品牌端更高的互動？這是否有機會在相對不久的時間內，成為可衡量的成長與利潤率槓桿？

Yeah. Our data confirms what Amazon indicated on the call. I guess it's not a surprise. It just allows you to get a better understanding of the consumer and what they're hoping the outcome is for the problem or solution they're looking to solve, which will be a product. The one thing that I think is important, as a general call-out for the future, or I think that you'll probably see a pretty incredible differentiation. The brands that are focused on quality, we're entering a world of much higher transparency.

是的。我們的數據證實了亞馬遜在電話會議上所提到的。我想這並不令人意外。它只是讓你能更好地理解消費者，以及他們希望針對正在嘗試解決的問題或需求，得到什麼樣的結果，而那最終會對應到某個產品。我認為有一點很重要，作為對未來的整體提醒，或者我想你們可能會看到相當驚人的差異化。專注於品質的品牌——我們正在進入一個透明度大幅提高的世界。

The brands that over the years have been obsessed about R&D, customer experiences, they will be paid back on. I think it's probably part of the thesis of why Procter & Gamble acquired Thorne, will be a tremendous asset for them, is that there's many other brands that also fit that category of just over the years, they've focused on great customer experiences, great products. If you think of an LLM world, rather than surfacing hundreds of pages of search results, you're starting to narrow in, okay, what did that customer really intend for? How might I shrink the surface area of what they have to review in order to make a decision, and product quality? All of those things. Also, the infrastructure bit, so say do factor on delivery timelines, the promise.

多年來痴迷於研發、客戶體驗的品牌，將會得到回報。我想這可能也是為什麼寶僑（Procter & Gamble）收購 Thorne 的論點之一；這對他們而言會是巨大的資產，因為還有許多其他品牌也符合這一類型——多年來，他們一直專注於優秀的客戶體驗與優質產品。如果你從 LLM 的世界來看，不再是呈現數百頁的搜尋結果，而是開始收斂到：好，這位客戶真正的意圖是什麼？我如何縮小他們為了做決策所必須檢視的範圍，以及產品品質？所有這些因素。另外還有基礎設施的部分，例如把交付時程、承諾納入考量。

Those things will all become, I think, paramount, and they will stay with the brands for the long term. Will be harder to shake negative experiences. I think where the world is going there will serve consumers better. And so I think that's where you're seeing Amazon's results. The more they invest there, the better their results will be. That's been great to see.

我認為這些都會變得至關重要，而且會長期跟著品牌走。負面體驗將更難被淡化或擺脫。我認為世界往這個方向發展，會更好地服務消費者。因此我想這就是你在亞馬遜的結果中看到的。他們在這方面投資越多，結果就會越好。看到這點很棒。

Maybe specifically on your second question, in terms of Pi, Dave talked about a lot of the roadmap. We're doing 24/7 feature offers, a lot of outcomes happening. It's built on 13 years of everything we've built. It's a better interface layer for the brands with more transparency. Our focus is primarily the effectiveness that comes out of that and driving revenue growth. But there will obviously be efficiencies that will come out of that in the future. We haven't put specific numbers on it, but we just know as we automate and make everything much more agentic, that that will be a natural output. And then what we do with those resources is completely up to us.

針對你的第二個問題更具體地說，就 Pi 而言，Dave 已談了很多路線圖。我們在做 24/7 的功能釋出，有很多成果正在發生。它建立在我們 13 年來所打造的一切之上。它為品牌提供一個更好的介面層，並帶來更高的透明度。我們的重點主要在於由此帶來的成效，以及推動營收成長。但未來顯然也會帶來效率提升。我們尚未給出具體數字，但我們知道，隨著我們自動化並讓一切變得更具代理化（agentic），那將會是自然的產出。接著我們如何運用那些資源，完全取決於我們。

Yeah.

是的。

Got it. Thank you.

了解。謝謝。

One bit of color I might add on the Pi point, Jason. Overall development in general. A unit of work that we think about for a developer would be a pull request or a unit of work that a developer would accomplish. For us, we've doubled those units' year to date as of last year. So you're starting to see Software Factories come into play and accelerate results in general. So I think it's just a fun place to be. The digital economy is growing, and I think we couldn't be more excited about the future, both for Pattern and consumers getting over the next 3, 5, 10 years.

關於 Pi 這點，我可能再補充一點背景，Jason。整體開發而言。我們對開發者所謂的一個工作單位，可能是一個 pull request，或是一個開發者能完成的一個工作單位。就我們而言，截至目前為止，我們相較去年已將這些工作單位翻倍。所以你開始看到軟體工廠（Software Factories）開始發揮作用，並整體加速成果。所以我覺得這是一個很有趣的時點。數位經濟正在成長，而我想我們對未來感到無比興奮——無論是對 Pattern，或是對未來 3、5、10 年的消費者而言。

Thanks, Dave.

謝謝，Dave。

Thank you. This concludes the question-and-answer session. Thank you for your participation in today's conference. This does conclude the program, and you may now disconnect.

謝謝。問答環節到此結束。感謝各位參與今天的會議。本次議程至此結束，您現在可以斷線。