Criteo SA (CRTO) 2025 Q4 法說會逐字稿

Good morning and welcome to Criteo's fourth-quarter and fiscal year 2025 earnings call. (Operator Instructions) Please note this event is being recorded.

早安，歡迎參加 Criteo 2025 財年第四季及全年業績電話會議。（操作說明）請注意，本次活動正在錄影。

I would now like to turn the conference over to Melanie Dambre, Senior Vice President, Investor Relations and Corporate Communications. Please go ahead.

現在我謹將會議交給投資人關係和企業傳播資深副總裁梅蘭妮‧丹布雷 (Melanie Dambre)。請繼續。

Good morning everyone and welcome to Criteo's fourth quarter and fiscal year 2025 earnings call. Joining us on the call today, Chief Executive Officer Michael Komasinski; and Chief Financial Officer, Sarah Glickman are going to share some prepared remarks. Joining us for the Q&A session is Todd Parsons in his role as Chief Product Officer.

各位早安，歡迎參加 Criteo 2025 財年第四季及全年財報電話會議。今天參加電話會議的有首席執行官邁克爾·科馬辛斯基和首席財務官莎拉·格利克曼，他們將分享一些準備好的演講。參加問答環節的是首席產品長托德·帕森斯。

As usual, you will find our investor presentation on our IR website now as well as our prepared remarks and transcripts after the call. Before we get started, I would like to remind you that our remarks will include forward-looking statements which reflect Criteo's judgment, assumptions, and analysis as of today.

像往常一樣，您現在可以在我們的投資者關係網站上找到我們的投資者簡報，以及電話會議結束後我們準備好的發言稿和文字記錄。在正式開始之前，我想提醒各位，我們的發言將包含前瞻性陳述，這些陳述反映了 Criteo 截至今日的判斷、假設和分析。

Our actual results may differ materially from current expectations based on a number of factors affecting Criteo's business. Except as required by law, we do not undertake any obligation to update any forward-looking statements discussed today. For more information, please refer to the risk factors discussed in our earnings release, as well as our most recent forms 10-K and 10-Q filed with the SEC.

由於多種因素影響 Criteo 的業務，我們的實際業績可能與目前的預期有重大差異。除法律另有規定外，我們不承擔更新今天討論的任何前瞻性陳述的義務。有關更多信息，請參閱我們在盈利報告中討論的風險因素，以及我們最近向美國證券交易委員會提交的 10-K 表格和 10-Q 表格。

We will also discuss non-GAAP measures of our performance. Definitions and reconciliations to the most directly comparable GAAP metrics are included in our earnings release published today. Finally, unless otherwise stated, all gross comparisons made during this call are against the same period in the prior year.

我們也將討論我們績效的非GAAP指標。我們今天發布的獲利報告中包含了定義以及與最直接可比較的 GAAP 指標的調整說明。最後，除非另有說明，本次電話會議中所有毛利比較均與去年同期進行比較。

With that, let me now hand it over to Michael.

那麼，現在讓我把麥克風交給麥可。

Thanks, Melanie, and good morning, everyone. 2025 unfolded a bit differently than we had anticipated. Even so, we delivered solid execution and made good progress against our strategy, closing the year with momentum across our key initiatives. One year into my role, we're aligning the company around commerce intelligence and AI decisioning to simplify the business and scale our opportunity as a commerce AI platform for the future of shopping.

謝謝梅蘭妮，大家早安。 2025年的展開和我們預想的有些不同。即便如此，我們依然實現了穩健的執行，並在戰略方面取得了良好進展，以各項關鍵舉措的良好勢頭結束了這一年。上任一年來，我們正在圍繞商業智慧和人工智慧決策來調整公司，以簡化業務並擴大我們作為面向未來購物的商業人工智慧平台的機會。

Consumer attention is more fragmented than ever across websites, apps, social feeds, connected TV, and now AI powered assistants. That fragmentation increases complexity for advertisers, but it creates an opportunity for us to help them navigate it. We reach shoppers wherever they are, with personalized advertising and deliver the outcomes that matter most to brands.

如今，消費者的注意力比以往任何時候都更加分散，他們分散在網站、應用程式、社群媒體、智慧電視，以及現在的AI助理。這種碎片化增加了廣告商面臨的複雜性，但也為我們提供了一個幫助他們應對這種複雜性的機會。無論消費者身在何處，我們都能透過個人化廣告觸達他們，並帶來對品牌最重要的結果。

All of this is powered by our unique commerce data foundation, an AI-driven performance engine that increasingly acts as an orchestration platform and decisioning layer across the ecosystem operating at global scale. Our data foundation gives us visibility into over $1 trillion in e-commerce transactions annually, or roughly $3 billion per day.

這一切都歸功於我們獨特的商業數據基礎，這是一個由人工智慧驅動的性能引擎，它日益成為全球規模生態系統中的編排平台和決策層。我們的數據基礎使我們能夠了解每年超過 1 兆美元的電子商務交易，或每天約 30 億美元。

With reach across more than 3 billion daily active users across channels, including social, and 5 billion product skews, we have a global view of how shoppers discover and buy products. That scale and capability translate into $39 billion of commerce outcomes for more than 17,000 customers, reinforcing our critical role to help them grow their businesses.

我們擁有覆蓋超過 30 億日活躍用戶（包括社群媒體用戶）和 50 億產品偏好的全球視野，能夠了解購物者如何發現和購買產品。這種規模和能力轉化為超過 17,000 家客戶 390 億美元的商業成果，鞏固了我們在幫助他們發展業務方面發揮的關鍵作用。

Our diversified reach continues to stand out as a core strength and a differentiator, driving consistent, measurable returns across channels. Simply put, fragmentation favors platforms that can orchestrate decisions and outcomes holistically. That foundation positions us well for the next major shift shaping our industry, the rise of agentic commerce.

我們多元化的覆蓋範圍仍然是我們的核心優勢和差異化優勢，在各個管道都能帶來持續、可衡量的回報。簡而言之，碎片化有利於能夠整體協調決策和結果的平台。這一基礎使我們能夠很好地應對塑造我們行業的下一個​​重大轉變——代理型商業的興起。

Against this backdrop, we entered 2026 with conviction and focus. Despite low expected growth in 2026, that is already well understood by the market, we are strongly confident in the potential of our business beyond the current transition. Our priorities are clear, and we have a world class team to execute against them. Across each priority, our strategy is consistent: apply proprietary commerce intelligence and AI decisioning to optimize performance at scale.

在此背景下，我們滿懷信心和專注地邁入了2026年。儘管市場普遍預期 2026 年的成長較低，但我們對當前轉型期之後我們業務的潛力充滿信心。我們的工作重點很明確，我們擁有一支世界一流的團隊來執行這些工作。在每個優先事項上，我們的策略都是一致的：運用專有的商業智慧和人工智慧決策來大規模優化績效。

First, we aim to be at the forefront of agentic commerce as AI assistants open new pathways for discovery and purchasing. Second, we are focused on scaling our AI powered performance engine across channels, across the full funnel, and through self-service. Third, we are reinforcing our Retail Media leadership by leveraging AI to help retailers shape the future of commerce and deliver measurable outcomes for brands.

首先，我們的目標是走在智慧商務的前沿，因為人工智慧助理為發現和購買開闢了新的途徑。其次，我們專注於將我們的 AI 驅動的性能引擎擴展到各個管道、整個銷售漏斗以及自助服務管道。第三，我們正在利用人工智慧來鞏固我們在零售媒體領域的領導地位，幫助零售商塑造商業的未來，並為品牌帶來可衡量的成果。

Agentic commerce marks the next evolution of digital shopping, where intelligent assistants increasingly influence how consumers discover and choose products and services. Winning in this environment requires turning intent into outcomes informed by real purchasing behavior. As Large Language Model platforms scale and advertising becomes an important monetization model, delivering relevant product recommendations with measurable outcomes sits squarely in our wheelhouse.

智慧助理商務標誌著數位購物的下一個發展階段，在這個階段，智慧助理將越來越多地影響消費者發現和選擇產品及服務的方式。在這種環境下取得成功需要將意圖轉化為以實際購買行為為基礎的結果。隨著大型語言模型平台規模的擴大和廣告成為重要的獲利模式，提供具有可衡量結果的相關產品推薦正是我們的強項。

We are prioritizing agentic commerce now because our scaled commerce intelligence and AI decisioning give us a path to win, and we are moving decisively across three focused areas.

我們現在優先發展智慧商務，因為我們規模化的商務智慧和人工智慧決策為我們提供了一條通往成功的道路，我們正在三個重點領域果斷推進。

First, we are developing an Agentic Commerce Recommendation Service designed for prospective partners, including LLM platforms and personal shopping agents. The objective is to enable these platforms to surface product recommendations that are not only accurate, but highly relevant to purchase intent. In broad offline testing, our approach outperformed baseline large platform recommendations, delivering materially higher relevance.

首先，我們正在開發一種針對潛在合作夥伴（包括LLM平台和個人購物代理）的代理商務推薦服務。目標是使這些平台能夠提供不僅準確而且與購買意願高度相關的產品推薦。在廣泛的離線測試中，我們的方法優於基準大型平台推薦，提供了顯著更高的相關性。

We observed an average uplift of 60% in prioritizing the products most likely to be purchased. This reflects the value of Criteo's access to real-time purchase data and proprietary AI models optimized for buying behavior. This capability is intended to broaden our partnerships across the ecosystem, and we are already engaged with several potential partners. During the fourth quarter, our proof of concept with an LLM partner advanced into extended testing, and we are preparing additional testing with other partners.

我們觀察到，優先考慮最有可能被購買的產品，平均提升幅度為 60%。這體現了 Criteo 獲取即時購買數據和針對購買行為優化的專有 AI 模型所帶來的價值。這項功能旨在擴大我們在整個生態系統中的合作夥伴關係，我們已經與幾個潛在合作夥伴進行了接洽。第四季度，我們與LLM合作夥伴的概念驗證進入了擴展測試階段，我們正在準備與其他合作夥伴進行額外的測試。

Second, agentic environments create additional monetization opportunities. We are testing new conversational shopping experiences, including conversational ads and sponsored products within retailer agents. We have progressed a proof of concept and are seeing strong client interest, pointing to a path for monetization within these new experiences.

其次，智能體環境創造了額外的獲利機會。我們正在測試新的對話式購物體驗，包括零售商代理商中的對話式廣告和贊助產品。我們已經完成了概念驗證，並看到了客戶的濃厚興趣，這表明這些新體驗存在著盈利的可能性。

Over time, we see meaningful opportunity for sponsored discovery as retailers extend their agentic shopping experiences through retailer-owned apps embedded in LLM platforms. Together, these efforts position agentic commerce as a natural extension of our strategy, reinforcing Criteo's role at the center of discovery, relevance, and measurable outcomes as shopping continues to evolve.

隨著時間的推移，隨著零售商透過嵌入 LLM 平台的零售商自有應用程式擴展其代理購物體驗，我們看到了贊助發現的重大機會。這些努力共同將代理商務定位為我們策略的自然延伸，鞏固了 Criteo 在發現、相關性和可衡量結果的核心地位，而購物也在不斷發展。

And third, we are embedding agentic capabilities directly into our solutions for marketers. Our Model Context Protocol infrastructure allows external agents to interact with Criteo in new ways, enabling more dynamic demand creation, activation, and optimization. Agentic workflows are already live across multiple campaigns and expected to drive broader adoption of our solutions among agency partners and brands.

第三，我們將代理功能直接嵌入到我們為行銷人員提供的解決方案中。我們的模型上下文協議基礎架構允許外部代理以新的方式與 Criteo 進行交互，從而實現更動態的需求創造、激活和最佳化。代理工作流程已在多個行銷活動中投入使用，預計將推動代理商合作夥伴和品牌更廣泛地採用我們的解決方案。

We now have multiple agents live across the platform, including audience and insights agents, helping clients reduce manual effort and move more quickly from insight to action. We are extending these agentic capabilities into campaign creation and reporting to make activation more efficient and scalable. Agentic commerce is a strategic pillar for Criteo in 2026 and a long-term growth opportunity.

我們現在在平台上部署了多個代理，包括受眾代理和洞察代理，幫助客戶減少人工操作，並更快地從洞察轉化為行動。我們將這些代理功能擴展到活動創建和報告方面，以使啟動更加高效和可擴展。智慧商務是 Criteo 在 2026 年的策略支柱，也是其長期成長機會。

Turning to our second priority, scaling our performance engine. We believe Performance Media has a strong future, and we are re-energizing the business through focused execution across three areas: expanding self-service, increasing cross-channel activation, and extending performance further up the funnel to capture discovery budgets.

接下來是我們的第二個優先事項，即擴展我們的效能引擎。我們相信效果媒體擁有光明的前景，我們正在透過在三個領域集中執行來重振這項業務：擴大自助服務、增加跨渠道激活以及將效果營銷進一步延伸到銷售漏斗的上游，以獲取發現預算。

Early traction across each of these levers reinforces our confidence, and the addition of Ed Dinichert as Chief Customer Officer last month further strengthens execution. At the center is GO, our AI-powered automation and optimization tool set that enables advertisers to launch high performing cross-channel campaigns in just five clicks. GO is scaling and progressing as planned, and we remain on track to launch our full self-service offering at the end of the first quarter.

在這些槓桿的各個方面取得的早期進展增強了我們的信心，而上個月 Ed Dinichert 擔任首席客戶長進一步加強了執行力。核心是 GO，我們這款由人工智慧驅動的自動化和優化工具集，它使廣告商只需點擊五次即可啟動高效的跨渠道廣告活動。GO 正在按計劃擴大規模並取得進展，我們仍有望在第一季末推出完整的自助服務產品。

GO campaigns deliver stronger results with higher spend, lower churn, and better return on ad spend. On average, GO campaigns that include social activation deliver more than 20% higher ROAS than traditional campaigns. Today, 37% of GO campaigns include social, supporting our cross-channel strategy.

GO 廣告活動投入較多，顧客流失率更低，廣告支出報酬率更高，進而帶來更強勁的成效。平均而言，包含社交互動的 GO 活動比傳統活動帶來的 ROAS 高出 20% 以上。如今，37% 的 GO 活動都包含社群媒體，這支援了我們的跨通路策略。

Adoption continues to accelerate. In the United States, one in two campaigns from small clients now runs through GO. Our self-service launch expands our addressable market among small and medium-sized business advertisers and represents a significant multi-year growth driver for Performance Media. Cross channel execution remains a key differentiator and an important growth lever for Criteo. We dynamically allocate and optimize spend across channels through a single, performance-focused orchestration platform.

採用速度持續加快。在美國，現在每兩個小客戶的行銷活動中就有一個是透過 GO 進行的。我們的自助服務推出擴大了我們在中小企業廣告客戶中的目標市場，並成為效果媒體未來多年顯著增長的驅動力。跨通路執行仍然是 Criteo 的關鍵差異化優勢和重要的成長槓桿。我們透過以績效為中心的單一協調平台，動態地分配和優化跨通路的支出。

Social continues to scale with double-digit sequential growth in every quarter of 2025, supported by strong momentum with Meta and expanding engagement with additional partners. New video formats launching this quarter for Instagram and Facebook are expected to support continued growth. We are extending performance further up the funnel as brand performance becomes increasingly important.

在 Meta 的強勁發展勢頭和與更多合作夥伴不斷擴大的合作支持下，社交業務在 2025 年的每個季度都保持兩位數的連續增長。本季Instagram和Facebook推出的全新影片格式有望支持其持續成長。隨著品牌表現變得越來越重要，我們將把績效考核範圍進一步擴大到銷售漏斗的上游。

In the first half of 2026, we plan to test new discovery solutions designed to introduce brands to new customers with a broader rollout in the second half of the year. While still early, CTV is a growing part of our mix and represents a multi-year opportunity given our under penetration in one of the fastest growing segments of digital advertising.

2026 年上半年，我們計劃測試新的發現解決方案，旨在向新客戶介紹品牌，並在下半年進行更廣泛的推廣。儘管 CTV 仍處於早期階段，但它在我們廣告組合中所佔的比例越來越大，鑑於我們在數位廣告成長最快的領域之一的滲透率較低，它代表著一個多年的發展機會。

By combining the reach of traditional television with the precision of digital targeting, CTV is emerging as an increasingly important performance channel. We are launching new campaigns that reflect growing enterprise adoption and continue to onboard premium CTV publishers globally. Together, these strategic initiatives are expanding our addressable market and broadening adoption, positioning Performance Media for stronger growth in the second half of the year.

CTV 結合了傳統電視的覆蓋率和數位導向的精準性，正在成為越來越重要的效果行銷管道。我們正在推出新的推廣活動，以反映企業對 CTV 的日益增長的需求，並繼續在全球範圍內吸引優質的 CTV 發行商。這些策略舉措共同擴大了我們的目標市場和用戶群體，為績效媒體在今年下半年實現更強勁的成長奠定了基礎。

Turning to Retail Media. Retail Media is the fastest growing segment of digital advertising and a growth engine for Criteo. We have a clear leadership position and unmatched supply at scale, including 70% of the top 30 retailers in the US and half of the top 30 retailers in EMEA. We are executing with focus and expect revenue to return to growth in the fourth quarter as we move past previously communicated near-term headwinds related to scope changes at two clients.

轉向零售媒體。零售媒體是數位廣告領域成長最快的細分市場，也是 Criteo 的成長引擎。我們擁有清晰的領導地位和無與倫比的規模化供應能力，包括美國前 30 大零售商中的 70% 和 EMEA 前 30 大零售商中的一半。我們正集中精力執行各項工作，預計隨著我們克服先前公佈的與兩家客戶範圍變更相關的近期不利因素，第四季度營收將恢復成長。

As a trusted partner to 235 retailers, we see agentic commerce reshaping how consumers discover products, not where commerce happens. Agentic capabilities are emerging as an incremental layer on top of existing retail experiences, improving discovery while engagement and conversion remain anchored on retailer-owned sites across an expanding set of formats and touchpoints. This dynamic is driving retailers to invest further in their own digital storefronts, unlocking new surfaces for sponsored discovery and monetization, and supporting continued growth in e-commerce traffic.

作為 235 家零售商的值得信賴的合作夥伴，我們看到代理型商務正在重塑消費者發現產品的方式，而不是商務發生的地點。代理功能正在以增量層的形式在現有零售體驗之上湧現，在不斷擴展的形式和接觸點上，提高發現率，同時參與度和轉換率仍然以零售商擁有的網站為基礎。這種動態促使零售商進一步投資於自己的數位店面，從而開闢新的贊助發現和獲利管道，並支持電子商務流量的持續成長。

Retail Media remains integral to retailer economics and margin profiles, reinforcing the durability of the category. We help retailers deliver a superior product discovery experience compared with generic LLM outputs. To support this, we are developing an AI-driven optimization layer that allows retailers to balance organic and sponsored content, expanding monetization opportunities while maintaining full control over product selection and ranking. This positions retailers for agentic commerce and reinforces our role as a long-term strategic partner.

零售媒體仍是零售商經濟和利潤率的重要組成部分，增強了該品類的持久性。我們幫助零售商提供比通用 LLM 輸出更優質的產品發現體驗。為了支持這一點，我們正在開發一個由人工智慧驅動的優化層，使零售商能夠平衡自然內容和贊助內容，擴大獲利機會，同時保持對產品選擇和排名的完全控制。這使零售商在代理型商務中佔據有利地位，並鞏固了我們作為長期戰略合作夥伴的角色。

Agentic systems are designed to preserve this retailer decisioning, whether through retailer chatbots or integrations with third-party LLM platforms. This control is essential for long-term success and creates clear opportunities for Criteo to help retailers maximize value. We are seeing traction across our strategic initiatives, and we expect Retail Media growth in the underlying client base to accelerate in 2026 versus 2025.

代理系統旨在維護零售商的決策，無論是透過零售商聊天機器人還是與第三方 LLM 平台的整合。這種控制對於長期成功至關重要，並為 Criteo 幫助零售商實現價值最大化創造了明確的機會。我們看到各項策略措施都取得了成效，預計 2026 年零售媒體在現有客戶群中的成長速度將比 2025 年加快。

On the demand side, we are enhancing Commerce Max for brands and agencies with new features like search conquesting, advanced analytics, deeper insights, and AI powered optimizations to further simplify holistic activation regardless of budget source. Demand partnerships are unlocking incremental budgets. We completed our Google SA 360 integration in the fourth quarter and saw strong early performance, including 600% ROAS with a leading global CPG brand. We are working closely with Google to build adoption.

在需求方面，我們正在透過搜尋征服、高級分析、更深入的洞察和人工智慧驅動的優化等新功能來增強 Commerce Max 對品牌和代理商的功能，從而進一步簡化整體激活流程，而無需考慮預算來源。需求夥伴關係正在釋放額外的預算。我們在第四季度完成了 Google SA 360 的整合，並取得了強勁的早期業績，包括與一家領先的全球 CPG 品牌實現了 600% 的 ROAS。我們正與Google緊密合作，以提高用戶採用率。

Our Miraki partnership continues to expand long tail demand and support retailers as they scale their marketplaces. We expect our demand partnerships to contribute approximately two points of growth this year across the underlying client base. On the supply side, we are strengthening our global leadership with new wins such as Lidl, Europe's leading retailer, and JB Hi-Fi, a leading Australian consumer electronics retailer, alongside major multi-year retailer renewals.

我們與 Miraki 的合作將繼續擴大長尾需求，並支持零售商擴大其市場規模。我們預計今年我們的需求合作夥伴關係將為我們的基礎客戶群帶來約兩個百分點的成長。在供應方面，我們透過贏得歐洲領先的零售商 Lidl 和澳洲領先的消費電子產品零售商 JB Hi-Fi 等新客戶，以及與多家大型零售商續簽多年合同，進一步鞏固了我們的全球領先地位。

Auction-based display continues to lead our Retail Media growth. Media spend on this new solution increased 65% this quarter, building on last quarter's momentum. Adoption is accelerating with 49 retailers live today and eight new additions this quarter, including Ulta Beauty.

基於競價的展示方式繼續引領我們的零售媒體成長。本季媒體對這項新解決方案的投入增加了 65%，延續了上季的成長動能。目前已有 49 家零售商上線，本季還將新增 8 家，包括 Ulta Beauty，因此該服務的普及速度正在加快。

During Black Friday week, on-site display spend, powered by Criteo's technology, more than doubled year-over-year. We are also rolling out auction-based display to our API demand partners, expanding access to incremental advertiser demand. For retailers using auction-based display, format now accounts for 21% of onsite media spend, up from 13% last year, highlighting strong performance.

在黑色星期五期間，借助 Criteo 的技術，現場展示廣告支出較去年同期成長超過一倍。我們也將向 API 需求合作夥伴推出基於競價的展示廣告，擴大新增廣告主需求的取得管道。對於採用競價展示廣告的零售商而言，該廣告形式目前佔現場媒體支出的 21%，高於去年的 13%，表現強勁。

Shoppable video is also gaining traction, with video spend up 30% sequentially, and acceleration expected in 2026, as retailers move toward full funnel onsite strategies. Offsite, which extends Retail Media beyond retail properties, is becoming more strategic and increasingly always on. One of the world's largest computer brands partnered with us on our largest Commerce Max offsite activation to date this quarter. The campaign reached seven million unique Costco shoppers and delivered more than 2000% ROAS during Cyber Week.

可購物影片也越來越受歡迎，影片支出較上季成長 30%，預計到 2026 年將加速成長，因為零售商正在向全通路網站策略轉型。非現場行銷（將零售媒體擴展到零售物業之外）正變得越來越具有戰略性，並且越來越全天候運作。本季度，全球最大的電腦品牌之一與我們合作，開展了迄今為止規模最大的 Commerce Max 異地推廣活動。該活動涵蓋了 700 萬 Costco 獨立購物者，並在網路星期一期間實現了超過 2000% 的投資回報率。

All of this translated into discipline execution in the fourth quarter, with solid contribution ex-TAC and adjusted EBITDA supported by a strong holiday season. Our capital allocation reflects our confidence in the underlying value of the business and the path ahead. The Board increased our remaining share buyback authorization to up to $200 million and we continue to focus on portfolio and corporate structure optimization with the re-domiciliation process on track.

所有這些都轉化為第四季度的嚴格執行，在強勁的假日季的支撐下，剔除TAC後實現了穩健的貢獻，調整後的EBITDA也保持穩定。我們的資本配置反映了我們對公司內在價值和未來發展道路的信心。董事會將剩餘的股票回購授權額度提高至 2 億美元，我們將繼續專注於投資組合和公司結構優化，重新註冊流程也正在按計畫進行。

Looking ahead, we see meaningful opportunities as we move through 2026. By serving as the AI-driven commerce intelligence and orchestration platform across an increasingly complex ecosystem. Criteo is well positioned to benefit from one of the most important shifts in commerce and advertising. We are confident this will deepen our role with clients and partners and support durable growth and long-term shareholder value.

展望未來，我們看到在 2026 年將迎來許多意義重大的機會。透過作為人工智慧驅動的商業智慧和編排平台，服務日益複雜的生態系統。Criteo 已做好充分準備，將從商業和廣告領域最重要的變革之一中受益。我們相信這將加深我們與客戶和合作夥伴的關係，並支持永續成長和長期股東價值。

With that, I'll hand it over to Sarah for more detail on our financial results and outlook.

接下來，我將把發言權交給莎拉，讓她詳細介紹我們的財務表現和展望。

Thank you, Michael, and good morning everyone. We delivered record results in 2025 with strong margins and robust cash flow generation. Starting with our financial highlights for 2025.

謝謝你，邁克爾，大家早安。2025年，我們取得了創紀錄的業績，利潤率強勁，現金流充裕。首先是我們對2025年的財務展望。

Revenue was $1.9 billion and contribution ex-TAC grew 3.5% at constant currency to $1.2 billion, reflecting a year over year tailwind from foreign currencies of $14 million. In Performance Media, revenue was $1.7 billion and contribution ex-TAC was the $915 million up 4% at constant currency, with our Commerce Growth solution up 5% and AdTech services down 3%.

營收為 19 億美元，扣除 TAC 後的貢獻收入按固定匯率計算增長 3.5% 至 12 億美元，反映出外幣帶來的 1400 萬美元的同比利好。在成效媒體業務方面，營收為 17 億美元，扣除 TAC 後的貢獻收入為 9.15 億美元，以固定匯率計算成長 4%，其中我們的商務成長解決方案成長 5%，廣告科技服務下降 3%。

In Retail Media, revenue was $264 million and contribution ex-TAC was $260 million up 2% year over year at constant currency. Excluding the two clients with previously communicated scope changes, Retail Media contribution ex-TAC grew 16%. We delivered a strong adjusted EBITDA margin of 35%, driven by operational leverage enabled by top-line growth, operational productivity, while we strategically continue to invest in agentic AI to fuel our future growth.

在零售媒體領域，營收為 2.64 億美元，扣除 TAC 後的貢獻收入為 2.6 億美元，以固定匯率計算年增 2%。除先前已溝通過範圍變更的兩家客戶外，零售媒體的貢獻（不含TAC）成長了16%。在營收成長和營運效率提升帶來的營運槓桿作用下，我們實現了 35% 的強勁調整後 EBITDA 利潤率，同時我們將繼續策略性地投資於智慧人工智慧，以推動我們未來的成長。

We delivered free cash flow of $211 million up 16% year over year. This represents 52% of adjusted EBITDA. Our adjusted net income was $253 million and adjusted diluted EPS increased to $4.62 in 2025.

我們實現了 2.11 億美元的自由現金流，年增 16%。這相當於調整後 EBITDA 的 52%。2025 年，我們調整後的淨收入為 2.53 億美元，調整後的攤薄每股收益增至 4.62 美元。

Turning to our fourth quarter performance. Revenue was $541 million and contribution ex-TAC was $330 million including a year over year tailwind from foreign currencies of $8 million. At constant currency Q4 contribution ex-TAC was down 4%, as expected, reflecting a $25 million headwind, or about 700 basis points, related to previously communicated scope changes with two Retail Media clients.

接下來來看看我們第四季的業績。營收為 5.41 億美元，扣除 TAC 後的貢獻收入為 3.3 億美元，其中包括來自外匯的 800 萬美元的年增長。以固定匯率計算，第四季度（不含TAC）的貢獻下降了4%，正如預期，這反映出與先前溝通的兩家零售媒體客戶的範圍變更相關的2500萬美元的不利因素，即約700個基點。

Client retention remains high at 90%, underscoring the resilience of our model. Macro trends remain stable throughout the quarter with a solid holiday season. In Performance Media, revenue was $465 million and contribution ex-TAC was $255 million, up 2% at constant currency. This was driven by our Commerce Growth solution, up 3%, which uses large scale commerce data and AI powered audience modeling to drive cross channel, full funnel activation.

客戶留存率維持在90%的高位，凸顯了我們模式的韌性。本季宏觀趨勢維持穩定，假日季表現強勁。在效果媒體業務方面，營收為 4.65 億美元，扣除 TAC 後的貢獻收入為 2.55 億美元，以固定匯率計算成長 2%。這主要得益於我們的電商成長解決方案，成長了 3%，該方案利用大規模電商資料和 AI 驅動的受眾建模來推動跨通路、全通路活化。

We had a strong Cyber Peak. Travel remained our fastest growing vertical, with growth accelerating to 37%, followed by classifieds which grew 12%. Retail was softer overall, including a 13% decline in department stores and a 12% decline in fashion.

我們曾達到網路高峰期。旅遊仍然是我們成長最快的垂直領域，成長速度加快至 37%，其次是分類廣告，成長了 12%。零售業整體疲軟，其中百貨公司下滑 13%，時尚產業下滑 12%。

Overall, we continue to benefit from a diversified client base and a global footprint. By region, media spend growth accelerated in EMEA, while trends were softer in the US and Asia Pacific. Similar to last quarter, AdTech services reduced Performance Media contribution ex-TAC growth by approximately 100 basis points due to lower spend in our media trading marketplace.

總體而言，我們繼續受益於多元化的客戶群和全球業務佈局。按地區劃分，歐洲、中東和非洲地區的媒體支出成長加速，而美國和亞太地區的成長趨勢則較為疲軟。與上個季度類似，由於媒體交易市場的支出減少，廣告科技服務導致績效媒體貢獻（不包括 TAC）成長下降了約 100 個基點。

In Retail Media, revenue was $76 million and contribution ex-TAC was $75 million reflecting the previously communicated $25 million headwind. Excluding this impact, trends were consistent with last quarter, and contribution ex-TAC grew 20% in Q4 across the underlying client base. This growth was driven by continued strength in Retail Media onsite. We benefited from the traction of our auction-based display offering and newly signed retailers.

在零售媒體領域，收入為 7,600 萬美元，扣除 TAC 後的貢獻收入為 7,500 萬美元，反映了先前公佈的 2,500 萬美元的不利因素。排除此影響後，趨勢與上一季一致，第四季度基礎客戶群的貢獻（不包括 TAC）增加了 20%。這一成長主要得益於零售媒體現場業務的持續強勁發展。我們受益於基於拍賣的展示方式的吸引力以及新簽約零售商。

Growth from existing clients was strong, with the same-retailer contribution ex-TAC retention at 99% or 110%, excluding our largest retailer, driven by multi-year contracts and exclusive partnerships with most of our retailer clients. We had another strong holiday season and saw advertising spend grow in all categories and across all regions during the traditional Cyber 6 peak. Media spend in Q4 grew 25% year over year, as our 4,100 global brands are prioritizing Retail Media as a key channel for their investments to reach relevant audiences and sell more products.

現有客戶的成長勢頭強勁，同店零售商貢獻率（不包括 TAC）達到 99% 或 110%（不包括我們最大的零售商），這主要得益於與大多數零售客戶簽訂的多年合約和獨家合作關係。我們又迎來了一個強勁的假期季，在傳統的網路行銷高峰期（6月），所有類別和所有地區的廣告支出都有所增長。第四季媒體支出年增 25%，因為我們的 4,100 個全球品牌將零售媒體作為其投資的關鍵管道，以接觸相關受眾並銷售更多產品。

We delivered an adjusted EBITDA of $120 million in Q4 2025. Non-GAAP operating expenses increased 12% year over year, driven by planned growth investments and a foreign exchange headwind on our euro-based costs, partially offset by our continued focus on productivity.

我們在 2025 年第四季實現了調整後 EBITDA 1.2 億美元。非GAAP營運費用年增12%，主要受計畫成長投資和以歐元計價的成本受外匯逆風影響，部分被我們持續關註生產力所抵銷。

Moving down the P&L, depreciation and amortization was $31 million in Q4 2025, and share-based compensation expense was $7 million down 67% year over year. Our income from operations was $73 million in Q4 2025, and our net income was $46 million. Our weighted average diluted share count was 53.1 million. This resulted in diluted earnings per share of $0.90 and adjusted diluted earnings per share of $1.30 in Q4 2025. In the fourth quarter we canceled a total of 2.2 million shares.

從損益表來看，2025 年第四季的折舊和攤提費用為 3,100 萬美元，股權激勵費用為 700 萬美元，年減 67%。2025 年第四季度，我們的營業收入為 7,300 萬美元，淨收入為 4,600 萬美元。我們的加權平均稀釋股份數為 5,310 萬股。這使得 2025 年第四季稀釋後每股收益為 0.90 美元，調整後稀釋後每股收益為 1.30 美元。第四季我們共註銷了 220 萬股股票。

We benefit from a strong financial position with solid cash generation and no long-term debt. We had $891 million in total liquidity as of the end of December, which gives us significant financial flexibility to execute our growth and capital allocation strategy. In the fourth quarter we generated $134 million in free cash flow and $161 million in cash from operating activities, driven by disciplined execution and record low days sales outstanding.

我們擁有穩健的財務狀況，現金流充裕，且無長期債務。截至 12 月底，我們擁有 8.91 億美元的流動資金，這為我們執行成長和資本配置策略提供了相當大的財務靈活性。第四季度，我們實現了 1.34 億美元的自由現金流和 1.61 億美元的經營活動現金流，這得益於嚴格的執行和創紀錄的低應收帳款週轉天數。

Criteo continues to be a resilient, cash generative business with the financial strength to invest for growth and return capital to shareholders. In 2025, we deployed $152 million of capital, or 72% of our free cash flow for the year to repurchase 5.4 million shares. As of December 31, 2025, there were $67 million remaining under the current authorized share repurchase program, and the board increased our remaining share buyback authorization to up to $200 million. Our capital allocation strategy demonstrates our confidence in our business strategy, financial strength, and our ongoing commitment to enhance shareholder value.

Criteo 仍然是一家具有韌性、現金流充裕的企業，擁有足夠的財務實力進行投資以實現成長並向股東返還資本。2025 年，我們投入了 1.52 億美元的資金，相當於當年自由現金流的 72%，回購 540 萬股股票。截至 2025 年 12 月 31 日，在目前授權的股票回購計畫下還剩下 6,700 萬美元，董事會將剩餘的股票回購授權額度提高到 2 億美元。我們的資本配置策略體現了我們對自身業務策略、財務實力以及持續提升股東價值的信心。

Turning to our financial outlook, which reflects our expectations as of today, February 11, 2026. For 2026, we expect contribution ex-TAC to be flat to up to 2% at constant currency. As you know, we anticipate low overall growth this year due to Retail Media client scope reductions. Excluding this $75 million headwind, underlying contribution ex-TAC is expected to grow high single-digit. We expect the low point in the first quarter, given the one-time tiered fee revenue recognized in January 2025, and we expect sequential improvement through the year with a return to growth in the second half.

接下來談談我們的財務展望，這反映了我們截至 2026 年 2 月 11 日的預期。預計 2026 年，扣除 TAC 後的貢獻率將保持平穩，以固定匯率計算最高可達 2%。如您所知，由於零售媒體客戶範圍縮小，我們預計今年整體成長將較為緩慢。排除這 7,500 萬美元的不利因素，預計扣除 TAC 後的基本貢獻將實現高個位數成長。我們預計第一季將達到低谷，因為 2025 年 1 月確認了一次性分級收費收入，我們預計全年將逐步改善，並在下半年恢復成長。

Our guidance does not assume any revenue contribution from agentic commerce initiatives, given their early stage. We estimate forex changes to drive a positive year over year impact of about $8 million to $12 million on contribution ex-TAC for the full year.

鑑於代理商務計劃尚處於早期階段，我們的指導意見並未假設這些計劃能帶來任何收入貢獻。我們預計外匯變動將對全年（不含TAC）的貢獻產生約800萬至1200萬美元的正面影響。

In Retail Media, we expect to drive median spend growth ahead of the market, and contribution ex-TAC is expected to decline year over year in the mid to high teens at constant currency due to the $75 million impact from client scope reductions. Excluding the two clients, the underlying Retail Media contribution ex-TAC for growth for 2026 is expected to accelerate into the high teens to 20% range compared to 16% in 2025.

在零售媒體領域，我們預計中位數支出成長將超過市場平均水平，但由於客戶範圍縮減 7,500 萬美元的影響，預計扣除 TAC 後的貢獻按固定匯率計算將同比下降 15% 至 10% 左右。除這兩位客戶外，預計 2026 年零售媒體（不含 TAC）的成長貢獻將加速至 20% 至 10% 的水平，而 2025 年為 16%。

In Performance Media, we expect contribution ex-TAC to grow mid-single-digit at constant currency in 2026. This reflects the expected ramp up of GO over the course of this year, while we are experiencing lower spend in certain categories including fashion and department stores in the US.

在效果媒體領域，我們預期 2026 年以固定匯率計算，扣除 TAC 後的貢獻收入將達到中等個位數成長。這反映了今年 GO 的預期成長，同時，我們在美國某些類別（包括時尚和百貨公司）的支出有所下降。

Overall, we anticipate an adjusted EBITDA margin of approximately 32% to 34% for 2026. This reflects continued disciplined investments in agentic commerce and AI innovation and key growth initiatives, costs associated with our return to office and a foreign exchange headwind on our euro-based costs, partially offset by productivity actions.

總體而言，我們預計 2026 年調整後的 EBITDA 利潤率約為 32% 至 34%。這反映了我們對智慧商務和人工智慧創新以及關鍵成長計畫的持續嚴謹投資，以及我們重返辦公室的相關成本和以歐元計價的成本面臨的外匯逆風，部分被提高生產力的措施所抵消。

We prioritize high return investments that lead to sustainable growth, enabling margin expansion. We anticipate that the investments we are making this year will position us for continued top-line growth and strong cash flow generation for the coming years. We expect a normalized tax rate of 27% to 32% under current rules, driven by our evolving revenue mix and certain one-time items related to our redomiciliation. As mentioned last quarter, we anticipate higher CapEx in 2026, primarily related to the renewal of certain data centers, with total CapEx expected to be approximately $190 million as we continue to invest in and to optimize our AI infrastructure. We expect a free cash flow conversion rate of about 40% of adjusted EBITDA before any non-recurring items. For modeling purposes, we assume a flat number of shares outstanding in 2026.

我們優先進行能帶來永續成長的高回報投資，進而擴大利潤率。我們預計，今年的投資將使我們在未來幾年內持續實現營收成長和強勁的現金流。我們預計，在現行規則下，受不斷變化的收入結構和與公司重新註冊相關的某些一次性項目的影響，正常稅率將達到 27% 至 32%。正如上個季度所提到的，我們預計 2026 年資本支出將增加，主要與某些資料中心的更新有關，隨著我們繼續投資和優化我們的人工智慧基礎設施，預計總資本支出將達到約 1.9 億美元。我們預計，在不計入任何非經常性項目的情況下，自由現金流轉換率約為調整後 EBITDA 的 40%。為了建模，我們假設 2026 年流通股數量保持不變。

For Q1 2026, we expect contribution ex-TAC of $245 million to $250 million down 9% to 11% at constant currency. As a reminder, Q1 is expected to mark the low point of the year, reflecting an approximately $27 million near-term headwind, or about a 10 percentage-point drag on growth related to the two Retail Media clients. Our Q1 guidance also reflects lower spend for department stores in the US, soft trends in Asia-Pac, and continued softness in AdTech services. We estimate forex changes to drive a positive year over year impact of about $8 million to $10 million on contribution ex-TAC in Q1. We expect adjusted EBITDA between $50 million and $55 million in a seasonally low quarter, reflecting lower top-line, the timing of certain corporate initiatives, and a foreign exchange headwind on our euro-based costs of $12 million.

預計 2026 年第一季度，不計 TAC 的貢獻收入為 2.45 億美元至 2.5 億美元，以固定匯率計算下降 9% 至 11%。提醒一下，預計第一季將是今年的低谷，反映出近期約 2700 萬美元的不利因素，或對與兩家零售媒體客戶相關的增長造成約 10 個百分點的拖累。我們第一季的業績預期也反映了美國百貨公司支出減少、亞太地區趨勢疲軟以及廣告科技服務持續疲軟。我們預計匯率變動將對第一季非TAC貢獻產生約800萬至1000萬美元的年比正面影響。我們預計，由於季節性淡季，調整後的 EBITDA 將在 5000 萬美元至 5500 萬美元之間，這反映了營收下降、某些公司舉措的時機以及以歐元計價的成本受到 1200 萬美元的外匯不利影響。

We are pleased that our proposed redomiciliation to Luxembourg and direct Nasdaq listing are progressing as planned with no material tax impact. We continue to expect completion of this redomiciliation to Luxembourg in the third quarter of 2026, subject to shareholder approval later this month and other conditions. Looking ahead, we plan to pursue a further redomiciliation to the United States as early as the first quarter of 2027 to broaden our access to US capital markets.

我們很高興，我們提出的將公司遷至盧森堡並直接在納斯達克上市的計劃正在按計劃進行，不會產生重大的稅務影響。我們仍然預計，在本月稍後獲得股東批准以及滿足其他條件後，該公司將於 2026 年第三季完成遷址至盧森堡。展望未來，我們計劃最早於 2027 年第一季將公司註冊地遷至美國，以擴大我們進入美國資本市場的管道。

In closing, we have strong conviction in our strategy, we are excited for agentic commerce, and we remain focused on disciplined execution and capital allocation while delivering strong margins and cash flow generation.

最後，我們對我們的策略充滿信心，我們對代理商業充滿熱情，我們將繼續專注於嚴謹的執行和資本配置，同時實現強勁的利潤率和現金流。

And with that, I will open up the call for questions.

接下來，我將開始接受提問。

(Operator Instructions)

（操作說明）

Thomas White, Davidson.

湯瑪斯懷特，戴維森。

Great, thanks for taking my question. Maybe two if I could, I guess first off, Michael, on the agentic offerings that you touched on at the top of the prepared remarks, could you maybe just talk a little bit more about kind of the your end targets or sort of prospects I guess for the AI recommendation service and you talked about helping surface sort of more personalized product recommendation.

太好了，謝謝你回答我的問題。如果可以的話，我想問兩個問題。首先，邁克爾，關於你在準備好的演講稿開頭提到的代理產品，你能否再多談談你的最終目標或人工智慧推薦服務的前景？你曾提到幫助用戶發現更個人化的產品推薦。

Like if it's the LLMs, I guess I'm just trying to understand like sort of how realistic it is that you're going to just be able to convince them to I guess help influence, you know the way that their search results look, in such a meaningful, kind of an important way.

例如，如果是法學碩士的話，我只是想弄清楚，你真的能夠說服他們去影響他們的搜尋結果，以如此有意義、如此重要的方式，這到底有多現實。

And then just on the maybe a follow-up on the department store weakness you cited, I'm just curious if there's any, if that's related to the Saks Global bankruptcy or maybe there's something broader going on there and maybe you could just quantify your exposure, Sarah, to Saks and all the various brands it has. Thanks.

關於您提到的百貨公司疲軟的問題，我很好奇這是否與 Saks Global 破產有關，或者是否存在更廣泛的問題。 Sarah，您能否量化一下您對 Saks 及其旗下所有品牌的風險敞口？謝謝。

Yeah, happy to address those, Thomas. Good questions.

是的，我很樂意解答這些問題，湯瑪斯。問得好。

Look, I'll start on the agentic topic and then hand it to Todd, and then Sarah will address the department store question. But, we, as we said in the remarks right, we've broadened out our partnership discussions on the product reco service now to be with multiple partners, and it's driven by the hypothesis that the fidelity and relevancy of product recommendations that comes out of those platforms is essential for them to compete for daily active users, and so it's a really important use case for them and the thesis goes further to say that there's no way for those platforms to surface high-quality recommendations without access to product recommendation engines that are grounded in commerce data, right.

聽著，我先講代理性這個話題，然後交給托德，之後莎拉再講百貨商店的問題。但是，正如我們在剛才的發言中所說，我們已經擴大了產品推薦服務的合作範圍，現在與多家合作夥伴展開合作。這是基於這樣的假設：這些平台提供的產品推薦的準確性和相關性對於它們爭奪每日活躍用戶至關重要。因此，這對它們來說是一個非常重要的用例。我們的論點進一步指出，如果沒有基於商業數據的產品推薦引擎，這些平台就無法提供高品質的推薦。

Semantic sort of scraping and compute alone won't get to the fidelity of answers that are required to really win in the commerce recommendation category. So it is early days in terms of how that revenue model might manifest, but it's important for us to expose the deep capabilities that we have and really embed ourselves deeply in that ecosystem as it evolves, and I'll let Todd talk a little bit about sort of the reco engine itself and kind of where we see that going with the partners.

單靠語意抓取和計算無法獲得在商業推薦領域真正獲勝所需的答案精度。所以，就這種收入模式的最終呈現方式而言，現在還處於早期階段，但對我們來說，重要的是要展現我們所擁有的深厚實力，並在生態系統發展的過程中真正深入地融入其中。接下來，我會讓 Todd 談談推薦引擎本身，以及我們希望它與合作夥伴共同發展的方向。

Yeah, I would just add to that, because the reco service in our system is application agnostic, Tom, it extends our data distribution advantage for merchants across the full ecosystem not just LLM agents. So everything that Michael said about improving the experience and the results of a query in an LLM agent is true. But the service itself is adaptable to all the rest of our cross-channel full funnel mix.

是的，我還要補充一點，因為我們系統中的推薦服務與應用程式無關，Tom，它擴展了我們在整個生態系統中為商家提供的資料分發優勢，而不僅僅是 LLM 代理商。所以，Michael 所說的關於改善 LLM 代理的查詢體驗和結果的一切都是正確的。但這項服務本身可以適應我們所有其他跨通路全通路組合。

So whether social platform CTV or other advertising related services, we're able to deliver outcomes in recommendation with an ad payload or not wherever consumers decide. So the flexibility of the service itself supports multiple monetization models as the agent ecosystems develop, not just the LLMs which gives us a great deal of leverage.

因此，無論是社群平台 CTV 或其他廣告相關服務，無論消費者如何選擇，我們都能夠提供有或沒有廣告負荷的推薦結果。因此，隨著代理生態系統的發展，該服務本身的靈活性支援多種貨幣化模式，而不僅僅是LLM，這給了我們很大的優勢。

Hello Tom. Just to comment on -- or maybe not comment on specific clients, I would say that in our prepared remarks we talked about department stores being down 13%, and also fashion, which is a protocol for us being down 12%, so we do continue to see, I would say headwinds that started in Q4 that will continue into 2026.

你好，湯姆。就具體客戶而言——或者說，不是針對具體客戶——我想說，我們在準備好的發言稿中提到，百貨公司的銷售額下降了 13%，時尚行業的銷售額也下降了 12%，所以我們確實繼續看到，從第四季度開始的不利因素將持續到 2026 年。

Okay, thanks, Michael. Thanks, Todd. Thanks, Sarah.

好的，謝謝你，麥可。謝謝你，托德。謝謝你，莎拉。

Mark Kelley, Stifel.

馬克凱利，史蒂費爾。

Great, thank you. Good morning, everyone. A couple of quick ones or maybe not quick. The first one just on Retail Media specifically, I appreciate all the color about the moving pieces for Q1 and as we move throughout the year. I guess maybe can we put maybe a finer point on how we expect underlying growth to kind of -- is it linear throughout the year or is it more of a step up in Q3 and then Q4 as you add these two new partners and we start to lap the other two partners that changed the way they work with Criteo? I guess that's the first one.

太好了，謝謝。各位早安。幾個簡單的，也可能不簡單。第一篇專門針對零售媒體的文章，我很欣賞文章中對第一季以及全年發展趨勢的詳細描述。我想我們或許可以更精確地說明一下我們對潛在成長的預期——是全年線性成長，還是隨著新增兩家合作夥伴，在第三季和第四季出現更大幅度的成長，從而超越之前兩家改變了與 Criteo 合作方式的合作夥伴？我想那是第一個。

And then the second one, can you just remind us the incremental opportunity in Commerce GO. As you move to a self-serve offering and that kind of removes the minimum spend requirements from some of the cohorts that maybe we're not able to use Criteo in the past, can you just walk through those dynamics for us one more time? Thank you.

其次，您能否提醒我們 Commerce GO 的增量機會？隨著你們轉向自助服務模式，取消了過去可能無法使用 Criteo 的部分用戶的最低消費要求，你能否再為我們詳細解釋一下這些變化？謝謝。

Sure, Mark, I'll just start the Retail Media question then hand off to Sarah, but there are several growth drivers for that business over the course of the year, and Sarah can then sort of talk about phasing, but we continue to scale the products that we introduced last year, in particular the auction auction-based display product. We continue to win new retailers, as you mentioned in the script and actually Q4 tends to be sort of a quieter quarter for winning new retailers, given their focus on holiday operations.

當然，馬克，我先開始談零售媒體的問題，然後交給莎拉。今年該業務有幾個成長動力，莎拉可以談談分階段實施的情況。我們將繼續擴大去年推出的產品規模，特別是基於拍賣的展示產品。正如你在腳本中提到的，我們一直在贏得新的零售商，實際上，鑑於零售商們專注於假日運營，第四季度往往是贏得新零售商比較平靜的季度。

And then, as we get into this year, the investments in Commerce Max, the additional rollout of features like conquesting, like we continue to have a really robust pipeline of products to roll out that are going to drive growth against that scaled client base, so a lot of confidence in how that business will perform even with the headwind that we've talked about. Sarah, you want to talk a little bit about the phasing and assumptions?

然後，隨著我們進入今年，對 Commerce Max 的投資，以及諸如征服客戶等功能的進一步推出，我們繼續擁有非常強大的產品線，這些產品將推動我們不斷擴大的客戶群的增長，因此，即使面臨我們之前提到的不利因素，我們對這項業務的表現也充滿信心。莎拉，你想稍微談談分階段實施和假設條件嗎？

Yeah. We, I believe we put in the slides a breakdown the quarter-by-quarter impact of the of the two largest clients, but it's certainly front loaded and yes, we do anticipate growth kind of pacing throughout the year, so the Q1 and Q2 more impacted, so it's about $27 million in Q1 and then it can start to ramp down, but we do continue to have an impact even into Q4 of next year.

是的。我相信我們在幻燈片中詳細列出了兩個最大客戶每個季度的影響，但影響肯定集中在前期，而且我們預計全年都會保持增長勢頭，因此第一季度和第二季度受到的影響更大，第一季度的影響約為 2700 萬美元，然後可能會開始逐漸減少，但即使到明年第四季度，我們仍然會繼續產生影響。

Also, I would say that Q1 is our smallest quarter and we obviously have more seasonality going into the second half, so there is a bigger impact if you will see on the percentages just given that dynamic.

另外，我想說的是，第一季是我們業績最差的季度，而且很明顯，下半年會受到更多季節性因素的影響，因此，考慮到這種動態因素，百分比會受到更大的影響。

Right, and look, I can just start the Go answer real quick and then hand to Todd for some extra color on that. I think -- when we think about GO, we do get very focused on the SMB opportunity, and the self-service rollout, and I think that's really the sort of new and exciting part. But the more fundamental part that I would share is GO starts to really be an expression of our cross-channel full funnel strategy. With this sort of very highly automated optimization and decisioning engine that sits in.

好的，你看，我可以先快速地給出圍棋的答案，然後交給托德補充一些細節。我認為——當我們想到GO時，我們確實非常關注中小企業的機會和自助服務推廣，我認為這才是真正令人興奮的新部分。但我更想分享的是，GO 真正開始體現我們跨通路全通路行銷策略。內建了這種高度自動化的最佳化和決策引擎。

It really starts to transition us from a managed service offering some self-serve, and as we really stand up supply channels at scale, the efficacy and power of that cross-channel optimization really starts to come through, and I think that's what we'll see across '26, and that actually will be applicable to our entire client base, not just limited to SMB, although that tends to be the focus with the self-service launch, but I believe that that optimization engine really gets expressed through GO in a more powerful way.

它真正開始將我們從託管服務過渡到一些自助服務，隨著我們真正大規模地建立供應鏈渠道，這種跨渠道優化的效率和力量開始真正顯現出來，我認為這就是我們將在 2026 年看到的，而且這實際上將適用於我們所有的客戶群，而不僅限於中小企業，儘管自助服務推出時往往側重於中小企業，但我相信，GO 以更強大的方式將透過 GO 展現出來。

I would only add one point to that which is that as we're seeing clients opt in and use the GO product, we are seeing some of those larger clients take advantage of the fact that they can advertise cross-channel full funnel as Michael said and maintain constant performance for whatever KPI it is that they're advertising against. That's what really sets GO and our cross channel full funnel strategy apart. It's making it impossible to think about all of the complexities of the single channel because the product is helping make those decisions and dynamically reallocating budgets for our advertiser and big and small advertisers are tending to like both of them.

我只想補充一點，那就是隨著我們看到客戶選擇加入並使用 GO 產品，我們看到一些較大的客戶利用了他們可以像 Michael 所說的那樣進行跨渠道全渠道營銷，並保持他們所針對的任何 KPI 的持續表現這一事實。這正是 GO 和我們的跨通路全通路行銷策略真正與眾不同的地方。這使得我們無法考慮單一管道的所有複雜性，因為該產品正在幫助我們做出這些決策，並動態地為我們的廣告商重新分配預算，而大大小小的廣告商似乎都很喜歡這兩者。

Thank you very much. I appreciate it.

非常感謝。謝謝。

Justin Patterson, KeyBanc.

Justin Patterson，KeyBanc。

Great, thank you. Could you expand a little bit more on how you're seeing retailers and brands adopt their own, kind of internal LLM agentic tools right now? I thought it was a good characterization you had of just kind of dimensioning between the broader chatbots that are in the market and then some more retail specific solutions. So would love to hear a little bit more about just what stage those, clients are in their journey.

太好了，謝謝。您能否再詳細談談您目前如何看待零售商和品牌採用他們自己的內部LLM代理工具？我認為你對市場上更廣泛的聊天機器人和一些更針對零售業的解決方案之間的差異做了很好的描述。所以很想了解這些客戶目前處於旅程的哪個階段。

And then just stepping back it does seem like Agentic has gotten very good at -- excuse me just broader AI has gotten very good at personalizing ads for international markets making them hyper localized. So I'd love to hear just kind of what experience Criteo had around that whether it's just local language, adoption, more kind of targeted scenery with an ad, so on and so forth and how that's influencing conversion rates. Thank you.

退一步講，Agentic 似乎在——抱歉，應該說更廣泛的人工智慧——方面做得非常出色，它能夠為國際市場個性化廣告，使其高度本地化。所以我很想聽聽 Criteo 在這方面有什麼經驗，無論是本地語言、使用者採納、廣告的定向投放等等，以及這些因素是如何影響轉換率的。謝謝。

Yes, great questions, Justin. I'll take the kind of AI retail question, and then Todd will take the global personalization question.

是的，賈斯汀，問得好。我將回答有關人工智慧零售的問題，然後 Todd 將回答有關全球個人化的問題。

I think what's interesting about retailers in this sort of AI and agentic conversation is, the starting point, right? They're already very bought in on serving relevant sponsored ads. As a core part of the experience and as part of their business model so that they don't really have any tension about whether that's something they're going to do or not. It's fundamental to retailing and the capabilities that have built up to show sponsored ads in like say on site, those capabilities are directly applicable to the new surfaces and so if you're a retailer. You're not going to sit back and surrender the discovery layer to other platforms for customer journeys that you want to own.

我認為，在人工智慧和智能體對話中，零售商最有趣的地方在於，他們的出發點是什麼？他們已經非常熱衷於投放相關的贊助廣告。作為體驗的核心部分，也是他們商業模式的一部分，所以他們對於是否要這樣做沒有任何顧慮。這對於零售業至關重要，而目前已經建立起在網站上展示贊助廣告的能力，可以直接應用於新的表面，因此，如果您是零售商。你不會坐視不管，將你想要掌控的客戶旅程的發現層拱手讓給其他平台。

And so they are all aggressively investing in shopping agents for more agentic AI enabled front ends that allow discoverability, a more curated experience, and Criteo is right there, powering the way that relevant sponsored opportunities show up in those surfaces. So in terms of innings, it is early innings, but we have several pilots underway with very large scale retail clients around different versions of that whether it's conversational ads, sponsorship inside of chatbots that are on site, or sponsored products that are going to show up in apps that are hosted on the LLM platforms.

因此，他們都在積極投資購物代理，以實現更具智慧化的 AI 前端，從而實現可發現性、更精心策劃的體驗，而 Criteo 正是其中的佼佼者，它推動著相關的贊助機會在這些介面上呈現。所以就目前而言，還處於早期階段，但我們正在與一些大型零售客戶開展多個試點項目，涉及不同版本，包括對話式廣告、網站聊天機器人內的贊助，以及將在 LLM 平台託管的應用程式中顯示的贊助產品。

So there's a few different flavors of that, but they're all rushing to make sure that they maintain share of customer journey, and our relevancy engine and data feeds power that. So we're excited about this, and I think you'll see retailers continue to make really good strides and with this across the year.

所以這方面有好幾種不同的做法，但它們都在爭先恐後地確保自己在客戶旅程中保持份額，而我們的相關性引擎和資料來源則為此提供了支援。所以我們對此感到興奮，我認為你會看到零售商們在今年繼續取得非常好的進展。

Yeah, on the creative piece so it's not just a regional thing that we're seeing show up. It's really affecting our entire business across the board. Just wanted to seize on Michael's mention of GO where we're using AI for not just localization and personalization but auto generated creation whether it is text to image or whether it's image to video. What we're seeing in practice is that our customers that are self serving into GO are using those products incredibly fast and that the results of using those products, take image to video for instance are really driving higher rates of performance across the board. So this is showing up across our whole business but in the self-service context, specifically it's very exciting.

是的，在創作方面，我們看到的不僅是區域性現象。這確實對我們整個業務產生了全面影響。我想藉用邁克爾提到的GO項目談談人工智慧，我們不僅將其用於本地化和個人化，還用於自動生成內容，無論是文字轉圖像還是圖像轉視頻。我們在實踐中看到的是，那些自助使用 GO 的客戶使用這些產品的速度非常快，而且使用這些產品（例如將圖像轉換為視頻）的結果確實全面提高了性能水平。所以這種情況在我們整個業務中都有所體現，但在自助服務方面尤其令人興奮。

Thank you both.

謝謝你們兩位。

Mark Zgutowicz, Benchmark Company.

馬克‧茲古托維奇 (Mark Zgutowicz)，基準公司。

Good morning guys. Thanks for taking the questions. This is Alex on for Mark. First, I was hoping you could qualify or quantify the implied decremental adjusted but down margin impact for Retail Media, predicated within guidance for the year, and perhaps what the incremental investment profile for agentic looks like, this year relative to last.

各位早安。謝謝您回答問題。這裡是Alex替Mark報道。首先，我希望您能根據今年的業績指引，對零售媒體業務隱含的調整後利潤率下降的影響進行定性或量化分析，並說明今年與去年相比，代理業務的增量投資情況如何。

Then second, with offset becoming more of an always on budget, if you will. What level of contribution is implied within guidance this year relative to '25, the large client transition and understanding you're not factoring incrementality from agentic? Thank you.

其次，偏移量正逐漸成為始終存在的預算。今年的績效指引中隱含的貢獻水準相對於 2025 年而言是多少？考慮到大型客戶的轉型以及您沒有將代理業務的增量考慮在內，這些貢獻水準又是多少？謝謝。

Yeah, hello Mark, just on -- we actually put in some slides into the deck related to the walk. We don't fill out the Retail Media related to EBITDA but we do talk about obviously the new, as I say, the more new growth, rate, so we have assumed that including the growth net of Retail Media is about 170 basis points kind of upside to the year. And then you'll see from the walk that we've also incorporated within our EBITDA guidance a growth investment to about 260 basis points kind of going the other way.

是的，你好馬克，我們剛剛——我們實際上在幻燈片中加入了一些與徒步旅行相關的幻燈片。我們不把零售媒體與 EBITDA 聯繫起來，但我們確實談到了新的，正如我所說，更多的新增長，增長率，所以我們假設，包括零售媒體在內的淨增長，對這一年來說大約有 170 個基點的增長空間。然後，從我們的預測中您會看到，我們還在 EBITDA 指導中納入了約 260 個基點的成長投資，這有點反其道而行。

So, our assumption is that we are highly focused on the productivity actions within the entire company to ensure that we -- I would say, mitigate and transform from where we are to move forward and that there has been a deliberate shift and focus into agentic AI investments and more productivity actions in terms of there being less now Criteo sold within our Retail Media pipeline.

因此，我們的假設是，我們高度重視整個公司的生產力提升措施，以確保我們——我想說，從我們目前的狀況開始緩解和轉型，從而向前邁進。我們已經有意識地將重點轉移到智慧人工智慧投資和更多生產力提升措施上，例如，現在在我們的零售媒體管道中，Criteo 的銷售額有所下降。

Got it. Thank you. And I'm just curious if you could elaborate on the implied mix of offsite within guidance this year just direction how to think about incremental drivers in '26, considering, you're not implying or factoring any incrementality from agentic.

知道了。謝謝。我只是好奇，您能否詳細說明今年指導意見中隱含的場外因素，以及在 2026 年如何考慮增量驅動因素，因為您並沒有暗示或考慮任何來自代理商的增量。

Yeah, like we didn't load the plan with any agentic upside from things like the product recommendation engine, but in terms of like retail offsite, I don't think we break that out specifically, but let me comment on just kind of offsite more broadly. The Retail Media offsite market is still in the early innings. Only about 15% of that market is offsite. But it's starting to get traction as platforms mature and look for additional revenue sources beyond sponsored product and on-site opportunities.

是的，就像我們沒有在計劃中加入任何來自產品推薦引擎之類的代理優勢，但就零售場外銷售而言，我認為我們沒有專門列出這部分，但讓我更廣泛地談談場外銷售。零售媒體場外市場仍處於起步階段。只有大約 15% 的市佔率來自異地。但隨著平台日趨成熟，並開始尋求贊助產品和網站機會以外的其他收入來源，這種模式開始獲得關注。

So we've got 40 retailers participating in our offsite program, and we mentioned the great, case study with the global computer brand and Costco in the prepared remarks. It's a great example. Our value prop is differentiated because agencies and brands basically get to approach retail marketing with a fully integrated capability where they can integrate offsite and onsite and have better measurement across how those two are working together, and we leverage unique third-party audiences for offsite. You've got full funnel closed loop measurement.

所以，我們有 40 家零售商參與了我們的場外活動計劃，我們在準備好的演講稿中提到了全球電腦品牌和 Costco 的精彩案例研究。這是一個很好的例子。我們的價值主張與眾不同，因為代理商和品牌基本上可以採用完全整合的能力來進行零售行銷，他們可以整合線上和線下行銷，更好地衡量這兩者是如何協同工作的，而且我們還利用獨特的第三方受眾進行線下行銷。您已完成全漏斗閉環測量。

And we also have several retailers running offsite monetization through our SSP, which gives us another way to approach that market depending on how retailers want to scale that. So, in the SSP we enable brands to access retailer audiences via third-party DSPs if that's the path we go. So early innings, but I think we'll get more focus on this from the industry in '26 and '27, and we've got two paths to execute on that between C-Max and the SSP.

此外，我們還有幾家零售商透過我們的 SSP 平台進行站外變現，這為我們提供了另一種進入該市場的方式，具體取決於零售商希望如何擴展規模。因此，在 SSP 中，如果這是我們選擇的道路，我們將使品牌能夠透過第三方 DSP 接觸零售商受眾。現在還處於早期階段，但我認為在 2026 年和 2027 年，業界會更加關注這一點，而我們在 C-Max 和 SSP 之間有兩條途徑可以實現這一點。

Very helpful, thank you guys.

非常有用，謝謝各位。

Tim Nolan, SSR.

提姆諾蘭，SSR。

Hi, thanks so much for taking the question. I'd like to dig a little bit more into a couple topics that have already come up on the call, and it's really about how do you make money from your clients doing these AI initiatives. So a previous question was about kind of focusing on the retailer client side and these new products that you're launching.

您好，非常感謝您回答這個問題。我想更深入地探討電話會議上已經提到的幾個主題，那就是如何透過客戶進行這些人工智慧專案來賺錢。先前的問題是關於如何專注於零售商客戶方以及你們即將推出的這些新產品。

Can you explain, what is the pricing model? How are you charging for that? Is it a transactional based business as you've typically done, or is there like a service element that you're making money off of?

請問一下，定價模式是什麼？你們打算如何收費？這是像你以往那樣以交易為主的業務，還是也包含一些服務方面的獲利元素？

And then on the kind of the big AI company side. I guess they're not necessarily customers of yours, but you need to be integrated with them to make sure that all these agentic processes work. Can you just help us understand, is it more the retailer traditional customer side that you will make money from, or is there a revenue opportunity from the native AI companies as well? Thanks.

然後是大型人工智慧公司那邊的情況。我猜他們不一定是你的客戶，但你需要與他們整合，以確保所有這些代理商流程都能正常運作。您能否幫我們理解一下，您的獲利主要來自傳統零售商的客戶群，還是也有機會從原生人工智慧公司獲得收入？謝謝。

Yeah, I'm happy to take that. I'm just making a couple of notes here. Yeah, the monetization opportunities obviously vary. So in retail, surfacing sponsored product in those interfaces is largely the same fee structure that we have today for serving those onsite. It's a take rate model, not dissimilar from how that works today for onsite products.

是的，我很樂意接受。我只是在這裡記下一些筆記。是的，獲利機會顯然各不相同。因此，在零售業中，在這些介面上展示贊助產品的收費結構與我們今天在現場提供服務的收費結構基本上相同。這是一種以成交率計算的模式，與目前網站產品的成交率計​​算方式並無太大差異。

As we think about the LLMs, like I said earlier, really the play right now is to get out to an industry leadership position with our offering, get deeply embedded into that ecosystem, and then as they develop their monetization models, we'll figure out how we participate in that. It could be a continuum from paying for that API access, almost like fee for content all the way to some kind of participation in their economic model as their monetization strategies evolve and mature. So it is early days there but there's a continuum of paths forward and there certainly needs to be a monetization on that at some point in time.

正如我之前所說，當我們考慮LLM時，現在的策略是憑藉我們的產品在行業中佔據領導地位，深入融入該生態系統，然後隨著他們開發其盈利模式，我們將弄清楚我們如何參與其中。這可能是一個連續的過程，從付費獲取 API 訪問權限（幾乎就像付費獲取內容一樣），一直到隨著其盈利策略的演變和成熟，以某種方式參與其經濟模式。所以現在還處於早期階段，但未來有很多發展道路，而且在某個時候肯定需要將其商業化。

Todd, you have anything you wanted to add?

托德，你還有什麼要補充的嗎？

Yeah. I would just add one thing, Tim, what we know from operating our business for years is that brands want to be discovered. They want new audiences to discover their product or discover them for the first time. Independent of whether it's a paid or an organic discovery. And what you hear Michael saying is that we have two paths to get to those outcomes of a brand being discovered and their product being bought.

是的。提姆，我只想補充一點，我們多年經營業務的經驗告訴我們，品牌都希望被發現。他們希望新的受眾群體發現他們的產品，或第一次發現他們的產品。無論它是付費發現還是自然發現。麥可的意思是，我們有兩種途徑可以實現品牌被發現和產品被購買的目標。

One is to take advantage of the traditional paid approach that we have done well for years, and that will show up in the LLMs that are developing advertising solutions. And the other one is organic and you're seeing that from the recommendation service today. We can get a brand discovered organically and that's where the pricing model is a little bit less clear. That has not been part of our core business. It is still representative of an outcome for that brand being discovered, and we know we'll monetize it, but it is an organic discovery rather than a paid discovery.

一種方法是利用我們多年來一直做得很好的傳統付費方式，這將在正在開發廣告解決方案的LLM中體現。另一種是自然流量，你今天就能從推薦服務中看到這一點。我們可以透過自然成長的方式讓品牌被發現，而這正是定價模式不太明確的地方。那並非我們的核心業務。這仍然代表著該品牌被發現的結果，我們也知道我們會從中獲利，但這是一種自然發現，而不是付費發現。

So that's why we're a little loose on what we're doing there. The experimentation now is focusing on making sure the discovery works better, and we know the monetization will come right after that.

所以這就是為什麼我們在那裡所做的事情有點隨意的原因。目前的實驗重點是確保這項發現能夠更好地發揮作用，我們知道商業化將在之後緊接著進行。

Thanks. And can I just follow-up to be clear, you mentioned -- I think Sarah, you said there's no assumption of any upside from Agentic tools. I guess you're referring to the new things that you've rolled out recently, some of the announcements that you've made. It feels like you may be holding back on the quantification of that given how enthusiastic you are about some of these tools and processes. Am I reading that correctly?

謝謝。我能否再確認一下，您提到過——我想是莎拉，您說過，代理工具沒有任何好處的假設。我想你指的是你最近推出的一些新東西，以及你發布的一些公告。鑑於你對其中一些工具和流程的熱情，感覺你可能在量化方面有所保留。我理解得對嗎？

I would say yes. But to Todd's point, we haven't monetized it yet, so it's -- we can't assume something that we haven't got, I guess, signed contracts for at this point. So we're very excited about it, and we believe we're in a really strong position to be -- I would say, the AdTech company of choice to partner with agentic commerce, but yes, I would hope that you will see more announcements and more upside going forward.

我會說是的。但正如托德所說，我們還沒有實現盈利，所以——我們不能想當然地認為我們還沒有簽署合約的事情就能成功。所以我們對此感到非常興奮，我們相信我們處於非常有利的地位，可以成為——我可以說是——代理商務首選的廣告技術公司，但當然，我希望你們在未來會看到更多的公告和更多的成長機會。

Great, thank you all.

太好了，謝謝大家。

(Operator Instructions)

（操作說明）

Richard Kramer, Arete Research.

Richard Kramer，Arete Research。

Thanks very much. Michael, maybe you could shed a little light on the lower take rates that we saw across Retail Media and Performance Media. Historically, Q4 would see stronger take rates in both segments, and what was driving that weakness? Is it competitive pressure? Is there some other factor here in terms of wanting to lock down business? I mean, what explains that?

非常感謝。Michael，或許你可以解釋一下我們在零售媒體和效果媒體領域看到的較低轉換率的原因。從歷史數據來看，第四季這兩個細分市場的成交率都會更高，那麼是什麼導致了這種疲軟態勢呢？這是競爭壓力嗎？是否有其他因素導致企業想要封鎖？我的意思是，這該如何解釋？

And maybe one for Sarah. Given the weak first quarter '26 guidance and your comments about Retail Media growth during the year, how much of the '26 budget do you have currently committed or do you have visibility on and how much still needs to be determined over the course of the year? Thanks.

或許還可以給莎拉一份。鑑於 2026 年第一季業績指引疲軟，以及您對零售媒體在年內成長的評論，您目前已承諾或能夠預見 2026 年預算的多少，還有多少需要在年內確定？謝謝。

Yeah, I'll take the take rate questions, and I just want to clarify your second question so I missed some of that. But in terms of take rate, we do continue to have strong take rates within Performance Media, as we move further up the funnel, those take rates change. So there is some mix related to we're serving more social, more CTV, so we kind of are moving into new areas with, I would say just a different mix of take rate, but we feel very good about the margins that we're getting in Performance Media.

是的，我會回答關於接受率的問題，另外，我想澄清一下你的第二個問題，因為我漏掉了一些內容。但就轉換率而言，我們在效果媒體方面仍然保持著強勁的轉換率，隨著我們向行銷漏斗的上游移動，這些轉換率會發生變化。所以，我們採取了一些混合策略，例如更多地投放社群媒體廣告、更多地投放連網電視廣告，我們正在進入新的領域，可以說，廣告投放比例有所不同，但我們對效果媒體的利潤率感到非常滿意。

On Retail Media, we do see the -- I would say the expected compression of the take rate, year on year. That does largely relate to the change that we did in the large retailer client and so that is a part of that impact. And we do also see a mixed issue there as we continue to add more display advertising that would be at a lower take rate than the onsite sponsored ads.

在零售媒體方面，我們確實看到了——我認為是預期的收支率逐年下降。這主要與我們為大型零售客戶所做的改變有關，因此這是其影響的一部分。我們也看到，隨著展示廣告轉換率的不斷提高，其轉換率低於站內贊助廣告，這方面也存在著一些問題。

And just to add one comment on the onsite, we see a massive note for us that we do own the onsite space so further to the questions kind of earlier in terms of as we continue to transition and move into these new spaces, they will likely have a lower take rate, however there's clearly more scale there.

關於現場空間，我想補充一點，我們注意到一個重要的訊息，那就是我們擁有現場空間。因此，對於先前提出的問題，隨著我們不斷過渡並搬入這些新空間，它們的入住率可能會降低，但顯然規模更大。

Could you just repeat the second question? I didn't hear the whole question.

您能再重複一次第二個問題嗎？我沒聽清楚問題的全部內容。

I mean given that obviously you're having a what's going to be a pretty slow start to the year and your comments about growth over the course of the year, how much of the '26 budget that you're looking for is currently committed, or do you have line of sight visibility on and I guess we're all concerned about how that progresses over the course of the year.

我的意思是，鑑於今年開局顯然會比較緩慢，而且你也提到了今年的增長，那麼你目前計劃的 2026 年預算有多少已經落實？或是你對預算的落實有所了解？我想我們都很關心預算在今年的進展。

Yeah, I mean I would say we have very strong forecasting capabilities, but as you know most of our revenue is reoccurring revenue and tends to shift season to season and product by product, so we don't have full visibility to the year's forecast, and that's very usual for us.

是的，我的意思是，我認為我們的預測能力非常強，但正如你所知，我們的大部分收入都是經常性收入，而且往往會隨著季節和產品而變化，所以我們無法完全了解全年的預測情況，這對我們來說非常正常。

However, we do have, especially with Ed coming in as well as the commercial leader, we have a very strong focus on our planning cycle with our large retailers and with our large customers kind of at start of the year, a lot of that starting now either for January or February as they move into their new year.

不過，尤其是在 Ed 和商業領袖加入之後，我們確實非常重視與大型零售商和大客戶的規劃週期，尤其是在年初，很多規劃工作現在就開始了，要么是針對一月份，要么是針對二月份，因為他們即將進入新的一年。

So the feasibility is strong. I would say the AI tools that we have to understand the forecasting along with our clients is also stronger, but same as usual in terms of a high recurring revenue base.

因此，可行性很高。我認為，我們與客戶一起理解預測所需的 AI 工具也更加強大，但在高經常性收入基礎方面，與往常一樣。

Okay. thanks.

好的，謝謝。

Thank you, Michael, Sarah and Todd. That concludes our call for today. Thanks again to everyone for joining. If you have any follow-up questions, the Investor Relations team is available to assist. Have a great day.

謝謝麥可、莎拉和托德。今天的通話就到這裡。再次感謝大家的參與。如果您有任何後續問題，投資者關係團隊隨時為您提供協助。祝你有美好的一天。

The conference has now concluded. Thank you for attending today's presentation. You may now disconnect.

會議已經結束。感謝各位參加今天的報告會。您現在可以斷開連線了。