Confluent Inc (CFLT) 2021 Q3 法說會逐字稿

Hi, everyone. Welcome to The Confluent Q3 2021 Earnings Conference Call. I'm Shane Xie from Investor Relations, and I'm joined by Jay Kreps, Co-Founder and CEO; and Steffan Tomlinson, CFO. During today's call, management will make forward-looking statements, including statements regarding our financial outlook for fiscal fourth quarter of 2021, fiscal year 2021, fiscal first quarter and fiscal year 2022, increased adoption of our platform, our ability and position to capitalize on the shift to cloud, growth in revenue, total customers, remaining performance obligations and dollar-based net retention rate, our market opportunity and our overall future prospects. These forward-looking statements are subject to risks and uncertainties, some of which are beyond our control, which could cause actual results to differ materially from those anticipated by these statements. Further information on risk factors that could cause actual results to differ is included in our SEC filings, including our Form 10-Q for the quarter ended June 30, 2021 and Form 10-K for the quarter ended September 30, 2021, that will be filed with the SEC. We assume no obligation to update these statements after today's call, except as required by law. As a reminder, certain financial measures used on today's call are expressed on a non-GAAP basis. We use these non-GAAP financial measures in generally to facilitate the analysis of financial and business trends and for internal planning and forecasting purposes. These non-GAAP financial measures have limitations. They should not be considered in isolation from or as a substitute for financial information prepared in accordance with GAAP. Our reconciliation between these GAAP and non-GAAP financial measures is included in our earnings press release and supplemental financials, which can be found on our Investor Relations website at investors.confluent.io. And with that, I'll hand the call over to Jay.

大家好。歡迎來到 Confluent 2021 年第三季度收益電話會議。我是 Investor Relations 的 Shane Xie，聯合創始人兼首席執行官 Jay Kreps 加入了我的行列；和首席財務官 Steffan Tomlinson。在今天的電話會議中，管理層將發表前瞻性陳述，包括關於我們對 2021 財年第四季度、2021 財年、第一財季和 2022 財年的財務展望、增加對我們平台的採用、我們利用的能力和地位的陳述向雲的轉變、收入增長、客戶總數、剩餘履約義務和基於美元的淨保留率、我們的市場機會和我們的整體未來前景。這些前瞻性陳述受到風險和不確定性的影響，其中一些超出我們的控制範圍，這可能導致實際結果與這些陳述的預期存在重大差異。有關可能導致實際結果不同的風險因素的更多信息包含在我們的 SEC 文件中，包括我們截至 2021 年 6 月 30 日的季度的 10-Q 表格和截至 2021 年 9 月 30 日的季度的 10-K 表格，這將是向美國證券交易委員會提交。除法律要求外，我們不承擔在今天電話會議後更新這些聲明的義務。提醒一下，今天電話會議中使用的某些財務指標是在非公認會計原則的基礎上表達的。我們通常使用這些非公認會計原則財務措施來促進財務和業務趨勢的分析以及內部規劃和預測目的。這些非公認會計原則財務措施有局限性。不應將它們與根據公認會計原則編制的財務信息分開考慮或替代它們。我們在這些 GAAP 和非 GAAP 財務指標之間的對賬包含在我們的收益新聞稿和補充財務數據中，可在我們的投資者關係網站 Investors.confluent.io 上找到。有了這個，我會把電話交給傑伊。

Thanks, Shane. Welcome, everyone, to our third quarter earnings call. I'm pleased to say our third quarter results exceeded our expectation and our guidance on all metrics, the outperformance we've seen is a result of great product markets fit and secular tailwinds driving the adoption of data in motion and cloud and a lot of hard work by our team. I'll get into these drivers in a moment. But first, let me touch on the numbers. Revenue in the third quarter totaled $102.6 million. This is significant for 2 reasons. First, this is our first quarter ever, achieving over $100 million in quarterly revenue, and we've managed to hit this milestone in less than 7 years. And second, revenue growth of 67% year-over-year represents a further acceleration from the growth we saw last quarter. Growth in Confluent Cloud revenue also accelerated, coming in at 245% year-over-year, while outpacing the growth of the overall business. Confluent Cloud represents 26% of total revenue in the quarter.

謝謝，謝恩。歡迎大家參加我們的第三季度財報電話會議。我很高興地說，我們第三季度的業績超出了我們的預期和我們對所有指標的指導，我們所看到的出色表現是由於產品市場契合度高，長期順風推動了運動和雲數據的採用，以及很多我們團隊的辛勤工作。我稍後會介紹這些驅動程序。但首先，讓我談談數字。第三季度的收入總計為 1.026 億美元。這很重要，原因有兩個。首先，這是我們有史以來的第一個季度，實現了超過 1 億美元的季度收入，而且我們在不到 7 年的時間裡就實現了這一里程碑。其次，收入同比增長 67%，比上一季度的增長進一步加速。 Confluent Cloud 收入的增長也加快了，同比增長 245%，同時超過了整體業務的增長。 Confluent Cloud 佔本季度總收入的 26%。

Our team has continued to execute well, and we feel very good about our position heading into the fourth quarter, which is reflected in our increased guidance for the balance of 2021. Steffan will elaborate further on the financials in a moment.

我們的團隊繼續表現良好，我們對進入第四季度的情況感到非常滿意，這反映在我們增加了對 2021 年餘額的指導中。Steffan 稍後將進一步詳細說明財務狀況。

First, I want to use the opportunity to talk a little bit more about our space. We have a very technical product. And I imagine many of you are still relatively new to our story. So I'll spend a little time today on the data in motion paradigm and our platform before we get into more of the specifics from the third quarter. I think the underlying macro change that's driving Confluent's success is that virtually all businesses are becoming defined end-to-end in software. It isn't just that there are more software applications but increasingly the core activities of the business. The production and distribution of goods and services and interactions with customers are driven by software. To do this, these software systems must connect end-to-end to the different parts of the business and operate in real-time as the business executes out in the real world. This transformation is really being driven by customer demand. In today's world, consumers expect the best digital experiences from the organizations that they interact with. These experiences go beyond simply a user interface or app, but touch everything from logistics to operations to back-end processes that together creates a customer experience. At the core of all the software that drives this is data. Data management has traditionally been about storage, that is, data at rest. Databases have long been the mainstay of data management and storage is exactly what they're built to do. They help you safely store your data and look up or process the right bits as you need it. But now the explosion of software and data means that there are more databases, applications, analytics, platforms and SaaS layers. And increasingly, all of these systems need to work together in real-time to help operate the business. It isn't enough anymore to have all these systems existing as disconnected storage silos. They all have to integrate in real time. Confluent's role is extending support in the infrastructure layer to cover the other half of data management, the management of data in motion. The databases have traditionally ignored. This problem is very different from a world with data at rest. It isn't about storing piles of data for individual disconnected applications. Data in motion requires a platform that helps support the real-time flow and processing of data between applications as it is generated to help drive the operations of the business. This new platform forms a key component of the emerging next-gen data architecture, which we think, over time, will be a requirement for every company. By using data in this way, companies can make it far easier to tap into data from any part of the company and operate on it in real time. The streams of data both trigger custom applications to take action as well as enabling real-time processing and enrichment of the data and integration into other data systems. Data in motion makes it possible to do this in a decoupled way because of the published subscriber nature of the usage. Any part of the business can publish the real-time stream of data at its own operation. And any other part of the business can choose to tap into that stream to react or processing. As a result, these capabilities are critical for companies to be able to build around software and data, which is, in turn, critical for their success in the modern competitive landscape. In this sense, the rise of data in motion is right at the heart of digital transformation. This kind of real-time architecture has long been desirable. What was long believed to be too difficult for all but the most critical applications. However, our platform makes real-time easy. Indeed, our goal is to make processing real-time data streams as easy and ubiquitous as the infrastructure around storage and batch process.

首先，我想藉此機會多談談我們的空間。我們有一個技術含量很高的產品。我想你們中的許多人對我們的故事還比較陌生。因此，在我們深入了解第三季度的更多細節之前，我今天將花一點時間在運動數據范式和我們的平台上。我認為推動 Confluent 成功的潛在宏觀變化是，幾乎所有業務都在軟件中定義為端到端。不僅僅是軟件應用程序越來越多，業務的核心活動也越來越多。商品和服務的生產和分銷以及與客戶的互動是由軟件驅動的。為此，這些軟件系統必須將端到端連接到業務的不同部分，並在業務在現實世界中執行時實時運行。這種轉變實際上是由客戶需求驅動的。在當今世界，消費者期望與其互動的組織提供最佳的數字體驗。這些體驗不僅僅是簡單的用戶界面或應用程序，而是涉及從物流到運營再到後端流程的方方面面，共同創造客戶體驗。驅動這一點的所有軟件的核心是數據。數據管理傳統上是關於存儲的，即靜態數據。長期以來，數據庫一直是數據管理的支柱，而存儲正是它們的目的所在。它們可幫助您安全地存儲數據並根據需要查找或處理正確的位。但現在軟件和數據的爆炸式增長意味著有更多的數據庫、應用程序、分析、平台和 SaaS 層。而且，所有這些系統都需要實時協同工作，以幫助運營業務。將所有這些系統作為斷開連接的存儲孤島存在是不夠的。它們都必須實時集成。 Confluent 的作用是擴展基礎設施層的支持，以覆蓋數據管理的另一半，即動態數據的管理。數據庫傳統上被忽略了。這個問題與數據靜止的世界非常不同。這不是為單個斷開連接的應用程序存儲大量數據。動態數據需要一個平台來幫助支持應用程序之間數據的實時流動和處理，因為它的生成有助於推動業務運營。這個新平台構成了新興的下一代數據架構的關鍵組成部分，我們認為，隨著時間的推移，這將成為每家公司的要求。通過以這種方式使用數據，公司可以更輕鬆地利用公司任何部分的數據並對其進行實時操作。數據流既可以觸發自定義應用程序採取行動，也可以實時處理和豐富數據並集成到其他數據系統中。由於使用的已發布訂閱者性質，動態數據可以以解耦的方式執行此操作。業務的任何部分都可以在自己的操作中發布實時數據流。業務的任何其他部分都可以選擇利用該流進行反應或處理。因此，這些能力對於公司能夠圍繞軟件和數據進行構建至關重要，而這反過來又對於他們在現代競爭格局中的成功至關重要。從這個意義上說，動態數據的興起正是數字化轉型的核心。這種實時架構一直很受歡迎。長期以來，人們認為除了最關鍵的應用程序之外，這對所有人來說都太難了。但是，我們的平台使實時變得容易。事實上，我們的目標是讓處理實時數據流像圍繞存儲和批處理的基礎設施一樣簡單和無處不在。

Our product strategy is laser-focused on this goal. To do this, we're building around 3 key pillars; being cognitive, being a complete offering for data in motion and being everywhere. These 3 pillars represent key aspects of making this new paradigm, the new default standard as well as key aspects of differentiation to separate us from competitors. We think this differentiation is key to our success in driving Confluent's high growth, and in particular, is the driving factor behind the outsized growth of our cloud offering.

我們的產品戰略專注於這一目標。為此，我們圍繞 3 個關鍵支柱進行構建；具有認知性，是動態數據的完整產品，並且無處不在。這 3 個支柱代表了製定這種新範式、新默認標準的關鍵方面，以及區分我們與競爭對手的關鍵方面。我們認為這種差異化是我們成功推動 Confluent 高速增長的關鍵，尤其是我們的雲產品大幅增長背後的驅動因素。

Let me go through each of these 3 pillars in more depth. First, being cloud-native. Confluent Cloud was architected from the ground up to operate as the last rescalable multi-tenant cloud service. These capabilities can sound superficial, but a true cloud-native service is far more than just putting open source Apache Kafka in the cloud. Cloud services are fundamentally different from solutions that might have been designed for an on-premise environment and different in a way that is transformative for the customers who adopt them. This flexibility gets customers out of the world in which their infrastructure dictates their pace of innovation. What's more of this difference isn't just added features. It cuts right to the heart of how these systems are designed and built. By being cloud-native we're able to offer the same protocol of Kafka on a platform that is completely serverless and elastic and also integrates natively with

讓我更深入地介紹這三個支柱中的每一個。首先，雲原生。 Confluent Cloud 是從頭開始構建的，可作為最後一個可重新擴展的多租戶雲服務運行。這些功能聽起來很膚淺，但真正的雲原生服務不僅僅是將開源 Apache Kafka 放在雲端。雲服務與可能為本地環境設計的解決方案根本不同，並且在採用它們的客戶方面具有變革性的方式不同。這種靈活性使客戶擺脫了基礎設施決定創新步伐的世界。更重要的是，這種差異不僅僅是增加的功能。它切入了這些系統的設計和構建方式的核心。通過雲原生，我們能夠在完全無服務器和彈性的平台上提供與 Kafka 相同的協議，並且還可以與

with many other software solutions and technologies on the major cloud providers to help customers manage their data easily and secure. Developing these capabilities required some of our deepest R&D investments and has enabled us to add a feature set that is a generation ahead of any competitor. Our cloud-native capabilities are particularly relevant for customers in tech who are often operating at large scale. Two great examples are Square and Gainsight, both of which have turned to Confluent Cloud for their data streaming platform. Square is a rapidly growing financial services software and payments organization with an ever-increasing volume of payments new customers joining daily. They needed a scalable, reliable and secure data streaming solution. When a merchant receives customer's order through their website or they need to process payroll, a lot needs to happen in real time. We're helping power Square's seller business with a number of use cases that connect the point-of-sale to payment processing to reporting.

與主要雲提供商的許多其他軟件解決方案和技術一起幫助客戶輕鬆安全地管理他們的數據。開發這些功能需要我們進行一些最深入的研發投資，並使我們能夠添加比任何競爭對手都領先一代的功能集。我們的雲原生功能與經常大規模運營的技術客戶特別相關。兩個很好的例子是 Square 和 Gainsight，它們都已將 Confluent Cloud 用於他們的數據流平台。 Square 是一家快速發展的金融服務軟件和支付組織，每天都有越來越多的新客戶加入。他們需要一個可擴展、可靠和安全的數據流解決方案。當商家通過他們的網站收到客戶的訂單或他們需要處理工資單時，很多事情需要實時發生。我們正在幫助 Square 的賣家業務提供大量使用案例，這些用例將銷售點、支付處理和報告聯繫起來。

Gainsight, a SaaS service that helps businesses build long-lasting relationships with customers uses Confluent for real-time data integration to bring together the insights to help maximize the customer experience and trigger actions to ensure their success. Our cloud-native foundation isn't just a factor for larger companies though. Confluent Cloud is seeing success across customers of all sizes. Our cloud-focused commercial sales organization was one of our fastest-growing segments of this quarter, which further speaks to the strong product market fit of Confluent Cloud for digital-native organizations. A great example of this is iFood, a mobile food delivery leader in Brazil, which has recently migrated to our fully managed cloud service running on AWS. They now manage more than 4 million orders per month. Everything from order placement to payment processing to delivery ETAs are handled with Confluent to make sure the right food gets to the right customer as fast as possible.

Gainsight 是一項 SaaS 服務，可幫助企業與客戶建立長期關係，它使用 Confluent 進行實時數據集成，將洞察整合在一起，幫助最大限度地提高客戶體驗並觸發行動以確保他們的成功。不過，我們的雲原生基礎不僅僅是大型公司的一個因素。 Confluent Cloud 在各種規模的客戶中都取得了成功。我們專注於雲的商業銷售組織是我們本季度增長最快的部門之一，這進一步說明了 Confluent Cloud 對數字原生組織的強大產品市場契合度。一個很好的例子是 iFood，它是巴西的移動食品配送領導者，它最近遷移到我們在 AWS 上運行的完全託管的雲服務。他們現在每月管理超過 400 萬份訂單。從下訂單到付款處理再到交貨 ETA 的所有事情都由 Confluent 處理，以確保盡快將正確的食物送到正確的客戶手中。

The second pillar of our product differentiation is completeness. In order for customers to be able to truly harness the power of data in motion, they need a complete offering that makes developing in this new paradigm easy. Kafka is a great foundational layer for building around data streams, but it's just one layer in a modern data and motion stack. Trying to assemble the stack in bitsy pieces and operate all of these is a huge challenge for companies. And that is why we've brought together the essential elements of this stack into a complete unified product. Our product gives customers a robust set of connectors that allow them to plug into the many systems holding data in their organization. We also give them a rich streaming SQL layer, KSQL, which allows them to extend the database skill set their team already has to the new world of real-time streaming. One solution, which brings together many of these capabilities is helping customers migrate and modernize their analytics platform. We recently launched our first priority partner solution focused on cloud data warehouse modernization. A key challenge for organizations is feeding these modern warehouses with data that may come from cloud environments or legacy systems on-premise. As part of that, organizations need a solution that connects data from any system while maintaining the flexibility and security to work across multi-cloud and hybrid environments. Our cloud data warehouse modernization solution enables customers to extract data in real time, transform it into the right structures to support analytics and transport it across cloud environments and regions to any one of a number of next-generation analytics technologies, including Google BigQuery, AWS Redshift, Snowflake or Azure Synapse. This solution truly illustrates the completeness of our platform. It is built on our ecosystem of over 120 connectors and our stream processing capabilities that make it easy to capture, transform and transparently deliver real-time data streams from existing databases, infrastructure layers and SaaS APIs. How we take this solution to market is a great example of our partnership with the major cloud providers. The ability to unlock data for ingestion into these next-generation data warehouses is a key blocker for the growth of these next-generation analytics services and a great point of collaboration and joint go-to-market between Confluent and the cloud providers.

我們產品差異化的第二個支柱是完整性。為了讓客戶能夠真正利用動態數據的力量，他們需要一個完整的產品來簡化在這種新範式中的開發。 Kafka 是圍繞數據流構建的重要基礎層，但它只是現代數據和運動堆棧中的一層。試圖以零碎的方式組裝堆棧並操作所有這些對公司來說是一個巨大的挑戰。這就是為什麼我們將這個堆棧的基本元素整合到一個完整的統一產品中。我們的產品為客戶提供了一組強大的連接器，允許他們插入組織中保存數據的許多系統。我們還為他們提供了豐富的流式 SQL 層 KSQL，這使他們能夠將團隊已經擁有的數據庫技能擴展到實時流式傳輸的新世界。一種將這些功能結合在一起的解決方案正在幫助客戶遷移和現代化他們的分析平台。我們最近推出了專注於雲數據倉庫現代化的首要合作夥伴解決方案。組織面臨的一個關鍵挑戰是為這些現代倉庫提供可能來自云環境或本地遺留系統的數據。作為其中的一部分，組織需要一種解決方案，該解決方案可以連接來自任何系統的數據，同時保持跨多雲和混合環境工作的靈活性和安全性。我們的雲數據倉庫現代化解決方案使客戶能夠實時提取數據，將其轉換為支持分析的正確結構，並將其跨雲環境和區域傳輸到多種下一代分析技術中的任何一種，包括 Google BigQuery、AWS Redshift、Snowflake 或 Azure Synapse。該解決方案真實地說明了我們平台的完整性。它建立在我們由 120 多個連接器組成的生態系統和我們的流處理能力之上，可以輕鬆地從現有數據庫、基礎設施層和 SaaS API 捕獲、轉換和透明地交付實時數據流。我們如何將此解決方案推向市場是我們與主要雲提供商合作的一個很好的例子。解鎖數據以攝取到這些下一代數據倉庫的能力是這些下一代分析服務增長的關鍵障礙，也是 Confluent 和雲提供商之間的協作和聯合上市的重要點。

Our third pillar is being everywhere. And in Q3, we have continued to strengthen the capabilities to support this pillar, most recently with the launch of cluster linking for Confluent Cloud. Organizations today are becoming increasingly geographically distributed with data living in a number of different heterogeneous environments throughout the business. Managing data securely across these environments is a huge challenge. Cluster linking introduces a new way to transparently link together Kafka clusters in different geographies, cloud providers or environments to create a unified fabric for data in motion that seamlessly spans all the parts of the company. An early user of this functionality is Namely. Namely is a fast-growing HR software company, and they have architectural requirements that span geographies in multiple cloud environments to serve their diverse customer base. Cluster linking makes real-time global data sharing, disaster recovery and workload migrations easy for them. A great example that puts all 3 of these pillars together is Instacart. Instacart experienced record growth at the onset of the pandemic, which required them to be able to scale elastically to meet the need. They evaluated a handful of different data and motion options and selected Confluent Cloud due to the combination of cloud-native capabilities that would let them scale elastically to support the growth of their business, the completeness of our offering and our ability to be everywhere, spanning the environments and availability zones their product required. They have complex problems to solve in real time with inventory, customer service and stacking.

我們的第三個支柱無處不在。在第三季度，我們繼續加強支持這一支柱的能力，最近推出了 Confluent Cloud 的集群鏈接。今天的組織在地理上越來越分散，數據存在於整個企業的許多不同的異構環境中。在這些環境中安全地管理數據是一項巨大的挑戰。集群鏈接引入了一種新方法，可以將不同地理位置、雲提供商或環境中的 Kafka 集群透明地鏈接在一起，從而為無縫跨越公司所有部分的動態數據創建統一的結構。此功能的早期用戶是 Namely。即是一家快速發展的人力資源軟件公司，他們的架構要求跨越多個雲環境中的地理區域，以服務於他們多樣化的客戶群。集群鏈接使實時全球數據共享、災難恢復和工作負載遷移對他們來說很容易。將所有這 3 個支柱放在一起的一個很好的例子是 Instacart。 Instacart 在大流行開始時經歷了創紀錄的增長，這要求他們能夠彈性擴展以滿足需求。他們評估了一些不同的數據和運動選項，並選擇了 Confluent Cloud，因為它結合了雲原生功能，可以讓他們彈性擴展以支持他們的業務增長、我們產品的完整性以及我們無處不在的能力，跨越他們的產品所需的環境和可用區。他們有復雜的問題需要實時解決，包括庫存、客戶服務和堆垛。

We believe our platform and the larger movement around data in motion represents a significant step change in making data accessible across an organization. But in order to truly unlock customers' data, it isn't enough just to make it easy, we also have to make it safe. As data powers more of the operations of a company, the quality, correctness, compliance and lineage of data becomes an existential concern. This is a difficulty that is compounded by the explosion of new data types as well as an explosion in the level of regulation and risk around data. When organizations had one monolithic team, they did everything and only a few places where data resided, it was much easier to control data integrity and governance. Customers now need to do this across a diverse organization that is innovating and changing in many areas in parallel, which is a substantial challenge. The key to addressing this challenge is a product offering we announced in Q3 called Stream Governance. This takes the best practices for decentralized federated data governance and applies it automatically with Confluent. Stream Governance is offered as a fully managed cloud solution and delivers a simple self-service experience that enables customers to discover, understand and trust their data flows. This new suite of capabilities establishes trust in the real-time data moving throughout our customers' businesses and empowers teams across any organization to quickly put event streams to work.

我們相信，我們的平台和圍繞動態數據的更大規模運動代表了使數據在整個組織中可訪問的重大變化。但為了真正解鎖客戶的數據，僅僅讓它變得簡單是不夠的，我們還必須讓它變得安全。隨著數據為公司的更多運營提供動力，數據的質量、正確性、合規性和譜系成為一個存在的問題。新數據類型的爆炸式增長以及圍繞數據的監管和風險水平的爆炸式增長加劇了這一困難。當組織擁有一個單一的團隊時，他們可以做所有事情，並且只在少數幾個地方存放數據，控制數據完整性和治理就容易得多。客戶現在需要在一個多元化的組織中做到這一點，該組織在許多領域同時進行創新和變革，這是一個巨大的挑戰。解決這一挑戰的關鍵是我們在第三季度宣布的稱為 Stream Governance 的產品。這採用了去中心化聯合數據治理的最佳實踐，並通過 Confluent 自動應用。 Stream Governance 作為完全託管的雲解決方案提供，並提供簡單的自助服務體驗，使客戶能夠發現、理解和信任他們的數據流。這套新功能建立了對在客戶業務中移動的實時數據的信任，並使任何組織中的團隊能夠快速將事件流投入使用。

There are 3 key capabilities of stream governance. The first is our stream catalog, which helps users find the data they need. This allows individuals across teams to collaborate within a centralized, organized digital library for data in motion, which allows any user at any skill level to put data to use right away. Our customers like Care.com really value this. When it comes to finding, managing and paying for family care, data confidence is critical. Next is stream lineage. This is like Google Maps for your data flows, created in real time of the actual flow of data. It provides an always up-to-date version of how data is flowing in the organization and who is consuming it. With a better understanding of where data originated and where it's going, how it's transformed and when it arrives, developers can move projects forward with assurance, their work won't cause negative or unexpected downstream impact. Take Ritchie Brothers., for example, the global industrial asset auctioneer and one of our great customers, depends on stream lineage for their real-time online bidding system. It makes buying and selling industrial equipment like giant earthmoving trucks easier. And finally, stream quality, which helps us avoid bad data, enforces schemas and ensures compatibility, which enables organizations to scale data integrity without adding operational complexity. We believe that standardizing around well-defined and agreed upon schema structures allows teams to develop resilient data in motion pipelines, prepared for safe compatible evolution over time.

流治理有 3 個關鍵功能。首先是我們的流目錄，它可以幫助用戶找到他們需要的數據。這允許跨團隊的個人在一個集中的、有組織的數字圖書館中進行協作，以獲取動態數據，從而使任何技能水平的任何用戶都可以立即使用數據。我們的客戶（如 Care.com）非常重視這一點。在尋找、管理和支付家庭護理費用時，數據可信度至關重要。接下來是流血統。這就像您的數據流的谷歌地圖，實時創建的實際數據流。它提供了關於數據如何在組織中流動以及誰在使用它的最新版本。通過更好地了解數據的來源和去向、如何轉換以及何時到達，開發人員可以放心地推進項目，他們的工作不會造成負面或意外的下游影響。以 Ritchie Brothers 為例，它是全球工業資產拍賣商，也是我們的重要客戶之一，他們的實時在線競價系統依賴於流血統。它使購買和銷售大型運土卡車等工業設備變得更加容易。最後，流質量可以幫助我們避免不良數據，強制實施模式並確保兼容性，這使組織能夠在不增加操作複雜性的情況下擴展數據完整性。我們相信，圍繞明確定義和商定的模式結構進行標準化可以讓團隊在運動管道中開發彈性數據，為隨著時間的推移安全兼容的演變做好準備。

We're really excited about Stream Governance as it extends the completeness of our platform with capabilities that are essential for a successful data in motion strategy. It represents one of the single most requested features from our customers. With better visibility into and control over data governance, customers can leverage our data and motion platform for mission-critical use cases without having to invest scarce software engineering resources into building tools for monitoring and managing the quality of their data.

我們對 Stream Governance 感到非常興奮，因為它擴展了我們平台的完整性，其功能對於成功的動態數據策略至關重要。它代表了我們客戶最需要的功能之一。通過更好地了解和控制數據治理，客戶可以將我們的數據和運動平台用於關鍵任務用例，而無需將稀缺的軟件工程資源投入到構建工具來監控和管理數據質量。

Stream Governance also serves to illustrate a more general point I'd like to make about our space. Data in motion is a genuinely new category. And though building a new category and paradigm comes with a number of challenges, it has one very big advantage, which is that unlike most of the data stack, all the white space around data in motion is not yet filled in. There are literally hundreds of use cases and product opportunities around data in motion that we see our customers and the larger ecosystem development. And because that ecosystem is just coming into being, Confluent has an enormous opportunity to expand our position and scope into these additional capabilities. This is quite different from a more mature space where all these niches have been filled in by at-scale mature competitors. This is something that for me makes our future product road map very exciting. You can ask yourself the question, if there was a central nervous system, which captured the real-time flow of data in a company, what else would naturally grow in and around that stack. The possibilities are endless. And if we execute quickly and effectively, many of the key elements of this ecosystem can be Confluent products. Stream governance is one small step into this larger problem of helping companies rebuild around real-time data. But it's a good illustration of how we can go beyond the raw infrastructure of moving and processing bytes and deeper into the larger problem of managing and using data in motion.

Stream Governance 還有助於說明我想就我們的空間提出的更一般的觀點。動態數據是一個真正的新類別。儘管建立一個新的類別和範式會帶來許多挑戰，但它有一個非常大的優勢，那就是與大多數數據堆棧不同，動態數據周圍的所有空白區域尚未填充。實際上有數百個我們看到我們的客戶和更大的生態系統發展圍繞動態數據的用例和產品機會。由於這個生態系統剛剛形成，Confluent 有巨大的機會將我們的地位和範圍擴展到這些額外的能力。這與一個更成熟的空間完全不同，在這個空間中，所有這些利基市場都被大規模成熟的競爭對手所佔據。對我來說，這讓我們未來的產品路線圖非常令人興奮。您可以問自己一個問題，如果有一個中樞神經系統，它可以捕獲公司中的實時數據流，那麼在該堆棧中和周圍自然會增長什麼。可能性是無止境。如果我們快速有效地執行，這個生態系統的許多關鍵元素都可以是 Confluent 產品。流治理是幫助公司圍繞實時數據進行重建這一更大問題的一小步。但它很好地說明了我們如何能夠超越移動和處理字節的原始基礎設施，更深入地解決管理和使用動態數據的更大問題。

As we continue to build a market-leading product in a new category, we've made a number of key hires to support our expansion. One of the recent notable hires was Chad Verbowski, our new SVP of Engineering, who joined us from Google. Chad was Head of BigQuery Engineering at Google and had extensive experience leading and scaling Google's cloud data analytics platforms, tools and machine learning services. Together with Ganesh Srinivasan, our Chief Product Officer, the team will continue to innovate and build great products for our customers to set their data in motion.

隨著我們繼續在新類別中打造市場領先的產品，我們聘請了許多關鍵員工來支持我們的擴張。最近值得注意的員工之一是 Chad Verbowski，他是我們新的工程高級副總裁，他從 Google 加入我們。 Chad 是 Google 的 BigQuery 工程主管，在領導和擴展 Google 的雲數據分析平台、工具和機器學習服務方面擁有豐富的經驗。與我們的首席產品官 Ganesh Srinivasan 一起，該團隊將繼續創新並為我們的客戶打造出色的產品，以啟動他們的數據。

And finally, I'm really pleased with the interest and participation we've seen at Kafka Summits around the world. More than 51,000 people registered to attend the 3 summits we held in America, EMEA and APAC this year, an increase of over 40% year-over-year. This momentum is also evident in the talent market. According to LinkedIn Talent Insights, more than 60,000 job hostings globally today are targeting those skilled in Kafka. This has been noted by industry observers as well. ThoughtWorks, a premier technology consultancy and analyst recently noted in their radar publication on tech trends that "Kafka continues towards its status as the de facto standard for asynchronous published subscribed messaging at volume." They further noted that they were seeing increasing standardization around these key platforms like Kafka, Kubernetes and the cloud service providers and far less interest in newer challengers to these emerging standards. This matches our observations. We think it's a very positive sign to the larger movement towards data in motion.

最後，我對我們在世界各地的 Kafka 峰會上看到的興趣和參與感到非常高興。今年我們在美國、EMEA 和 APAC 舉辦的 3 場峰會有超過 51,000 人報名參加，同比增長超過 40%。這種勢頭在人才市場上也很明顯。根據 LinkedIn Talent Insights 的數據，如今全球有超過 60,000 個職位託管針對那些熟練掌握 Kafka 的人。行業觀察家也注意到了這一點。領先的技術諮詢公司和分析師 ThoughtWorks 最近在他們關於技術趨勢的雷達出版物中指出，“Kafka 繼續朝著其作為異步發布的大量訂閱消息傳遞的事實標準的地位邁進。”他們進一步指出，他們看到圍繞這些關鍵平台（如 Kafka、Kubernetes 和雲服務提供商）的標準化程度越來越高，而對這些新興標準的新挑戰者的興趣要小得多。這符合我們的觀察。我們認為，對於動態數據的更大發展，這是一個非常積極的信號。

With that, I'll turn the call over to Steffan to walk through the financials.

有了這個，我會把電話轉給 Steffan 來看看財務狀況。

Thanks, Jay. Good afternoon, everyone. Q3 was another strong quarter as we exceeded the high end of our guidance on all metrics and continue to see accelerating growth in revenue, RPO and a robust net retention rate. These results underscore the power of our industry-leading platform for data in motion and our ability to execute and capture the large and growing market opportunity in front of us.

謝謝，傑。大家下午好。第三季度是另一個強勁的季度，因為我們在所有指標上都超過了我們指導的高端，並繼續看到收入、RPO 和強勁的淨保留率增長加速。這些結果凸顯了我們行業領先的動態數據平台的力量，以及我們執行和捕捉擺在我們面前的巨大且不斷增長的市場機會的能力。

Turning to our customer metrics. We added approximately 190 net new customers this quarter, bringing total customer count to approximately 3,020, up 75% year-over-year. Our growth in large customers continues to be robust. We ended the third quarter with 664 customers with at least $100,000 in ARR, up 48% year-over-year; and 74 customers with at least $1 million in ARR, up 90% year-over-year. The growth in our customers is driven by the complete cloud-native and everywhere aspects of our product differentiation, the network effects of data in motion and our continued improvement in go-to-market with broad-based momentum across enterprise and commercial segments, coupled with our strong partnership with cloud service providers, driven by strong gross retention and expansion across both of our product offerings. Q3 dollar-based net retention rate, or NRR, was greater than 130% for the second quarter in a row, exceeding our near-term target threshold of 120% and was in line with our long-term target of above 130%. Our thesis is that over time, Confluent Cloud with its elasticity and consumption-based model should have a higher NRR than Confluent Platform. The thesis played out nicely this quarter. As we noted in our last earnings call, we expect fluctuation in NRR to continue. With that said, we're pleased with the progress we've made in increasing NRR, and we remain operationally focused on driving it consistently above our near-term threshold and long-term target.

轉向我們的客戶指標。本季度我們新增了大約 190 個淨新客戶，使客戶總數達到約 3,020 個，同比增長 75%。我們在大客戶中的增長繼續強勁。我們在第三季度結束時有 664 名客戶的 ARR 至少為 100,000 美元，同比增長 48%； 74 位客戶的 ARR 至少為 100 萬美元，同比增長 90%。我們客戶的增長是由我們產品差異化的完整雲原生和無處不在的方面、動態數據的網絡效應以及我們在進入市場的持續改進以及企業和商業領域的廣泛勢頭推動的，再加上我們與雲服務提供商建立了牢固的合作夥伴關係，這得益於我們兩種產品的強勁保留和擴展。第三季度基於美元的淨保留率 (NRR) 連續第二個季度超過 130%，超過了我們 120% 的近期目標閾值，符合我們 130% 以上的長期目標。我們的論點是，隨著時間的推移，具有彈性和基於消費的模型的 Confluent Cloud 應該比 Confluent Platform 具有更高的 NRR。這篇論文在本季度表現得很好。正如我們在上次財報電話會議中指出的那樣，我們預計 NRR 將繼續波動。話雖如此，我們對我們在提高 NRR 方面取得的進展感到高興，並且我們在運營上仍然專注於將其持續高於我們的近期閾值和長期目標。

Turning to revenue. Q3 total revenue was $102.6 million, growing 67% year-over-year, an acceleration from 64% in Q2. Subscription revenue was $92.4 million, accelerating to 70% growth year-over-year and it accounted for 90% of total revenue. The 2 components of our subscription revenue continued to exhibit strong growth. Confluent Platform revenue was $65.6 million, up 40% year-over-year and accounted for 64% of total revenue. Confluent Cloud revenue was $26.8 million, accelerating to 245% growth year-over-year and accounted for 26% of total revenue, up from 22% of total revenue last quarter and up from 13% in the year ago quarter. Our strong performance in Confluent Cloud revenue was driven by the themes Jay discussed earlier, including the secular trend of cloud migration, the enhancements and features we've added to our cloud product in recent quarters and great execution by our go-to-market team.

轉向收入。第三季度總收入為 1.026 億美元，同比增長 67%，高於第二季度的 64%。訂閱收入為 9240 萬美元，同比增長 70%，佔總收入的 90%。我們訂閱收入的兩個組成部分繼續呈現強勁增長。 Confluent 平台收入為 6560 萬美元，同比增長 40%，佔總收入的 64%。 Confluent Cloud 收入為 2680 萬美元，同比增長 245%，佔總收入的 26%，高於上一季度的 22%，高於去年同期的 13%。我們在 Confluent Cloud 收入方面的強勁表現是由 Jay 之前討論的主題推動的，包括雲遷移的長期趨勢、我們在最近幾個季度添加到雲產品中的增強功能和功能以及我們的上市團隊的出色執行.

Turning to the geographic mix of revenue. We saw robust demand for our data in motion platform worldwide, with revenues from outside the U.S. outpacing total revenue growth. Revenue from the U.S. grew 59% year-over-year to $66.3 million, representing 65% of total revenue. Revenue from outside the U.S. grew 82% year-over-year to $36.3 million, representing 35% of total revenue, up from 32% in the year ago quarter.

轉向收入的地理組合。我們看到全球對我們的運動數據平台的強勁需求，美國以外的收入超過了總收入的增長。來自美國的收入同比增長 59% 至 6630 萬美元，佔總收入的 65%。美國以外的收入同比增長 82% 至 3630 萬美元，佔總收入的 35%，高於去年同期的 32%。

Turning to remaining performance obligations or RPO. We ended the third quarter with $385 million in RPO, up 75% year-over-year. Current RPO, which is estimated to be 67% of RPO, was approximately $256 million, up 65% year-over-year, an acceleration from 63% last quarter. The main driver of RPO growth was the broad-based strength across both of our subscription offerings. We also saw a modest increase in multiyear deals, which is an endorsement of how our value proposition is resonating with customers. Before turning to gross margins and profitability, I'd like to note that I'll be discussing non-GAAP results unless otherwise noted. Q3 total gross margin was 69.4%, down from 71.6% a year ago. Subscription gross margins were 76.8%, down from 78.9% a year ago. The decline in gross margin was anticipated due to the higher mix and strong growth in Confluent Cloud revenue. As discussed on our last earnings call, Confluent Cloud has a lower gross margin profile than Confluent Platform. We've seen continued improvement in our cloud gross margin, and we remain in our early days of achieving leverage and scale for the infrastructure that supports our cloud offering. As Confluent Cloud continues to scale and account for a larger share of total revenue, we anticipate total gross margin to fluctuate near our midterm target of approximately 70%.

轉向剩餘的履約義務或 RPO。我們在第三季度結束時的 RPO 為 3.85 億美元，同比增長 75%。當前的 RPO 估計為 RPO 的 67%，約為 2.56 億美元，同比增長 65%，比上一季度的 63% 有所加速。 RPO 增長的主要驅動力是我們兩種訂閱產品的廣泛實力。我們還看到多年交易的適度增長，這是對我們的價值主張如何與客戶產生共鳴的認可。在談到毛利率和盈利能力之前，我想指出，除非另有說明，否則我將討論非公認會計原則的結果。第三季度總毛利率為 69.4%，低於一年前的 71.6%。訂閱毛利率為 76.8%，低於一年前的 78.9%。由於更高的組合和 Confluent Cloud 收入的強勁增長，預計毛利率會下降。正如我們在上次財報電話會議上所討論的，Confluent Cloud 的毛利率低於 Confluent Platform。我們已經看到我們的雲計算毛利率持續提高，而且我們仍處於為支持我們的雲產品的基礎設施實現槓桿和規模的早期階段。隨著 Confluent Cloud 繼續擴大規模並佔總收入的更大份額，我們預計總毛利率將在接近 70% 的中期目標附近波動。

Turning to profitability. Operating loss was $42.6 million, representing an operating margin of negative 41.6% compared to negative 32.1% a year ago. Free cash flow margin was negative 20.1% compared to negative 16.7% a year ago. Net loss per share was negative $0.17 using 259.2 million basic and diluted weighted average shares outstanding. As a reminder, our profitability in FY '21 is being impacted by our plan to catch up on hiring given the pause we prudently took in FY '20. Additionally, we're continuing to invest for growth driven by our robust competitive positioning in the large market and the strong unit economics in our model.

轉向盈利能力。營業虧損為 4260 萬美元，營業利潤率為負 41.6%，而一年前為負 32.1%。自由現金流利潤率為負 20.1%，而一年前為負 16.7%。使用 2.592 億股基本和稀釋加權平均流通股，每股淨虧損為負 0.17 美元。提醒一下，我們在 21 財年的盈利能力正受到我們趕上招聘計劃的影響，因為我們在 20 財年謹慎地採取了暫停措施。此外，由於我們在大型市場中強大的競爭地位和我們模型中強大的單位經濟性，我們將繼續投資以實現增長。

Moving on to the balance sheet. We ended the third quarter with $1.3 billion in cash, cash equivalents and marketable securities.

轉到資產負債表。我們在第三季度結束時擁有 13 億美元的現金、現金等價物和有價證券。

Turning now to guidance. For the fourth quarter of 2021, we expect revenue to be in the range of $108 million to $110 million, representing growth of 54% to 56% year-over-year; non-GAAP operating loss in the range of negative $59 million to negative $57 million; and non-GAAP net loss per share in the range of negative $0.23 to negative $0.21 using approximately 264 million weighted average shares outstanding. For fiscal year 2021, we are raising our guidance and now expect revenue to be in the range of $376 million to $378 million, representing growth of 59% to 60% year-over-year; non-GAAP operating loss in the range of negative $170 million to negative $168 million; and non-GAAP net loss per share in the range of negative $0.92 to negative $0.90 using approximately 189 million weighted average shares outstanding.

現在轉向指導。對於 2021 年第四季度，我們預計收入將在 1.08 億美元至 1.1 億美元之間，同比增長 54% 至 56%；非美國通用會計準則運營虧損在負 5900 萬美元至負 5700 萬美元之間；非 GAAP 每股淨虧損在負 0.23 美元至負 0.21 美元之間，使用約 2.64 億股加權平均流通股。對於 2021 財年，我們正在提高我們的指導，現在預計收入將在 3.76 億美元至 3.78 億美元之間，同比增長 59% 至 60%；非美國通用會計準則運營虧損在負 1.7 億美元至負 1.68 億美元之間；非公認會計原則每股淨虧損在負 0.92 美元至負 0.90 美元之間，使用約 1.89 億股加權平均流通股。

I'd also like to provide some modeling points which remain consistent with what we discussed last quarter. We expect FY '21 non-GAAP taxes to be in the range of $2 million to $3 million and FY '21 capital expenditures and amounts capitalized for internal use software costs to be approximately 2% to 3% of total revenue. And finally, while we're still early in the planning process for 2022, I'd like to provide an initial read on our expectations for revenue growth next year. For fiscal year 2022, we expect revenue to grow approximately 36% to approximately $511 million from our FY '21 revenue guidance, which has been raised to $377 million at the midpoint this afternoon. From a seasonality standpoint, we expect Q1 '22 sequential revenue growth to be roughly flat. As our business continues to grow, we expect to see more pronounced seasonality with Q4 '22 being our strongest sequential quarter.

我還想提供一些與我們上季度討論的內容保持一致的建模點。我們預計 21 財年非公認會計原則稅收將在 200 萬美元至 300 萬美元之間，21 財年資本支出和內部使用軟件成本資本化金額約為總收入的 2% 至 3%。最後，雖然我們仍處於 2022 年規劃過程的早期階段，但我想初步了解一下我們對明年收入增長的預期。對於 2022 財年，我們預計收入將比我們 21 財年的收入指導增長約 36% 至約 5.11 億美元，該指導在今天下午的中點已上調至 3.77 億美元。從季節性的角度來看，我們預計 22 年第一季度的連續收入增長將大致持平。隨著我們的業務繼續增長，我們預計會看到更明顯的季節性，22 年第四季度是我們最強勁的連續季度。

In closing, we're pleased with our third quarter results, which reflect our team's ability to execute to capitalize on our significant opportunity, coupled with the growing market adoption and secular tailwinds for data in motion. We remain confident in our ability to build on our momentum in the quarters ahead. With that, Jay and I will take your questions.

最後，我們對第三季度的業績感到滿意，這反映了我們團隊利用我們重大機遇的執行能力，以及不斷增長的市場採用率和動態數據的長期順風。我們仍然對我們在未來幾個季度繼續發展勢頭的能力充滿信心。有了這個，傑伊和我會回答你的問題。

Thank you, Steffan. (Operator Instructions) We will now take our first question from Mark Murphy of JPMorgan followed by Morgan Stanley.

謝謝你，斯特凡。 （操作員說明）我們現在將向摩根大通的馬克墨菲提出第一個問題，然後是摩根士丹利。

Thank you, Shane. Congratulations on just a tremendous growth rate. Jay, I'm wondering if you're possibly seeing any effect from global supply chain issues on the positive side? What I mean by that is perhaps driving retailers to more continual or real-time processes to just try to get a better handle on inventories, right, which have been fluctuating so much, maybe reworking some of their distribution centers to try to handle these challenges out there in the environment.

謝謝你，謝恩。恭喜你取得了巨大的增長率。傑伊，我想知道您是否可能從積極的方面看到全球供應鏈問題的任何影響？我的意思可能是促使零售商採用更連續或實時的流程，以試圖更好地處理庫存，對，庫存波動很大，也許重新改造他們的一些配送中心以應對這些挑戰在外面的環境中。

Yes, it's a great point. We definitely see a set of use cases around retail that have this much more real-time ability to know what's where, when, how and act faster off of it. We don't see any immediate impact from this in the short term, right? And this is generally the case, I think, with these infrastructure layers that it's not the next day after this that you rework all your inventory management, but it often does provoke a longer-term reaction and reflection on how this will be done going forward. We saw that with the pandemic, where maybe Zoom, you would buy it the day after. But Confluent in this larger investment and the infrastructure to support the digital side of the business, that comes a little bit later as you're really thinking through that strategy. It takes a little bit more time. But in some cases, can be a little bit longer lasting because even after people are back at work or even after the ports are unclogged, you still -- this is the architecture you want to have and build against going forward. So it's certainly possible that we'll see some positive impact from it, but nothing immediate.

是的，這是一個很好的觀點。我們肯定會看到一組圍繞零售業的用例，這些用例具有更實時的能力，可以知道什麼在哪裡、何時、如何以及更快地採取行動。我們在短期內看不到這會產生任何直接影響，對吧？我認為，通常情況就是這樣，對於這些基礎設施層，您不會在第二天重新處理所有庫存管理，但它通常會引起長期反應和思考，以了解未來將如何完成這項工作.我們看到在大流行中，可能是 Zoom，你會在第二天購買它。但是融合在這項更大的投資和支持業務數字方面的基礎設施中，當您真正考慮該戰略時，這會稍晚一些。這需要更多時間。但在某些情況下，可能會持續更長時間，因為即使在人們重返工作崗位或端口暢通之後，您仍然 - 這是您想要擁有並構建的架構，以防止向前發展。因此，我們當然有可能從中看到一些積極的影響，但不會立即產生影響。

Okay, great. Just as a quick followup. Steffan, is there any effect in these results where you would see someone who decided they don't want to run -- or they don't want to manage Kafka internally, you compelled them to migrate to Confluent Cloud? And if so, is there any type of conversion ratio or uplift that you're seeing, right, in terms of, I think, presumably their annual spend in that scenario would see some increase.

好，太棒了。就像快速跟進一樣。 Steffan，在這些結果中，你會看到有人決定不想運行——或者他們不想在內部管理 Kafka，你強迫他們遷移到 Confluent Cloud 有什麼影響嗎？如果是這樣，您是否看到了任何類型的轉化率或提升，對，我認為，在這種情況下，他們的年度支出可能會有所增加。

Well, we see a couple of different angles to that question. The first is we see our pay-as-you-go customers that, that contribution is robust. And those are folks who are typically Kafka users who then migrate to our platform for the first time. We make it easy, frictionless to sign up. And once we land those pay-as-you-go customers, we have a customer journey that we've outlined for our customer base and investors to see them progress. And that would naturally lead to more expansion dollars over time, and that's definitely a factor. For the other angle, which is customers who may be running some Kafka and some Confluent at the same time, in this environment, we are seeing folks migrate over to us from the free version to Confluent, mainly due to all of the product feature and functionality enhancements that the team has been working on, candidly, over the last 5 years, in particular, over the last year, all of the things we put into that Confluent Cloud product, in particular, is making it very easy for customers to come over to us and have a very valuable experience.

好吧，我們看到這個問題有幾個不同的角度。首先是我們看到我們的現收現付客戶，這種貢獻是強勁的。這些人通常是 Kafka 用戶，然後他們第一次遷移到我們的平台。我們讓註冊變得輕鬆、順暢。一旦我們找到了這些現收現付的客戶，我們就有了一個客戶旅程，我們已經為我們的客戶群和投資者概述了他們的進步。隨著時間的推移，這自然會導致更多的擴張資金，這絕對是一個因素。另一方面，可能同時運行一些 Kafka 和一些 Confluent 的客戶，在這種環境下，我們看到人們從免費版本遷移到我們，這主要是由於所有產品功能和坦率地說，過去 5 年，特別是去年，團隊一直在努力的功能增強，我們在 Confluent Cloud 產品中投入的所有東西，特別是讓客戶很容易來到交給我們並獲得非常寶貴的經驗。

Thanks, Mark. Our next question comes from Sanjit Singh of Morgan Stanley followed by UBS.

謝謝，馬克。我們的下一個問題來自摩根士丹利的 Sanjit Singh，然後是瑞銀。

All right. Congrats to the team, Jay and Steffan on another really strong quarter. That 245% cloud growth is very impressive. My question was sort of thinking about where we are against what I think are the big opportunity. And correct me if I'm wrong, if I sort of bucket, Jay, the big sort of opportunity you're going for, I think one would be sort of reimagining the data pipeline with streaming data, that being one. The second, a whole new class of streaming apps on sort of the other side of Kafka, this downstream applications. And then third, maybe using Confluent as a data store and analytics engine as well. So the 2 questions is, one, do I have those buckets right, a? And then secondly, like which -- across those 3 buckets, where are you seeing the most traction today? And what do you think could be emerging opportunities as you guys advance the product road map.

好的。祝賀球隊，傑伊和史蒂芬在另一個非常強勁的季度。 245% 的雲增長令人印象深刻。我的問題是在思考我們在哪裡反對我認為的大機會。如果我錯了，請糾正我，如果我有點桶，Jay，你正在尋找的大機會，我認為一個會是用流數據重新構想數據管道，那就是一個。第二，在 Kafka 的另一端，即下游應用程序上的一類全新的流式應用程序。第三，也許也可以使用 Confluent 作為數據存儲和分析引擎。所以有兩個問題，一個，我有這些桶嗎，一個？其次，就像哪個——在這三個桶中，你今天看到的最大牽引力在哪裡？當你們推進產品路線圖時，您認為可能出現的機會是什麼。

I think that's a great way of bucketing it. The first 2, I think, are very mainstream. So we see companies building out pipelines for real-time flow of data between data systems, SaaS layers, whatever needs to connect. We see a whole set of real-time applications being built. Those are very common in our customers. The prevalence might vary from one customer to another, but virtually anybody at scale has both use cases. So if you're -- if you have the stream of sales that are occurring in the business, there are both data systems that, that might need to be delivered to maybe your data warehouse. There's also real-time activity and processing applications that work off of that. And so those are very common. The newer thing is certainly what you described, which is we have in our platform the ability to store data indefinitely. And particularly for more kind of database data sets like the changes of all user profiles, really store that complete data set and be able to do not just real-time processing but actually load the historical data and work off of that. So that is a pattern that we see in customers. I would say that's kind of the -- maybe the second wave people start for the real time and then that retention and longer-term use case becomes very natural because you're just kind of turning the knob on what you're retaining and as use case is demanding.

我認為這是一個很好的方法。我認為前兩個是非常主流的。因此，我們看到公司為數據系統、SaaS 層以及任何需要連接的東西之間的實時數據流構建管道。我們看到正在構建一整套實時應用程序。這些在我們的客戶中很常見。流行度可能因客戶而異，但幾乎任何大規模的人都有這兩種用例。因此，如果您是——如果您有業務中發生的銷售流，那麼可能需要將這兩個數據系統交付到您的數據倉庫。還有實時活動和處理應用程序可以解決這個問題。所以這些很常見。更新的東西肯定是您所描述的，即我們在我們的平台上擁有無限期存儲數據的能力。特別是對於更多類型的數據庫數據集，例如所有用戶配置文件的更改，真正存儲完整的數據集，並且不僅能夠進行實時處理，而且能夠實際加載歷史數據並對其進行處理。這就是我們在客戶身上看到的模式。我會說這有點——也許第二波人們會實時開始，然後保留和長期用例變得非常自然，因為你只是在轉動你所保留的東西的旋鈕，並且作為用例要求很高。

Makes a total sense. And Jay you devoted a good part of your script talking about the new capabilities of platform. You mentioned the governance being a particular one of notable importance. And I wanted to get your particular view as what is this supposed to mean for the existing customer base as either as a conversion from Kafka to the paid Confluent and paid Confluent Cloud or just migrating from platform to cloud rather. How key are these developments in terms of that larger enterprise opportunity?

完全有道理。 Jay，您在腳本中花了很大一部分時間談論平台的新功能。您提到治理是一個特別重要的問題。我想了解您的特定觀點，即這對於現有客戶群意味著什麼，要么是從 Kafka 到付費 Confluent 和付費 Confluent Cloud 的轉換，要么只是從平台遷移到雲。就更大的企業機會而言，這些發展有多重要？

Yes. I mean these are key features, right? So I mentioned cluster linking, virtually every company needs to span geographies, environments, cloud providers, hybrid setups. That's incredibly important. Stream governance, as I mentioned, is probably the most requested feature set in our customer base. And it makes very natural sense, if you're going to unlock data, you have to do it in a way that's safe. And by doing that, it actually encourages the use of this platform growth into central nervous system. It's only possible if the mechanism is really safe and well governed in the modern world. And so yes both of these capabilities are unique to Confluent and are great reasons to adopt our product, right? They are all under that complete and everywhere bucket. These are kind of key reasons with what we have. And so yes, we think it absolutely adds to that.

是的。我的意思是這些是關鍵功能，對吧？所以我提到了集群鏈接，幾乎每家公司都需要跨越地域、環境、雲提供商、混合設置。這非常重要。正如我所提到的，流治理可能是我們客戶群中最需要的功能集。這很自然，如果你要解鎖數據，你必須以安全的方式進行。通過這樣做，它實際上鼓勵使用這個平台增長到中樞神經系統。只有當該機制在現代世界中真正安全且管理良好時，才有可能。所以是的，這兩種功能都是 Confluent 獨有的，是採用我們產品的重要原因，對吧？他們都在那個完整且無處不在的桶下。這些是我們所擁有的關鍵原因。所以是的，我們認為它絕對增加了這一點。

Our next question comes from Karl Keirstead of UBS, followed by Barclays.

我們的下一個問題來自瑞銀的 Karl Keirstead，其次是巴克萊。

Okay. Great. Thanks, Shane. Steffan, maybe I'll direct this one to you and see if you might be able to elaborate on that comment you made when you were talking about seasonality next year about Q1 maybe being sequentially flat and maybe having a little bit more of a 4Q SKU. I think most people on the line are used to hearing that from the likes of Workday and Salesforce, much larger organizations. So I'm curious why you would be seeing it and whether the fact that you're hiring a lot of sales reps today and maybe they're hitting full productivity in the second half might be contributing to that dynamic? Or maybe it's the fact that you're signing a lot more multiyear deals and that sort of pushes you into a fourth quarter cadence. Anyway, I'd love to hear a little bit more color on that.

好的。偉大的。謝謝，謝恩。 Steffan，也許我會把這個發給你，看看你是否能詳細說明你在談論明年的季節性時所做的評論.我認為在線上的大多數人都習慣於從 Workday 和 Salesforce 等更大的組織那裡聽到這一點。所以我很好奇你為什麼會看到它，以及你今天僱傭了很多銷售代表，也許他們在下半年達到全生產力這一事實是否有助於這種動態？或者也許是因為你簽了更多的多年合同，這讓你進入了第四季度的節奏。無論如何，我很想听到更多的色彩。

We're still in the early stages of planning for FY '22. But the way that our hybrid rev rec model works is we have a portion of our business for Confluent platform that gets recognized upfront, as folks know. And then the hybrid component of our model, we have consumption base that's coming off of Confluent Cloud. And depending on the deals structure and the deals that we have in the pipeline, Q4 tends to be a seasonally strong quarter for us, and that's applicable to both Confluent Platform and Confluent Cloud. And when you look at other companies in the enterprise space, even with the similar model of ours, like a MongoDB as an example, they see similar dynamics between a Q4 to Q1 bridge. But we'll see how Q4 ends and we'll give updated guidance post our Q4 earnings call around what obviously Q1 looks like in an update for the year.

我們仍處於 22 財年規劃的早期階段。但是，眾所周知，我們的混合 rev rec 模型的工作方式是，我們有一部分業務用於 Confluent 平台，並且預先得到認可。然後是我們模型的混合組件，我們有來自 Confluent Cloud 的消費基礎。根據交易結構和我們正在進行的交易，第四季度對我們來說往往是一個季節性強勁的季度，這適用於 Confluent Platform 和 Confluent Cloud。當您查看企業領域的其他公司時，即使使用我們的類似模型（例如 MongoDB），他們也會看到 Q4 到 Q1 橋樑之間的類似動態。但我們將看到第四季度如何結束，我們將在第四季度財報電話會議後提供更新的指導，圍繞今年第一季度的更新情況。

The second part of your question around seeing Q4 even being kind of more pronounced next year in terms of sequential growth, we have made a lot of investments in the sales organization. It does take roughly 4 quarters for folks to get fully productive. And as part of the modeling and capacity analysis that we do, we're basically kind of forecasting that to really see the seasonality pattern increase in Q4 of next year. And that's why we called it out on this call.

您的問題的第二部分是關於明年第四季度在連續增長方面甚至會更加明顯，我們已經對銷售組織進行了大量投資。人們確實需要大約 4 個季度的時間才能充分發揮作用。作為我們所做的建模和容量分析的一部分，我們基本上是在預測明年第四季度的季節性模式會真正增加。這就是為什麼我們在這次電話會議上呼籲。

That seems pretty reasonable. And Steffan and Jay, congrats on the amazing numbers.

這似乎很合理。還有 Steffan 和 Jay，祝賀他們取得了驚人的數字。

Thanks so much, Karl.

非常感謝，卡爾。

Our next question comes from Raimo Lenschow of Barclays, followed by Goldman Sachs.

我們的下一個問題來自巴克萊的 Raimo Lenschow，其次是高盛。

Congrats from me as well. I have 2 quick questions. First for Jay.

我也祝賀你。我有 2 個簡單的問題。首先是傑。

If you think about the real time as the kind of the emerging new normal, what's the appetite from your customers around all the kind of more established kind of use like kind of batch processing use cases to modernize that. So at the moment, when we're hearing a lot around like reimagining things around real time, et cetera, but like -- but there's a whole new world out there that is kind of crazy and grow over 20, 30 years, where you think like you could probably do it a lot better in the new way. So maybe speak to that one.

如果您將實時視為一種新興的新常態，那麼您的客戶對所有更成熟的使用（如批處理用例）的需求是什麼，以使其現代化。所以目前，當我們聽到很多關於實時重新想像事物的聲音時，等等，但是就像 - 但是那裡有一個全新的世界，它有點瘋狂，並且會在 20 到 30 年的時間裡成長，在那裡你想想你可能可以用新的方式做得更好。所以也許和那個人談談。

And then for Steffan, like if I look at the numbers, like everything accelerated this quarter, like kind of just the bookings smart as well, which looked like very, very strong. Just one simple question. Was there anything specific in terms of one large deal or anything that drove that? Because this is -- looks like a very, very amazing quarter all around.

然後對於 Steffan，如果我看一下數字，就像本季度的一切都在加速，就像預訂量也很聰明，看起來非常非常強勁。只是一個簡單的問題。就一筆大交易或任何推動該交易的因素而言，是否有任何具體內容？因為這是 - 看起來像一個非常非常令人驚嘆的季度。

Yes, I'll take the first part of that. The question that you're kind of getting at is around. Can you remind me...

是的，我將採取第一部分。你想問的問題是周圍的。能不能提醒我...

Oh, yes. Like if you think about like there's a world of batch. Like what's the appetite to kind of think about kind of modernizing that world.

哦是的。就像如果你想像有一個批處理的世界。就像考慮使這個世界現代化的胃口是什麼。

It's a good question. So I would say there's really no reason you want things to be processed in batch. Like it doesn't -- there's nothing in the world that happens batch. Things in the world happen all the time, continuously and in real time. The migration happens as there's new demands. What's the new need that kind of drives this. And as it gets easier and easier to do it this way. So as this platform has gotten more mature, as the set of people who know how to use it expands, then more things get migrated. And then also the kind of driving force, I touched on this a little bit in the call, is it's kind of just the digitization of business, like as the actual operation of the business moves into software, not just the analysis, not just the report you get in the morning, but the actual carrying out of the business as it plugs into e-commerce things, as it drives operations out in the world, as IoT bridges into other parts of the world, as machine learning kind of closes the loop on some of the decision-making and processes, that's really where you have to do it. Like it's always better, but that's where you must, you must. And so it's a combination of both. And I think over time, yes, I think the majority of that will move.

這是個好問題。所以我想說你真的沒有理由要批量處理事情。就像它沒有一樣——世界上沒有什麼是批量發生的。世界上的事情無時無刻不在發生，不斷地、實時地發生。遷移發生在有新的需求時。這種驅動的新需求是什麼。隨著這種方式變得越來越容易。因此，隨著這個平台變得越來越成熟，隨著知道如何使用它的人的數量不斷增加，更多的東西會被遷移。然後還有那種驅動力，我在電話中談到了一點，是不是只是業務的數字化，就像業務的實際運營轉移到軟件中一樣，不僅僅是分析，不僅僅是報告你早上得到的，但是實際執行的業務是因為它插入電子商務的東西，因為它推動了世界的運營，因為物聯網連接到世界的其他地方，因為機器學習有點關閉循環一些決策和流程，這確實是你必須要做的。就像它總是更好，但那是你必須的，你必須的。所以這是兩者的結合。我認為隨著時間的推移，是的，我認為其中大部分都會發生變化。

And on the second question, we really saw broad-based strength across the board. There was not 1 large deal that came in and kind of swayed the numbers. You look across the details, Raimo, and you look at our commercial and enterprise segments, did incredibly well. They both did incredibly well. You look at our products, Confluent Cloud continues to really shine for us at 245% growth year-over-year. It basically doubled its revenue mix from a year ago. And that's really a testament to the investments we've made in those parts of the organization. The first is in our product. That team has delivered really great products and features and functionality that our customers want. So that's resonating. And then in our go-to-market organization, from sales to marketing, to field operations, we've been finally tuning the machine relative to how to pitch it, how to go in and win big business. And also win small. What we've seen in terms of dollar-based net retention rate being above 130% for a second quarter in a row is really a testament to the ecosystem working. And so we're very pleased with the broad-based strength.

在第二個問題上，我們確實看到了全面的廣泛實力。沒有 1 大筆交易進來，有點影響數字。你看看細節，Raimo，你看看我們的商業和企業部門，做得非常好。他們倆都做得非常好。您看看我們的產品，Confluent Cloud 繼續以 245% 的年增長率繼續為我們閃耀。它的收入組合基本上比一年前翻了一番。這確實證明了我們在組織的這些部分所做的投資。首先是在我們的產品中。該團隊已經交付了我們客戶想要的非常棒的產品、特性和功能。所以這引起了共鳴。然後在我們的市場營銷組織中，從銷售到營銷，再到現場運營，我們最終一直在調整機器，以了解如何推銷它，如何進入並贏得大生意。而且還小贏。我們所看到的基於美元的淨保留率連續第二個季度超過 130%，這確實證明了生態系統的運作。因此，我們對廣泛的實力感到非常滿意。

Our next question comes from Kash Rangan of Goldman Sachs followed by Wells Fargo.

我們的下一個問題來自高盛的 Kash Rangan，然後是富國銀行。

So Jay, in your prepared remarks, there are 2 things that struck me among several things. One was how you're getting tied into the data warehousing ecosystem. Can you talk about what could likely be Confluent's involvement in that ecosystem? Do you see yourself ultimately being a data integrator of choice to that ecosystem, maybe even start with some light analytical workloads.

所以傑伊，在你準備好的評論中，有兩件事讓我印象深刻。一個是您如何與數據倉庫生態系統聯繫在一起。您能談談 Confluent 參與該生態系統的原因嗎？您是否認為自己最終會成為該生態系統的首選數據集成商，甚至可能從一些輕型分析工作負載開始。

And also the other thing that caught my attention you had a chart on the left-hand side, there were data infrastructure use cases and business applications use cases. You had a very observation that you could see in one of these product and some of these could contribute to growth opportunities in the future. Can you just expand on that one? For Steffan, the long-term deferred revenue is certainly much better than expected. Despite the rise of the cloud value proportion, one would think that the opposite would happen, but long-term deferred revenues went up very nicely. What's driving that dynamic?

另一件引起我注意的事情是左邊有一個圖表，有數據基礎設施用例和業務應用程序用例。您有一個非常明顯的觀察結果，您可以在其中一個產品中看到，其中一些可能有助於未來的增長機會。你能擴展那個嗎？對於Steffan來說，長期遞延收入肯定比預期好很多。儘管雲價值佔比上升，但有人認為會發生相反的情況，但長期遞延收入增長非常好。是什麼推動了這種動態？

Yes to the first question, yes, we see it as very complementary. There's a number of really amazing next-generation data warehouse and analytics platforms. And there's actually a number of strong players there. And we partner with all of them. So any streams that go into Confluent, that can power not just real-time applications, not just the kind of operational side of the business. It can flow into some of the analytical systems as well. And that strengthens the value of our platform and the data that flows through. And it's obviously a huge win for these platforms, of course. Build themselves up with useful data and be able to drive analysis on top of that. So it's great. Obviously, as part of that, there is a fair amount of processing of data as it needs to land, and that's the stream processing component. And of course, some of those pipelines need to be purely real-time and live around our platform already today. And yes, there is a role for some real-time analysis as well. And we see customers building out solutions around that, that feed off of Confluent for sure. And there's actually a whole portfolio in the ecosystem of data stores that really specialize in trying to be by that. So yes, I think that's a fantastic area for us.

第一個問題是的，是的，我們認為這是非常互補的。有許多非常了不起的下一代數據倉庫和分析平台。實際上那裡有很多強大的球員。我們與他們所有人合作。因此，任何進入 Confluent 的流，不僅可以為實時應用程序提供支持，還可以為業務的運營方面提供支持。它也可以流入一些分析系統。這增強了我們平台和流經數據的價值。當然，對於這些平台來說，這顯然是一個巨大的勝利。用有用的數據建立自己，並能夠在此基礎上推動分析。所以這很棒。顯然，作為其中的一部分，有相當數量的數據處理需要登陸，這就是流處理組件。當然，其中一些管道需要純粹是實時的，並且今天已經存在於我們的平台周圍。是的，還有一些實時分析的作用。我們看到客戶圍繞這個構建解決方案，這些解決方案肯定是從 Confluent 中獲得的。實際上，數據存儲生態系統中有一個完整的產品組合，它們真正專注於嘗試做到這一點。所以是的，我認為這對我們來說是一個很棒的領域。

And on the long-term revenue question, Kash, it was -- it grew 59% year-over-year. Some of the drivers that go back to large multiyear deals that we've done. And for Confluent Platform deals that are multiyear in nature, that will definitely positively impact long-term deferred revenue. To give you a sense, though, I mean, proportionately speaking, we have $202 million of total deferred revenue. And long-term deferred revenue is about $22.4 million. So it's a relatively small portion, but you see -- when we see more multiyear deals coming in for Confluent Platform, that's going to impact long-term deferred revs.

在長期收入問題上，Kash 是 - 它同比增長了 59%。回到我們已經完成的大型多年交易的一些驅動因素。對於多年性質的 Confluent 平台交易，這肯定會對長期遞延收入產生積極影響。不過，為了給你一個感覺，我的意思是，按比例來說，我們有 2.02 億美元的總遞延收入。長期遞延收入約為2240萬美元。所以這是一個相對較小的部分，但你看 - 當我們看到更多的 Confluent Platform 多年期交易時，這將影響長期延期轉會。

The second part of your question, you asked a little bit about the growth into the ecosystem. Yes, it's a bit of a nuance point, but I think it's important, which is in an established category, everything that's adjacent to the category is filled in by a strong competitor. And in a new category, it's not. It's open. Historically, if you look at an older data platform like Oracle Databases, they could kind of jump relatively easily up into a lot of the business applications that were driving that. We have offerings in HR and financials and all kinds of other things, right? And now, of course, today, if you make a next-gen operational database, maybe you're MongoDB, you don't necessarily get to be working as there's already somebody doing that. And so that's the difference, I think, in a category that's emerging versus a category that's established. What's interesting in this space is there's so much excitement and use cases getting built in customers around data and motion, I think that's just a very target-rich environment for us to grow into. And I think it's something that's exciting about the space. It's something that makes me excited about the opportunities for the company. So I just wanted to share that point.

你問題的第二部分，你問了一點關於生態系統的增長。是的，這有點細微差別，但我認為這很重要，這是在一個既定類別中，與該類別相鄰的所有內容都由強大的競爭對手填補。而在一個新的類別中，它不是。它是開放的。從歷史上看，如果您查看像 Oracle 數據庫這樣的舊數據平台，它們可以相對容易地跳入許多推動它的業務應用程序。我們提供人力資源和財務以及各種其他方面的服務，對嗎？現在，當然，今天，如果你製作下一代可操作數據庫，也許你是 MongoDB，你不一定要開始工作，因為已經有人這樣做了。我認為，這就是新興類別與已建立類別的區別。這個領域的有趣之處在於，圍繞數據和運動在客戶中構建瞭如此多的興奮和用例，我認為這只是我們成長的一個目標豐富的環境。我認為這對這個空間來說是令人興奮的。這讓我對公司的機會感到興奮。所以我只想分享這一點。

Great. Thanks, Kash. Our next question comes from Michael Turrin of Wells Fargo, followed by JMP Securities.

偉大的。謝謝，卡什。我們的下一個問題來自 Wells Fargo 的 Michael Turrin，然後是 JMP Securities。

Congrats on another impressive quarter. On the retention rates, the long-term target of more than 130% is actually higher than the near-term target of 120%. Steffan, you mentioned the mix towards cloud as a key driver there. But can you also expand on the confidence that gives you and the durability of the revenue growth profile, given you're saying the long-term targets are more than 130%. It's not something we often see in software, and I think that's uniquely worth highlighting.

祝賀另一個令人印象深刻的季度。在留存率上，130%以上的長期目標實際上高於120%的近期目標。 Steffan，您提到混合雲是其中的關鍵驅動因素。但是，鑑於您所說的長期目標超過 130%，您能否擴大對您的信心和收入增長的持久性。這不是我們在軟件中經常看到的東西，我認為這是特別值得強調的。

Well, we are very pleased with the progress we've made over the last several quarters and operationally being focused on driving NRR higher. And if you recall, going back to even our IPO, the quarter in which we went public, our NRR for Q2 was 117%. -- sorry, for Q1 was 117%. And so we had work to do, and there were some crosscurrents that were going on back in those times when NRR was fluctuating in the lower levels. There was a transition from -- in our cloud business to usage-based billing. We had some timing of large initial deal sizes, not expanding within a 12-month time period. And there's a real testament to, I think, the go-to-market team and the product team on emphasizing operationally how we can get customers through the life cycle and journey of their deployment of Confluent. And so when we look at the near-term target of greater than 120%, we've been operating for 2 quarters in a row now above 120%. We're actually at our longer-term target of above 130% and the profile of like in the dynamics of NRR as it relates to our 2 main products, Confluent Platform and Confluent Cloud, our thesis is that Confluent Cloud will have a higher NRR profile over time because it's elastic, it's consumption-based and there's very little friction in terms of expansion, whereas with Confluent Platform, we're renegotiating deals, et cetera. And when we look at the profile of our revenue base and the durability of revenue growth for cloud, I don't think there is very much debate that cloud is where it's at going forward. But we also have to be mindful to be the central nervous system, we have to view both on-prem and in the cloud. So there's going to be a dynamic in play where we have both platform and cloud. We think cloud from a durability standpoint is going to be very strong, and that should be a tailwind for NRR, which is why we have all of that kind of sum up into basically saying we have this 120% near-term target. We've been above that for 2 quarters in a row. We want to be above 130% consistently, but it takes time. So before we set new benchmarks in place, we want to see, call it, 3, 4 quarters in a row of great performance in NRR before we look to revisit the targets.

好吧，我們對過去幾個季度取得的進展感到非常滿意，並且在運營上專注於提高 NRR。如果你還記得，甚至回到我們的首次公開募股，也就是我們上市的那個季度，我們第二季度的 NRR 是 117%。 -- 抱歉，第一季度是 117%。所以我們有工作要做，當 NRR 在較低水平波動時，有一些交叉流正在發生。從我們的雲業務過渡到基於使用的計費。我們有一些初始交易規模較大的時機，而不是在 12 個月的時間段內擴大。我認為，這確實證明了上市團隊和產品團隊在運營上強調我們如何讓客戶度過他們部署 Confluent 的生命週期和旅程。因此，當我們看到大於 120% 的近期目標時，我們已經連續兩個季度在 120% 以上運行。實際上，我們的長期目標是 130% 以上，並且 NRR 的動態與我們的兩個主要產品 Confluent Platform 和 Confluent Cloud 相關，我們的論點是 Confluent Cloud 將具有更高的 NRR隨著時間的推移，因為它是彈性的，它是基於消費的，並且在擴展方面幾乎沒有摩擦，而對於 Confluent Platform，我們正在重新談判交易等等。當我們查看我們的收入基礎概況和雲收入增長的持久性時，我認為雲計算未來的發展方向並沒有太多爭論。但我們也必須注意成為中樞神經系統，我們必須同時查看本地和雲端。因此，在我們同時擁有平台和雲的情況下，將會出現動態變化。我們認為，從持久性的角度來看，雲將會非常強大，這應該是 NRR 的順風，這就是為什麼我們將所有這些總結為基本上說我們有這個 120% 的近期目標。我們已經連續兩個季度高於這個水平。我們希望始終保持在 130% 以上，但這需要時間。因此，在我們設定新的基準之前，我們希望在重新審視目標之前看到，稱之為 NRR 連續 3、4 個季度的出色表現。

And just to provide a little bit about the product space that I think kind of helps drive that, the question is like, okay, is that expansion, whether it's 120, 130 now? Can you have high expansion over a very long period of time in a customer base? I think one of the things to understand about this domain is infrastructure is a little bit slow. And we talked about this with the supply chain stuff, right? Like you don't just put it in tomorrow and you're done. In reality, what happens is usage is driven by each application that moves over. The batch process that moved into real time, right? And how quickly do enterprises turn over their set of applications? Well, it's -- I mean it's over decades in reality, right? Like a lot of these things end up lasting much, much longer than anyone would like. And so that does mean that the bad news is we can't just come in and install the central nervous system over the weekend. The good news is we can grow as long as we're successful and we remain the platform of choice in this space, we can grow in these customers over a very long period of time.

只是提供一點關於我認為有助於推動這一點的產品空間，問題是，好吧，是擴張嗎，現在是 120 還是 130？您能否在很長一段時間內在客戶群中實現高度擴張？我認為關於這個領域需要理解的一件事是基礎設施有點慢。我們談到了供應鏈方面的問題，對吧？就像你不只是明天把它放進去，你就完成了。實際上，發生的情況是使用由每個移動的應用程序驅動。轉移到實時的批處理，對吧？企業交出他們的應用程序集的速度有多快？嗯，它 - 我的意思是實際上已經有幾十年了，對吧？就像很多這樣的事情最終會持續很久，比任何人想要的要長得多。所以這確實意味著壞消息是我們不能只是在周末進來安裝中樞神經系統。好消息是，只要我們成功並且我們仍然是該領域的首選平台，我們就可以成長，我們可以在很長一段時間內在這些客戶中成長。

That's great. Congrats to the team.

那太棒了。祝賀球隊。

Our next question comes from Patrick Walravens of JMP Securities, followed by Cowen.

我們的下一個問題來自 JMP Securities 的 Patrick Walravens，其次是 Cowen。

Great. So let me add my congratulations. And then, Jay, as you look out to 2022, what are your top 2 or 3 priorities for Erica to get done and for the sales organization to get done.

偉大的。所以讓我補充一下我的祝賀。然後，傑伊，當你展望 2022 年時，你要完成埃里卡和銷售組織的首要 2 或 3 項優先事項是什麼。

Well, Erica has her own fantastic list of priorities. There's a lot happening in our sales organization. I would say this move to a consumption model for our cloud product, which we undertook a little over a year ago, that's great. That's been a boon to us. It's something our customers love. It allows us to land and expand faster. But I think we're just in the very early days of taking advantage of that and really building out this journey from the lightest, lowest friction possible land in the product to expansion over time, the central nervous system vision that can take many years. And we've built a lot of supporting elements to that. What services do people need as they get to scale. What do we need to do at the very top of the funnel to land in volume and make that really easy to get going with, how do people progress and what are the trigger points. But I think despite all that, we're a young company, and there's a lot more we can do there. And as we do, I think we can land in higher volume. We can expand faster with customers. All of that is going to make us better and better. So that's certainly one of the areas that's really on all our minds is how to do that really well, along with just the continued expansion and tuning of the business as we go.

嗯，埃里卡有她自己的優先事項清單。我們的銷售組織發生了很多事情。我想說的是，我們一年多前就開始了我們雲產品的消費模式，這很棒。這對我們來說是一個福音。這是我們的客戶喜歡的東西。它使我們能夠更快地著陸和擴張。但我認為我們正處於利用這一優勢的早期階段，並真正建立起從產品中最輕、摩擦最低的土地到隨著時間的推移而擴展的旅程，可能需要很多年的中樞神經系統願景。我們為此構建了很多支持元素。人們在擴大規模時需要什麼服務。我們需要在漏斗的最頂端做什麼才能大量落地並使其真正容易上手，人們如何進步以及觸發點是什麼。但我認為，儘管如此，我們還是一家年輕的公司，我們可以做的還有很多。正如我們所做的那樣，我認為我們可以登陸更高的數量。我們可以與客戶一起更快地擴展。所有這些都會讓我們變得越來越好。因此，這當然是我們真正關心的領域之一，即如何真正做好這一點，以及隨著我們不斷擴展和調整業務。

And then can I just ask, I mean your international is growing faster than the U.S., right? Are there different requirements there from a market point of view.

然後我可以問一下，我的意思是你的國際增長速度比美國快，對吧？從市場的角度來看，那裡有不同的要求。

Yes. I mean there's obviously a lot to operate in all these countries. I think we're lucky that we have a product which is inherently very international. And so Kafka usage is kind of everywhere. And so in many ways, it's really just on us and execution to get into all those areas and make sure that we have a sales team that speaks the language and is able to operate on the right accounts in those areas, and we're still in the early days of a lot of that expansion, which is why you see the faster growth there, and we would expect that to continue until monetization looks a little bit closer to the adoption that we see internationally.

是的。我的意思是，在所有這些國家顯然有很多事情要做。我認為我們很幸運，我們擁有一個本質上非常國際化的產品。因此，Kafka 的使用無處不在。因此，在許多方面，進入所有這些領域並確保我們擁有一支會說該語言並能夠在這些領域為正確客戶開展業務的銷售團隊，真的只需要我們和執行力，我們仍然在很多擴張的早期，這就是為什麼你看到那裡增長更快的原因，我們預計這種情況會持續下去，直到貨幣化看起來更接近我們在國際上看到的採用。

Our next question comes from Derrick Wood of Cowen followed by Bank of America.

我們的下一個問題來自 Cowen 的 Derrick Wood，然後是美國銀行。

Great. And congrats as well from me. I'd like to ask about, the new IPO companies, often an IPO can be a catalyst for brand awareness. Have you guys seen any change in inbound interest, lead generation or any discernible impact to win rates? Or is it too early to tell?

偉大的。我也祝賀你。我想問一下，新的IPO公司，往往IPO可以成為品牌知名度的催化劑。你們有沒有看到入境興趣、潛在客戶產生或對贏率有任何明顯影響的變化？還是現在說還為時過早？

Yes, it's hard, of course, to pinpoint anything to something that's a little bit softer like this. But I think for us, it is a bonus. I think because it's a new category that we're trying to bring awareness of, especially in the kind of senior technology leadership ranks of a larger set of companies, I think just raising the profile of the company is really important. And we've tried to do a little bit to aid that. Stephanie joined as our CMO not long before the IPO and has really invested at both as part of the IPO in a sense. And I think that awareness is actually a real boon to us. And we're just starting to get to the scale where that's possible. As an early company, it's hard to really have broad awareness. But as we do, I think that's really important. I think with new categories, there's often an early company that becomes closely associated with that, and I think Confluent is becoming that for data in motion.

是的，當然，很難將任何東西精確定位到像這樣更柔軟的東西。但我認為對我們來說，這是一種獎勵。我認為，因為這是一個我們正在努力提高知名度的新類別，尤其是在更多公司的高級技術領導隊伍中，我認為僅僅提高公司的形象非常重要。我們已經嘗試做一點幫助。 Stephanie 在 IPO 前不久加入了我們的首席營銷官，從某種意義上說，作為 IPO 的一部分，Stephanie 確實在這兩家公司進行了投資。而且我認為這種意識實際上對我們來說是一個真正的福音。我們才剛剛開始達到可能的規模。作為一家早期公司，很難真正擁有廣泛的知名度。但正如我們所做的那樣，我認為這非常重要。我認為對於新的類別，通常會有一家早期的公司與之密切相關，我認為 Confluent 正在成為動態數據的公司。

And a quick follow-up. It's a tough hiring environment out there. How are you guys feeling about attracting and retaining good talent and driving new onboards to productivity as you go into next year?

并快速跟進。這是一個艱難的招聘環境。進入明年，你們對吸引和留住優秀人才以及推動新員工提高生產力有何感想？

Yes. It's as competitive as I have seen it in the tech world for talent. We feel good. We're on track with our plans but it's definitely worked. There is a silver lining on the war for talent on the technical side, which is every company in the world is struggling to hire engineers and the natural complement to hiring a bunch of distributed systems experts to run open sources, cloud services. And we're definitely seeing that shift in some of our customers where they're like, okay, we only have so many brilliant people in the organization and all of their salaries are going up quite significantly and it's really hard to hire more. How can we move those on to the parts of the business that are unique to us and how can we get some of the other staff for a reasonable price. And that's definitely a tailwind that's driving, I think, all of these cloud services, but certainly us.

是的。它與我在科技界看到的人才競爭一樣激烈。我們感覺很好。我們正在按照我們的計劃進行，但它確實奏效了。技術方面的人才爭奪戰有一線希望，那就是世界上每家公司都在努力招聘工程師，這是招聘一群分佈式系統專家來運行開源雲服務的自然補充。我們肯定看到我們的一些客戶發生了這種轉變，他們喜歡，好吧，我們組織中只有這麼多才華橫溢的人，他們的薪水都在大幅上漲，而且很難僱傭更多人。我們如何才能將這些員工轉移到我們獨有的業務部分，我們如何才能以合理的價格獲得其他一些員工。我認為，這絕對是推動所有這些雲服務的順風，當然也包括我們。

Thanks, Derrick. Our next question comes from Brad Sills of Bank of America.

謝謝，德里克。我們的下一個問題來自美國銀行的 Brad Sills。

Congratulations on a real nice quarter here. I wanted to ask about the cloud. I think the focus there was on really driving that SMB or lower TCO kind of initial footprint to kind of get your foot in the door with some mid-market accounts. Is that what's driving the success here would you say? Or is it a balance of that plus large organizations kind of getting started with the cloud, moving data in motion more to the cloud as well?

恭喜這裡有一個非常好的季度。我想問一下云。我認為那裡的重點是真正推動 SMB 或更低 TCO 的初始足跡，以便讓您涉足一些中端市場客戶。你會說這就是推動這裡成功的原因嗎？或者它是一種平衡，加上大型組織開始使用雲，將更多的動態數據移動到雲中？

It really is the full journey. And for this kind of product, it isn't the case that the cloud offering is for the kind of low-end commercial companies only and big companies, what they really want is to do it all in their data centers. There was certainly a period of time where that was true for cloud infrastructure, but that's changed. And so if we look at this last quarter, a good chunk of our largest deals were cloud, surprising chunk of them. And I think that's great. And again, the whole journey matters. Of course, yes, we want people to be able to start quicker with lower friction, but we want that because we want them to graduate to where this is really a strategic platform in the company over time. And so the challenging part for us is to really nail all the steps in that journey and make it happen consistently and regularly. But yes, it is a mix. So we saw great success at the top end of the market. We saw great success at the low end. Our commercial sales organization really killed it this quarter. So yes, it really is a mix of company types.

這真的是完整的旅程。而對於這種產品，雲產品並不是只為低端商業公司和大公司提供的，他們真正想要的是在他們的數據中心裡做這一切。對於雲基礎架構而言，確實有一段時間確實如此，但情況已經改變。因此，如果我們看一下最後一個季度，我們最大的交易中有很大一部分是雲交易，其中的一部分令人驚訝。我認為這很棒。再一次，整個旅程都很重要。當然，是的，我們希望人們能夠以更低的摩擦更快地開始，但我們希望這樣，因為我們希望他們隨著時間的推移畢業到這確實是公司的戰略平台。因此，對我們來說，具有挑戰性的部分是真正確定這一旅程中的所有步驟，並使其始終如一地、定期地發生。但是，是的，它是一個混合體。所以我們在高端市場看到了巨大的成功。我們在低端看到了巨大的成功。我們的商業銷售組織在這個季度真的把它搞砸了。所以是的，它確實是公司類型的混合體。

Great to hear, Jay. One more, if I may, please. I know the global SI channel has been a big focus here. Can you remind us kind of where you are in that development? How critical is that channel to delivering on some of these bigger expand deals? You're obviously already seeing really good net revenue retention and expand activity. But could we anticipate that as a potential catalyst as that channel becomes more online?

很高興聽到，傑伊。如果可以，請再來一份。我知道全球 SI 頻道一直是這裡的一大焦點。你能提醒我們你在那個發展中的位置嗎？該渠道對於實現其中一些更大的擴展交易有多重要？您顯然已經看到了非常好的淨收入保留和擴展活動。但是，隨著該渠道變得更加在線，我們是否可以預期這會成為潛在的催化劑？

Yes. I think we're still early, right? We are putting effort into that. I think over time, this is going to be really important for us to grow into the role that we imagine with our customers. And so we're certainly investing in it and I think it could contribute a lot over time, but it's still in the early days for us now.

是的。我想我們還早，對吧？我們正在為此付出努力。我認為隨著時間的推移，這對於我們成長為我們想像的客戶角色將非常重要。所以我們肯定會投資它，我認為隨著時間的推移它會做出很大的貢獻，但現在對我們來說還處於早期階段。

Thanks, Brad. This concludes the Q&A portion of our call. We will now turn it back over to Jay for closing remarks.

謝謝，布拉德。我們電話的問答部分到此結束。現在，我們將把它交還給 Jay 做閉幕詞。

Yes. Thanks, everyone, for being with us on the call today. Huge thanks to our team, to our partners, to our investors and especially to our customers. And thank you all for being part of this journey, and we'll see you again next quarter.

是的。謝謝大家，今天與我們一起參加電話會議。非常感謝我們的團隊、我們的合作夥伴、我們的投資者，尤其是我們的客戶。感謝大家參與這次旅程，我們下個季度再見。