Pinterest (PINS) 2026年第三季度业绩电话会：欧洲市场压力拖累增长

核心要点

• Pinterest公布2026财年第三季度营收为11.74亿美元，同比增长12%，低于公司的业绩指引区间。管理层将这一不及预期主要归咎于欧洲微量免税（de minimis）豁免政策变化后，跨境卖家缩减预算的幅度大于预期。
• 全球月活跃用户数（MAU）达到创纪录的6.51亿，同比增长10%。这也是Pinterest连续第13个季度创下用户数新高，以及连续第12个季度实现两位数的用户增长。
• 美国和加拿大地区营收增长14%至8.94亿美元。欧洲地区营收约为1.93亿美元，同比基本持平，其中包括跨境卖家带来的预计11个百分点的拖累。
• 调整后EBITDA为3.39亿美元，利润率为29%。自由现金流总计3亿美元，而过去12个月的自由现金流达到近12.6亿美元。
• Pinterest指引第四季度营收为14.25亿美元至14.75亿美元，相当于同比增长8%至12%。调整后EBITDA预计为5.38亿美元至5.88亿美元。
• 管理层表示，强劲的效果广告、国际市场去向（go-to-market）转型的推进以及人工智能驱动的购物产品，旨在支撑公司营收增速随着时间的推移逐步回归至历史上10%中高段（15%-19%）的水平。

关键财务数据

业务与运营表现

Pinterest表示，本季度美国和加拿大的搜索量、创建的图板数量以及出站点击量的增速均快于用户增长速度。在零售、金融服务和医疗健康领域的推动下，该地区营收增长14%。今年迄今为止，美国和加拿大的营收以10%中高段的速率增长，较2025年有所加速。

欧洲是主要压力来源。跨境卖家的预算缩减对该地区营收同比增长造成了约11个百分点的拖累。在全球营收增速从二季度的18%放缓6个百分点至三季度的12%中，管理层将其中约3个百分点归咎于该监管影响，2个百分点归咎于Prime Day和世界杯相关支出的时间安排，1个百分点归咎于汇率对比基数。

Pinterest继续将自身定位为视觉搜索与购物平台。用户每月进行超过800亿次搜索，其中绝大多数被描述为视觉搜索，且超过一半具有商业意图。约70%的搜索旅程始于人工智能支持的主动推荐。

Z世代占用户总数的57%，仍是Pinterest增长最快且互动度最高的群体。管理层表示，三分之二的Z世代用户称他们使用Pinterest进行购物，而他们在美国和加拿大的每用户平均收入（ARPU）与千禧一代相当。

Pinterest Assistant现已在美国全面上线。使用该产品的搜索中约有80%具有商业意图，而Pinterest整体的商业搜索比例为一半以上。公司还推出了Restyle测试版，并正在测试“视觉搜索广告”（Visual Search Ads），这是一种针对搜索结果和Pin图放大查看的下漏斗（转化层）广告形式。

大型广告主对Pinterest Performance+的采用率持续上升。在一项早期测试中，某领先电商平台根据其自有测量系统记录到新客户广告支出回报率（ROAS）提升了77%。

公司还在通过覆盖DSP市场三分之二份额的DSP平台扩大程序化需求，包括Google DV360、Amazon DVA+和Yahoo! DSP，并以Magnite作为其供应方合作伙伴（SSP）。管理层提醒称，这些合作关系的成效释放仍需时日。

管理层业绩指引

管理层预计美国和加拿大地区的营收将继续保持强劲增长，但会被来自跨境卖家的持续国际业务压力所抵消。预计汇率因素将在第四季度构成轻微逆风。

Pinterest扩大了其第四季度营收指引区间，以反映更广泛的潜在结果。公司表示，其国际市场转型的效益已开始显现，但预计在第四季度无法完全抵消跨境业务带来的压力。

风险与关注领域

• 欧洲微量免税（de minimis）豁免政策的变化导致跨境卖家的支出缩减幅度超出预期。管理层预计该支出在假日季期间不会出现复苏。
• 国际销售团队的重组造成了短期干扰。尽管最困难的人员调整已基本完成，但许多新领导者仍处于业务熟悉与磨合阶段。
• 由于广告展示量增速超过营收增速，第三季度广告价格下降了4%。
• 由于用户增长、互动度提升以及GPU产能投资，基础设施支出有所增加。
• 管理层表示，监管政策变化和广告主的适应过程本质上难以预测，尤其是在政策变化于季度内生效的情况下。

分析师问答环节亮点

管理层表示，来自跨境卖家的压力在第四季度已趋于稳定。公司还透露，本季度迄今，其核心欧洲企业业务的增长已重新加速至高于二季度的水平。不过公司指出，卖家的适应过程（包括更多使用本地仓库）可能需要多个季度。

在国际变现方面，管理层强调，尽管购物行为类似，但海外的ARPU仍显著低于美国水平。Pinterest预计其效果广告打法和销售重组将随着时间的推移提升变现能力，但未提供第四季度以外的业绩指引。

Pinterest将Assistant描述为一个集成的对话层，而非独立产品。公司意在利用人工智能来支持产品搜索、比较和规划，同时让购买决策保持在用户的控制之下。

管理层表示，“视觉搜索广告”（Visual Search Ads）首次向广告主开放高意向的视觉搜索互动。该广告形式仍处于测试阶段，公司正在继续与广告主共同测试。

业绩电话会议完整文字记录

---

完整财报电话会议逐字稿

管理层陈述

Operator

Hello, everyone. Thank you for joining us, and welcome to Pinterest's Third Quarter 2026 Earnings Conference Call. [Operator Instructions] I will now hand the conference over to Andrew Somberg, Vice President of Investor Relations and Treasury. Please go ahead.

Andrew Somberg

Good afternoon, and thank you for joining us. Welcome to Pinterest's earnings call for the third quarter ended September 30, 2026. Joining me to discuss our results today is Bill Ready, Pinterest's CEO; James Dibbo, who recently joined as Pinterest's CFO, will briefly introduce himself. The statements we make on this call reflect management's view as of today and will include forward-looking statements. Such statements involve a number of assumptions, risks and uncertainties, and actual results may differ materially. We disclaim any obligation to update these statements. For information about assumptions, risks, uncertainties and other factors that could affect our results, please refer to our earnings press releases and the periodic reports we file with the SEC and are available on our Investor Relations website at investor.pinterest.com.

During this call, we will present both GAAP and non-GAAP financial measures. A reconciliation of non-GAAP to GAAP measures is included in today's earnings press release and presentation, which are distributed and available to the public through our Investor Relations website. Lastly, all growth rates discussed today are on a year-over-year basis, unless otherwise specified. And now I'll turn the call over to Bill.

William Ready

Thanks, Andrew. Good afternoon, and thank you for joining our third quarter 2026 earnings call. Before we begin, I want to officially welcome James Dibbo, who joined Pinterest as Chief Financial Officer earlier this week. James joins us from Amazon, where he most recently served as CFO of their ads and media businesses. Before that, he was CFO of Amazon's worldwide consumer business. James brings deep experience from one of the largest and fastest-growing ads platforms in the industry and one of the best shopping experiences in the world. Both are highly relevant to what we are building, and we are thrilled to have him here. Later in the call, James will share a few words and introduce himself. Welcome, James.

Before I turn to the specifics of the quarter, I want to take a moment to step back and frame where we are in Pinterest's journey. Pinterest is where users come and search is something they love. We have become an increasingly important destination for visual search and shopping, powered by differentiated AI capabilities and built around commercial intent. Unlike platforms built around passive content consumption, Pinterest helps people discover what they love and act on it. At a time when search and consumer expectations are changing rapidly, we continue to win consumer shopping demand at record levels for our platform, including across geographies and generations and especially with Gen Z.

In our core UCAN market, engagement is deepening as searches, boards created and outbound clicks all grew faster than users in Q3. Users and engagement are the clearest indicators of the strength of a platform and that gives us conviction in our long-term opportunity. Our focus over the last year has been to translate that strong engagement into more durable monetization. We are doing that in 2 ways: continuing to improve our performance ads platform and strengthening our go-to-market motion, including new leadership across the organization.

Now let me turn to the quarter. In Q3, we delivered our 13th straight quarter of record users and our 12th straight quarter of double-digit user growth. On revenue, we flagged 2 headwinds coming into the quarter affecting our international business. regulatory changes to the de minimis exemption in Europe affecting cross-border sellers and the restructuring of our international go-to-market organization. The regulatory impact was greater than we anticipated as it is inherently hard to forecast when changes take effect during the quarter. The pullback from cross-border sellers accounted for approximately 3 points of the sequential change in our global revenue growth rate from Q2 to Q3.

The impact of our international go-to-market restructuring played out about as we expected and was a smaller contributor. Overall, Q3 revenue came in below our range. We are not happy with that result, and we have clear plans in place to accelerate our revenue diversification efforts. Despite these headwinds, we delivered adjusted EBITDA within the range we expected. This quarter reinforces the importance of the work already underway to build a broader, more durable revenue base. The reinvigoration of our UCAN business provides a clear proof point for the approach we are taking.

Last year, changes in the regulatory environment in the United States affected large retailers and cross-border sales into the U.S. Despite those headwinds, continued improvements to our performance ads platform and go-to-market execution helped us drive a reacceleration in UCAN revenue growth. Year-to-date, UCAN revenue has grown at a mid-teens rate, a clear acceleration from 2025. We are now applying that same playbook internationally, and we are beginning to see early signs that these changes are taking hold internationally. That gives us confidence that the actions we are taking are right for the long-term health of the business. As we said at the beginning of the year, progress would not always be linear, and we might face some short-term headwinds as we make changes that are right for the long term.

Moving forward, our priorities remain clear. First, continue building a differentiated visual search, discovery and shopping experience; second, keep AI at the core of everything we do. Third, accelerate monetization through improved go-to-market and measurement. With that context, I'll start with engagement, which underscores the distinct role Pinterest plays for users. Pinterest has become a destination for search, particularly visual search. Users conduct more than 80 billion searches on Pinterest each month. The vast majority are visual and more than half are commercial. This is a different kind of search experience from traditional text-based search. People come to Pinterest with a point of view, but often without the exact words, brand or product in mind. At the core of this search experience is our approach to agentic discovery. We use AI to proactively bring users relevant ideas, products and next steps based on their taste and intent.

For many people, shopping is not a task to be automated. It is a joyful activity in the past time. Our AI-assisted experiences are designed to enhance a journey the user wants to engage in rather than take that journey away from them. We use AI trained on our unique data to understand people's preferences and anticipate what might inspire them next. But taste is personal and the ultimate choice belongs to the individual. We believe the best AI experiences will expand human choice, not replace it. This is why Pinterest is built for all the wonderful parts of shopping that people do not want to give up in the same way that people don't seek to automate away their entertainment.

Approximately 70% of search journeys on Pinterest begin with a proactive recommendation made by our AI-enabled systems before a user types a word. From there, users explore possibilities, compare options visually, and turn inspiration into action. That experience continues to drive user momentum. Global MAUs grew 10% in Q3. It is resonating particularly well with Gen Z, our fastest-growing cohort. Gen Z now represents 57% of our users and is our most engaged cohort across monthly retention, saves and sessions. They also come with intent. 2/3 of Gen Z users say they come to Pinterest to shop. That intent carries through to monetization as ARPU for Gen Z users is comparable to millennials and UCAN which bodes well for ARPU growth as Gen Z begins to enter their peak spending years. What makes our engagement particularly valuable is the intent behind it. As users search, save, and build boards they reveal evolving tastes and interests that help us anticipate what they want next.

This creates a valuable feedback loop. The more people engage, the more relevant and personalized their search experience becomes, giving them more reasons to return. These distinctive signals built over years create an increasingly differentiated experience for users and a powerful commercial opportunity for advertisers.

Pinterest Intelligence is the foundation that connects our proprietary first-party signals, visual understanding and purpose-built AI models. With it, we can deliver more personalized discovery for users and better performance for advertisers. We pair our proprietary data with compact fit-for-purpose models and open way models post trained in our own environment.

As I said last quarter, open-weight models can outperform comparably sized closed third-party models for our use cases because we can post train them on Pinterest unique data. We are also achieving a cost per transaction at less than 8% of comparable closed proprietary models. Importantly, we are scaling these AI capabilities efficiently and applying them to highly monetizable use cases such as Pinterest Assistant. Pinterest Assistant is now in general availability in the U.S., and we are encouraged by the early signals. While more than half of all searches on Pinterest are commercial, approximately 80% of searches using Pinterest Assistant are commercial in nature. Within the assistant, we are building more agentic experiences designed to assist users in their discovery journey rather than automating them out of it. including Ask mode, which is beginning to scale in the U.S.

We are also deepening capabilities across our core use cases with products such as beauty guides and Restyle. Restyle, which launched in beta in Q3, helps users reimagine a room from a photo by adding a new piece, replacing an existing 1 or exploring a different look. It combines Pinterest Intelligence's understanding of style with shoppable products from our catalog, helping users see how an idea could work in their own space. Much of the Restyle experience is currently powered by Pinterest Canvas, our proprietary AI image generation model. The same intelligence that improves the user experience is also helping drive compounding gains for advertisers. We continue to improve how we retrieve and rank ads using stronger relevant signals and predicted purchase intent. We are also increasing the pace of these enhancements.

So gains reach advertisers sooner. Visual Search Ads is a new lower funnel format currently in beta, which helps performance advertisers reach users in some of Pinterest's highest intent moments, search results and pen close-ups. The format is designed to drive measurable outbound action, giving advertisers a new way to turn visual discovery into conversion. Advertisers can use their existing catalogs and keyword targeting making adoptions straight forward. We are extending the same focus on performance to app promotion, our new objective for app installs, now in beta with qualified advertisers globally.

Pinterest has not historically been a meaningful participant in mobile app advertising, and this opens an important new budget pool. Alongside these launches, we are making Pinterest easier for advertisers to use, optimize, and scale. We have consistently said that Pinterest Performance+, our AI-powered campaign solution would be a multiyear product and adoption cycle. Advertisers can now add keywords and interests that matter most to them and track results for each, with priority products, they can also emphasize the product groups they value most.

The result is automation with greater control and transparency. Larger advertisers have asked for both, and they are driving continued Pinterest Performance+ adoption. We are also advancing our measurement and bidding capabilities to help advertisers see the full value Pinterest drives. In Q3, more of our most sophisticated advertisers directly integrated their in-house measurement systems with our AI bidding tools. These integrations allow us to optimize toward the outcomes each advertiser values most. For example, in early testing with a leading e-commerce platform, that advertiser saw 77% higher new customer return on ad spend reflected in their proprietary measurement source of truth.

We are pairing these platform improvements with a more disciplined go-to-market motion focused on selling performance advertising. Stepping back, we are only about 3 years into our performance advertising journey, which started in UCAN, we built our lower funnel playbook around shopping in a deliberate order, starting with making Pentasa place where users shop and take action. We then built a performance ad platform so advertisers could tap into that intent and aligned our sales and go-to-market teams behind it. We are now running the same playbook internationally. On go-to-market, we are building across 3 areas: people, organizational structure and operational rigor.

Starting with people. We have leveled up our sales leadership to align with our shift toward performance selling , which takes different skills than selling upper funnel brand ads. In our international regions, we started this process more recently, and more changes were required across multiple markets. Overall, while the toughest parts of our personnel changes are behind us, many of our new leaders are still reading up; second, organizational structure. We are refining account coverage around our largest opportunities and bringing our [indiscernible] sales teams closer together. That gives advertisers more consistent access to the performance, measurement and product expertise they need to scale with Pinterest. Third, operational rigor.

We are creating clear seller accountability through more consistent training, incentives and sales motions. And can this has helped us package key commercial moments more effectively and connect Pinterest product performance more directly to advertiser outcomes. Internationally, we are bringing the same discipline to each market. These actions are designed to improve client coverage, seller productivity and our ability to convert strong advertiser performance into more durable revenue growth.

Ultimately, the goal is a more resilient revenue base that reflects both the strength of our users and engagement and the performance we deliver for advertisers. In summary, I am confident in the work because our foundation is strong. Our users are growing and engaged. Our platform is delivering performance for advertisers, and we are strengthening our go-to-market engine. The work we are doing now positions Pinterest to reaccelerate growth over time and return to the mid- to high teens revenue growth we have delivered historically. Before I walk through our results and outlook, I'd like to turn it over to James to share a few words about why he joined Pinterest.

James Dibbo

Thank you, Bill. I'm delighted to have joined Pinterest on Monday. And since it's only been a few days, I'll keep my remarks brief for this call. I came to Pinterest because I have huge admiration for what Bill and the team have built, but primarily because of the enormous upside opportunity still ahead of us.

I first discovered Pinterest in 2014 and have used it for house moves and remodels, planning my dream cheap Wrangler build, and just last weekend, I had some fun searching for outfit ideas for my first day as Pinterest CFO. I'm looking forward to getting to know the analyst and investor community and working with Bill, the Board and the whole team to realize Pinterest's full potential as a world-leading visual search, shopping and advertising platform.

Pinterest is a place where a growing number of users from around the world come looking for inspiration and a safe space online to discover the ideas, products and adventures that bring them joy off-line. At the moment, when a growing number of people are actively trying to limit their use of social media and as a father of 2 daughters, I believe Pinterest focus on positivity and inspiration is especially timely and important as well as a significant business opportunity. The AI-powered innovations the team continues to develop helping more and more brands to reach those high-intent customers at exactly the right moment with relevant, useful, actionable ads that are complementary to users

Pinterest experience and in success will create accelerating growth flywheel of higher advertiser value, deeper user engagement and increasing value for shareholders. I'm looking forward to the journey and to our discussions on future earnings calls. Thank you, and I'll pass it back to you, Bill.

William Ready

Thank you, James. Now I'll discuss our third quarter 2026 financial results and provide an update on our fourth quarter 2026 outlook. All financial metrics except for revenue, will be discussed in non-GAAP terms unless otherwise specified. And all comparisons will be discussed on a year-over-year basis unless otherwise noted. .

We ended the quarter with 651 million global monthly active users, or MAUs, growing 10% and reaching another record high. We continue to demonstrate user growth across all of our geographic regions. In Q3, our U.S. and Canada region had 107 million MAUs growing 4%. Our Europe region had 161 million MAUs, growing 7%. And in the rest of world markets, we had 394 million MAUs growing 13%.

Shifting to revenue. In Q3, our global revenue was $1.174 billion, up 12% on a reported and on a constant currency basis. We saw strength from our consideration objectives, particularly our performance-driven CPC advertisers and, to a lesser extent, our awareness objective. Across verticals, we continue to see strength in retail as well as smaller but faster-growing emerging verticals on our platform, including financial services, health and travel.

Turning to our geographical breakouts for Q3. Revenue in the U.S. and Canada was $894 million, growing 14%. Growth came from retail, financial services, and health. In Europe, revenue was $193 million, roughly flat on a reported basis or up 1% on a constant currency basis. The pullback from cross-border sellers represented roughly an 11-point drag on Europe's year-over-year growth. Revenue from Rest of World was $87 million, growing 24% on a reported basis or 21% on a constant currency basis.

I want to be clear about the primary drivers of the 6-point sequential deceleration in global revenue growth from 18% in Q2 to 12% in Q3, approximately 3 points reflected the pullback from cross-border sellers in response to the EU regulatory change on the de minimis exemption. Approximately 2 points reflected the timing of Prime Day and World Cup related spend that we called out last quarter. The remaining point reflected foreign exchange, which added 1 point to Q2 growth and was neutral in Q3.

In Q3, overall ad impressions grew 17%, ad pricing in Q3 declined 4% as ad impressions growth outpaced revenue growth. Moving to expenses. In Q3, cost of revenue was $250 million, up 21% year-over-year and up 2% versus Q2 driven by increased infrastructure spend related to our user and engagement growth and investments in GPU capacity as well as the impact of tvScientific. Our non-GAAP operating expense was $591 million, up 9%. The increase was primarily driven by sales and marketing reflecting higher marketing expense and sales headcount as well as R&D to support our AI and product initiatives.

We ended Q3 with 5,192 employees, down slightly year-over-year in Q3 including employees who joined Pinterest through tvScientific acquisition. Excluding that acquisition, headcount was down 4% year-over-year in Q3. We will continue to invest in specialized talent where we see a clear impelling ROI but an ongoing focus for us is finding the right mix of talent, infrastructure and AI-enabled productivity tooling to produce durable growth and expand margins over time.

In Q3, we delivered $339 million in adjusted EBITDA with an adjusted EBITDA margin of 29% down 30 basis points versus Q3 last year. We also delivered Q3 free cash flow of $300 million. On a trailing 12-month basis, we've generated nearly $1.26 billion of free cash flow representing 90% free cash flow conversion. We ended the quarter with cash, cash equivalents, and marketable securities of $1.5 billion.

Now I'll discuss our guidance for the fourth quarter. We expect Q4 revenue to be in the range of $1.425 billion to $1.475 billion, representing 8% to 12% year-over-year growth. Based on current spot rates, we expect foreign exchange to be a modest headwind in Q4. Our outlook assumes that the puts and takes that impacted our Q3 revenue growth will continue in Q4.

First, we expect to see continued strong revenue growth in UCAN, our largest region. However, we expect the pullback from cross-border sellers to continue to weigh on international revenue growth. We are beginning to see early signs that our international go-to-market changes are taking hold but their associated revenue benefits will take time to build and are not expected to fully offset this pressure in Q4.

To be prudent, we have widened our Q4 revenue guidance range to reflect the broader range of potential outcomes we see for the quarter. We are confident in the underlying drivers of the business and have made significant progress to simplify how we operate and invest intentionally in the areas that matter most that give us the flexibility to navigate a range of scenarios. For Q4, we expect adjusted EBITDA to be in the range of $538 million to $588 million. We anticipate Q4 2026 non-GAAP cost of revenue to grow sequentially from Q3 2026 at a low teens rate, partially driven by our investment in GPU capacity. As a reminder, Q3 2026 non-GAAP cost of revenue benefited from the accelerated recognition of certain contractual benefits associated with our recently executed multiyear infrastructure agreement.

Within Q4 non-GAAP operating expense, the primary areas of year-over-year investment will be sales and marketing and R&D to support our AI and product initiatives. We continue to expect full year 2026 adjusted EBITDA margins of approximately 30%. While our near-term revenue growth is not where we want it to be, I'm proud of the experience we continue to deliver for users. I have full confidence in our team, strategy and the actions we are taking to durably reaccelerate growth over time. With that, Andrew and I will now move to Q&A.

Operator

[Operator Instructions] Your first question comes from the line of Jason Helfstein with Oppenheimer.

分析师问答

Jason Helfstein

So Bill, I think you were pretty clear about the headwinds with 3Q. I'm intrigued by your comment about seeing signs of internal signs that international can accelerate. Obviously, you've got the fourth quarter guide, but I guess, when -- help us maybe like when do you think that the initiatives could start to turn international revenues is just we're kind of thinking out? And I guess if there's any other color you want to give us on kind of had a kind of process what we saw in third -- 3Q and how that follows through the model in the fourth quarter of next year.

William Ready

Thanks for the question. Yes. So I'll give a little bit more color on Q3 and then give you some more color on our Q4 outlook as well. So in Q3, the sources of puts and takes in the business were consistent with what we described on our last earnings call, namely strong UCAN growth continuing and 2 headwinds impacting international, the 1 being regulatory changes in Europe and the other being our own go-to-market restructuring.

However, the magnitude of the European regulatory change was greater than we forecast and contributed 3 points of deceleration in the quarter. Let me give you a little more color on all those puts and takes. We continue to drive strong user engagement growth, globally and in UCAN, users grew 10% to $661 million, and engagement in UCAN deepening with searches, boards created and outbound clicks, all growing faster than users. UCAN revenue was up 15% year-to-date, a meaningful acceleration from 2025. And globally, revenue growth decelerated sequentially from 18% in Q2 to 12% in Q3, 3 points of that came from factors we mentioned on our last earnings call, those being the World Cup, timing of Prime Day shifting from Q3 into Q2 and 1 point from foreign exchange. The other 3 points were due to the removal of the European de minimis exemption resulted in cross-border sellers pulling back more than we anticipated.

Regulatory changes like those are inherently hard to forecast, particularly as they took effect during the quarter. We also saw some near-term disruption from the transition of our international go-to-market organization, although that impact played out largely as we expected. And I'd add to that, that in recent months, we've made a number of leadership changes as we transition toward a more performance-oriented selling model. We're actually quite encouraged by the [indiscernible] we've added, and we're largely through the most difficult parts of that transition, and we're starting to see early signs of those changes taking hold. For example, the growth rate in our core European enterprise business, which excludes cross-border sellers that growth rate halved in Q3 relative to Q2. But quarter-to-date in Q4, we have since reaccelerated that growth above its Q2 level. So while we and our advertisers were dealing with major regulatory changes and the implementation of those played out through the quarter [indiscernible] a miss in our revenue guidance, and we saw the deviation from the forecast was greater in the later weeks of the quarter. That's now stabilized in Q4.

And again, we're seeing the green shoots with our international sales restructuring. For Q4, based on our quarter-to-date results, Q4 is tracking broadly in line with the growth that we saw in Q3. We continue to see strong global engagement along with continued strong revenue growth in UCAN offset by pressure on international growth from cross-border sellers, although that pressure has stabilized. That said, as mentioned earlier, we have provided a wider guidance range to reflect a broader range of potential outcomes as well as a desire to provide a prudent baseline that we feel highly confident we can meet or exceed.

Operator

Your next question comes from the line of John Blackledge with TD Cowen.

John Blackledge

Great. And I think you just kind of addressed it, but I don't know if you could expand or discuss the impact of the cross-border sellers from the regulatory change, the de minimis change and kind of following up to Jason's question, thinking about a return to growth in Europe, kind of as we enter next year? Any color there would be helpful.

William Ready

Thanks for the question. Just as I mentioned, regulatory changes like these could be inherently difficult to forecast in the short term as advertisers adapt to a new reality. I mentioned that there was a 3-point sequential change in our overall revenue grade at the global level. And we do anticipate that we'll still weigh on international growth in Q4, though it stabilized. But I'd say it's also important to recall what happened with these advertisers last year when tariffs went into effect in the U.S., many rerouted their spend into Europe and Lat Am.

Now there are barriers in both the U.S. and Europe, and there are fewer large markets to reroute to. So in the short term, we're seeing these advertisers pull back globally to protect margins, but we do expect these sellers to adapt over time. For example, some of those retailers have talked about their plans to leverage local warehouses to navigate those new regulations, but those shifts take some time. So hard for us to predict that. But in the meantime, our new sales leaders are ramping. They're helping to bring more advertisers across more categories, and that will diversify our revenue over time.

And so while that doesn't fully offset the headwinds that we see, we are quite encouraged by that. And while we're not guiding next year, and we don't expect cross-border spend to come back for the holiday, as we look into H1 of next year, we think those -- that allows more time for those advertisers to make those kinds of shifts, while we also continue to build out our revenue diversification as we look into H1.

Operator

Your next question comes from the line of Colin Sebastian with Baird.

Colin Sebastian

Great. I'll add my welcome to James. Bill, thanks for the comments on the agentic initiatives. How are you thinking about the role of Pinterest Assistant alongside core discovery within the app. And I'd be curious to understand a little bit better what you're learning about the role of agents or assistance in helping users move from inspiration towards a purchase.

William Ready

Yes, thanks for the question. Well, first thing I'd say, having just put up our 13th straight quarter of record high users. As there have been multiple ways of AI-driven experiences and agentic experiences, we have continued to show strong user and engagement growth. And at the core of that is that we are delivering an AI-powered [indiscernible] experience that we think really meets our users' needs in the ways that they want to see it. .

And I mentioned a little bit in my prepared remarks. From the beginning, we've been focused on the aspects of shopping that the user wants to engage in. You all heard me say many times before, the first 25 years of e-commerce sort of solved [indiscernible] but killed shopping in some ways and that being the shopping part being the journey that the user wants to engage in the joy of the discovery sort of figuring out what you want your look to be refining your own taste. And so we've designed Pinterest Assistant, one to be woven in throughout the experience. It acts as a copilot, taking users from inspiration to comparison and research while the user stays in control. So Pinterest Assistant, which is our AI conversational layer, we're not approaching it as a separate experience. It's well it throughout the user experience where every pen is already served by AI.

But it does bring large language model capabilities into our visual personalized experience. So users can research, compare and decide without leaving Pinterest. That experience is now generally available in the U.S., and we're really encouraged by what we're seeing as users engage with it. One of the most encouraging signs that we see is that roughly 80% of assistant searches are commercial, and that compares to half of searches across the rest of Pinterest. So that demonstrates how we are aligning AI to highly monetizable use cases. We feel really good about that alignment of our use of AI towards monetization and our ability to monetize that. So we're doing $80 billion-plus monthly searches.

And around this issue of agentic, 70% of the searches on Pinterest today start with a proactive recommendation. So that gives you a sense of how much we have woven this into the core experience of making proactive recommendations before the user had to type or ask anything, and that sort of gets to the process of agentic, but in a way that is built for mass adoption in terms of how users everyday users think of engaging with these things in a way that meets them where they are. So the assistant makes recommendations easier to digest, meet users where they are, and as we look forward, we do see an opportunity to expand further into more agentic experiences like our ask mode that I mentioned. That's now available to millions of users in the U.S., and that helps with multistep journeys such as planning a trip.

And so things like planning a trip, you can ask about planning a trip, and it will come back with recommendations aligned with your style and taste around where to stay, what activities you might engage in, what restaurants you might dine at, what outfits you might plan for? These are all things that are taking us deeper into that agent experience, but again, keeping the user in control, where they're getting great recommendations, but the user is still siding. So we're just getting started around that, but we continue to focus on scaling actionability and monetization, and we continue to test and innovate and we're adding assisted capabilities. But again, we feel really good about how we're aligning those AI-driven capabilities with good commercial experiences, which bodes well for monetization as we look forward.

And we feel really good about the way that we're designing these for mass user adoption, meeting the users where they are, which is helping them with the decisions that they want to make but still allowing the user to ultimately decide for themselves, which we think is really important, that people don't want to lose control of decisions. They may want help with the decisions, but they don't want to lose control of the decision. So we think that's really at the core of how we're driving really great growth in users and engagement.

Operator

Your next question comes from the line of Justin Patterson with KeyBanc.

Justin Patterson

I was hoping you could talk a little bit more about the decision to expand programmatic partnerships. How do you see this for 1 thing advertiser reach and [indiscernible]? And so some of these before [indiscernible] Could you just talk about how your initial learnings were there and how that's been playing out thus far?

William Ready

Yes. Thanks for the question. So yes, we're expanding the partners we work with programmatically, and it's several ways we've been broadening demand on Pinterest. And it helps us reach advertisers where they are. So up to this point, our goal has been to complement our first-party business with more sources of demand to round out gaps in our auction, drawing on the commercial intent that we see on the platform. However, we also see a meaningful opportunity to make it easier for our customers to buy where they want, as we increasingly see advertisers buying programmatically, we want to meet them more where they are.

So as we announced this week, we're now working with DSPs that represent 2/3 of the DSP market, including Google DV360 and Amazon DVA+, Yahoo! DSP with Magnite as our supply-side partner. And as we look forward, we think there's more of those types of opportunities to come. As we've shared over the last few quarters, there's also a meaningful opportunity to leverage our audience beyond our own platform. We started this with our acquisition of tvScientific earlier this year, which extended our audience into Connected TV. Now as these platforms advance, there are more privacy safe ways to leverage our unique audience data and partnership with these DSPs.

So that opens up the door to bring more demand programmatically while still highlighting value of our unique signal and their unique highly commercial audience that we have on Pinterest. And I'll close with a reminder as I've shared on these partnerships before, this demand takes time to tune and ramp, as you've seen with prior program partnerships. So we think this is an important step to make it easier for more partners to bring more demand on the platform. But again, expect these things to take time to ramp.

Operator

Your next question comes from the line of Lloyd Walmsley with Mizuho.

Lloyd Walmsley

Just wanted to ask about the new ad unit you guys announced in Visual Search in the last sort of month or so. it looked like it was -- the lease talked about some pretty bullish stats. I think it said the vast majority of searches are visual, more than half of them are commercial. But help us frame up how meaningful this could be maybe equivalent to your other big services, how incremental this potential ad unit could be?

William Ready

Yes. Thanks for the question, Lloyd. So yes, visual search is a whole new way that people are searching online, and we're the leader in visual search. It those 80 billion-plus monthly searches, the vast majority of them being visual. What we're doing with visual search ads is giving advertisers a direct way to engage users in that experience for the first time. And we think this is a big and meaningful opportunity.

So those visual surgeons that are occurring on the platform, high commercial intent, more than half commercial and now really opening up more of a way for the advertiser to engage at the top of those interactions. So we focus on building those experiences organically first and hadn't really made ads featured in that visual experience. Now we're opening that up for advertisers to be much more prominent, much more top of placement, so visual search ads will appear at the top of those visual search results and on [indiscernible] where users are looking at what they want to buy. And so this unlocks that user interaction for advertisers and they can now engage with users in a way that they can't on other platforms, particularly given that we skew so significantly towards these digital interactions. This is a major sort of user need that we are solving that's different than what they get from traditional search or from chat bots, where 1 of the biggest issues is that users so often don't have the words to express what it is they're looking for.

They have this loosely formed idea. And so that's why I mentioned earlier that 70% plus of our searches start with a proactive recommendation that we are making. So when you pair that proactive recommendation that we make without the user having to ask, which is sort of the promise agentic experiences sort of helping the user enacting on user behalf and then combine that with their ability to then go engage and refine visually and now get great product placements and great ad places that are loving to them right at the top of those results. We think this is a compelling experience for both users and advertisers. It debut Pinterest in September. It's now in beta. And so we're quite excited about it. We're learning with advertisers, and we're excited to share more progresses.

Operator

Your next question comes from the line of Rich Greenfield with LightShed Partners.

Richard Greenfield

A Couple. One, I just wanted to follow up on the DSPs. Why now? I mean, obviously, you could have done this last quarter or last year, Bill, like just -- what was the catalyst to put these into effect earlier this week? And then just I think last quarter, there was a lot of excitement amongst investors that your revenue growth in the U.S., the key market sort of accelerated from lower teens up into the higher teens. And you sort of slowed back down towards that 14% this quarter. Were there any sort of unique items that sort of drove the acceleration, the slowdown or just anything to think about of like, is there the ability to sort of sustain that higher level of growth you had last quarter as we think about the go forward?

William Ready

Yes. Thanks, Rich. Appreciate the questions. So I'll start with the first part and then give Andrew the second part. So on the DSP deals, you've heard us about for some time that we feel like there's a lot of opportunity to monetize our audience in other context. And we started that with bringing third-party demand onto our platform, and we extended that beyond our platform with the acquisition of tvScientific, part of what's enabling us to do more is both our own learnings from how we do that first with bringing demand onto our platform, the learnings we've had with how to monetize our audience on other new services like connected TV, and across the industry, there's been an advance of privacy-preserving technologies that is really helping us think about how we can work with others to meet advertisers where they are. help bring that really unique audience that we have and that great personalization that we're able to do on to more surfaces or in partnership with others in their data.

And part of that why now is like we've been testing and learning our way towards this. and the advance of privacy-preserving technology that give us the confidence that we can do these things in a way that helps deliver better experiences for users and advertisers, but also in a privacy-preserving way. and in a way that make sure that we capture the value from our audience as we engage in these partnerships. And I'll give to Andrew for the second part of your question.

Andrew Somberg

Yes. Thanks, Rich. So UCAN grew 14% in Q3, which represents strong reacceleration from where we ended 2025. As we noted on our last call, Q2 included a few unique onetime benefits, including elevated spend from the World Cup as well as the shift of Prime Day into Q2. So in Q3, we did see some softness in food, beverage and alcohol, which resulted in about a point of sequential pressure in UCAN. At the same time, as we've noted, we are successfully broadening our advertiser base across categories to help offset some of this pressure in food and beverage.

So in addition to retail, we've also seen health inclusive of GLP-1 advertisers growing nicely, and this has helped to partially offset some of the pressure we see from fruit and beverage, so looking to Q4, we feel overall good about UCAN, but those are some of the things -- some of the puts and takes that point too, Richard.

William Ready

Yes, just in closing on that, just say, again, the reacceleration of UCAN along with our really strong user growth and engagement are real highlights in the business. We navigated this sort of regulatory change in the U.S. last year. It's analogous to what we're now going through in Europe, but we've clearly reaccelerated that from last year. And it wasn't just us, you've heard others sort of talk about World Cup and Prime Day and these things sort of shifting the shape of Q2 versus Q3. That's pretty consistent with what you're seeing out there, but when you do the comparison to where UCAN was last year, it's still a very clear reacceleration. We feel good about that. And as you heard in my comments on the guide, we see that strength in UCAN continuing into Q4.

Operator

Your next question comes from the line of Eric Sheridan with Goldman Sachs.

Eric Sheridan

Maybe 2, if I could, quickly. On the back of the user growth you saw, and I just want to sort of pull on this we've been talking about on the call. How should investors think about the duration against the opportunity set to close in ARPU or monetization gap relative to the user growth you've closed, just so we understand how that might correlate with each other over the next 12 to 24 months.

And then in the answer, I think you gave earlier, you were talking primarily about the Pinterest Assistant, I believe. If we widen out conversation to think about horizontal agents. How are you thinking about both the opportunity set that they present as well as some of the friction points that have to be solved for?

William Ready

All right. Thanks for the question, Eric. So on the growth in ARPU, I think if you look at our U.S. business, our UCAN business where we started our shopping journey first. And as a reminder, we only launched our first product, our first ad product designed this and click to advertisers about 3 years ago. So we're pretty early in that monetization journey but we had really great progress. And as we've rebuilt interest into an AI-driven shopping platform that's clearly bringing users here, and you've seen ARPU improvement in UCAN, as we look internationally, we've been making good progress there. But if you look at the ratio of ARPU in Europe or international relative to our U.S. business, there is a meaningful gap for us on that. the user behavior, the shopping behavior we see as a global phenomenon.

So our user growth and engagement being driven by shopping daters is a global phenomenon for us. So in Europe and international, where we started in our largest home market with the ad opportunity, we're now expanding into those markets. And so while we're earlier on, we're running that playbook. We're seeing it take hold, we've had a bit of a speed bump with the European regulatory change, but that's not impacting the user growth and engagement around shopping. And so ultimately, if you have the user engagement on shopping, advertisers will meet that engagement. We're making that more available to advertise both through our ad product and our go-to-market realignment I mentioned the early green shoots there.

So we're excited about that progress. But we're also quite excited that there's a lot of upside. When you look at how low the ARPU is sort of in a -- the relative positioning of that ARPU or the ratio of that to the U.S. and what that ratio would be on other platforms, it says there's a lot of upside in us addressing that ARPU internationally. And as we look into next year, we see the go-to-market changes that we're making internationally starting to take hold. I shared some of those examples, and what we saw in the U.S. was that the sort of tariff adjustments by advertisers can take a couple of quarters.

We can't perfectly predict how this plays out with like the European de minimus exemption change but it's probably a multi-quarter event, but also 1 that just as we reaccelerate in the U.S., we think will reaccelerate in international. And the biggest thing that gives us a [indiscernible] to that is that we see that core your shopping behavior on that. So as we look into next year, we -- again, we don't guide beyond the quarter. But as we look into next year, we expect that we'll manage to reaccelerate that just as we did in the U.S.

On the same part of your question, you mentioned horizontal agents. It's -- I think it's really interesting to note that if you look at how search has played out over the last 20-plus years, it's, while there were a few winners on horizontal search, there were many winners on more verticalized search experiences, Think about how many product searches would start on an Amazon versus on a horizontal search platform, how many travel searches would start on a booking [indiscernible] versus a horizontal search platform. And as we look into future world of agents, we think that there will be broad horizontal agents, but there will be opportunities vertical-specific winners in those agents. And we're very focused on making sure that we deliver a great shopping experience and 1 designed to help the user continue to drive those experiences.

So we feel really good about that progress, and we've been designing that in a way that meets the user where they are, driving mass adoption of that. We're delivering agent experiences in a way that the user feels like is assisting them and helping them but not taking the decision out of their hands. I shared some of the stats around that. And I think that's why as we've been through multiple rounds of improvements in chatbots now the launch of other horizontal agents we have not seen any impact in our UCAN engagement since the launch of some of those popular consumer agentic experiences because those tend to be more horizontal, and users are coming to us for a specific purpose around shopping and 1 where we're able to go really drive experiences that are deeply personalized to them and they feel like letting them control those experiences. So said simply, users -- they're happy to automate away drudgery. They don't want to automate away their joy. They don't want to automate away their entertainment and their past time. And so we've done some research on there's a new research that was conducted by [ YouGov ] on behalf of Pinterest that more than 8 in 10 Americans say they'd missed something about shopping if AI did it for them from discovering things that they weren't looking for to figuring out what they like.

So this is why at Pinterest, we've been building for all the wonderful parts of shopping that people don't want to get up in the same way that people don't seek to automate away their entertainment. And so that's been at the very core of that DP engagement. I mentioned in the U.S., we continue to see deep engagement, especially with Gen Z, the most digitally native audience. We see deep engagement even as these other tools are becoming available. And the big for that is that we're in an expansionary moment for search. And I think we're carving out a unique space for ourselves, particularly around shopping in a way that is highly personalized and it's striking the right balance between the user retaining their own agency, in participating in those decisions versus just giving the decision up to others, which, as I just mentioned in that research, users aren't necessarily wanting to give up the decision. They want help in making decisions, but they still want to be there for the decision. And so that will continue to play out. There's more work to do but we feel really good about the progress we're making there.

Operator

Your last question comes from the line of Shweta Khajuria with Wolfe Research.

Shweta Khajuria

Can you hear me?

Andrew Somberg

We can hear you Shweta.

Shweta Khajuria

Thanks, Andrew. I was on mute. I guess I have 1 on bidding optimization. So AI-driven bidding optimizations that you deployed late in Q1 and then through Q2 we're originating really well with larger advertisers. Any more color you could provide in terms of how that has progressed through the quarter and how we should be thinking about the potential impact tailwinds, if there are any between now and year-end and then to the degree you can going forward beyond that?

William Ready

Yes. Thanks for the question, Shweta. So as you know, we are aligning our AI bidding systems to the advertisers measurement sources of truth. And as we -- and that allows us to deliver hyper specific outcomes that they care about, and we're seeing strong results from that. We've clearly advanced significantly our ability to deliver great personalized recommendations to the user in terms of what they're shopping for, but every advertiser values different events in different ways and has different outcomes that they want to see as those advertisers give us more specifics about how they value different events, whether do they want to focus on a specific product sale? Do they want to get a new customer versus deepening engagement with an existing customer.

As they give us more information about what they desire in the same way that AI is able to be hyper specific about the recommendations that we make to the consumer we can now be hypersialation comes that we drive for the advertiser. So we feel really good about the progress that we've made there. Our focus to date has been on more sophisticated advertisers with proprietary measurement, and as we said earlier this year, we've seen those advertisers leaning in as a result of this. In Q3, as planned, we extended these integrations to a small number of additional sophisticated advertisers, for example, nearly testing with the leading e-commerce platform, that advertiser saw 77% higher new customer ROAS reflected in their proprietary measurement source of truth.

So we feel good about that. we will continue to broaden that out. And it's at an important moment in time because the industry is moving beyond last click attribution, which is ultimately good for a platform like Pinterest but different advertisers are moving at different speeds. And we're continuing to learn from these direct integrations extending that into partnerships with third-party measurement providers to reach the longer tail of advertisers. So for the more sophisticated advertisers, these are direct integrations or taking those learnings to work more with the third-party measurement providers that the longer tail of advertisers use, those adoption curves naturally take time, but we're very encouraged about what we can do there.

Again, we are much further along in the consumer and engagement side of this, but those same techniques about driving hyperpersonalized results that are easy for consumers are allowing us to match hyper-specific results for advertisers as they open up more of their specific attribution around different events. So again, these things take time to play out, but we feel really good about what that means for the long term of the business.

So with that, I want to thank everyone for joining our call today and for your questions. Before we conclude, I also want to thank Julia Donnelly for her contributions to Pinterest. Finally, I want to thank our teams at Pinterest, our advertising partners and all the people that come to Pinterest to find inspiration and take action. I hope you have a great rest of your day.

Operator

This concludes today's call. Thank you for attending. You may now disconnect.