互联网广告正进入一个“复杂而易变的时代”。

技术发展、市场碎片化以及人工智能的快速崛起，让广告商既看到了效率提升的曙光，也感受到了前所未有的压力。广告主既要在全球范围内寻找新的增量市场，又必须时刻守护品牌适用性、避免落入“黑箱”算法带来的不透明陷阱。

作为全球头部的媒体质量与广告效果平台，DoubleVerify（DV）不仅在纽约证券交易所上市，还通过收购和持续的技术创新让自己从“验证服务商”升级为集品牌适用性、优化与归因于一体的全栈平台。

10月3日，Morketing Group CEO曾巧受邀参与DoubleVerify（简称DV）新加坡的IMPACT 2025峰会，并对话了DoubleVerify CEO Mark Zagorski，梳理了DV的发展逻辑与解决方案，并针对中国品牌出海的痛点提出建议，解答为什么越来越多国际品牌选择 DV 作为出海必备伙伴。

WaveGlocal 曾巧：在过去几年里，我们发现海外市场的数字广告投放越来越复杂，品牌适用性、流量欺诈等问题层出不穷。您觉得这一趋势是如何形成的？

DoubleVerify CEO Mark：确实，这些年全球广告生态发生了巨大的变化。首先，媒体环境高度碎片化，广告渠道从搜索、电商到社交短视频，从开放网络到各类“围墙花园”（例如社交平台、视频平台）层出不穷。碎片化带来的直接结果就是管理复杂度提升，广告主要在众多平台中进行投放、监测与优化，需要投入大量人工和时间。

DoubleVerify《2025 全球洞察报告》指出，营销人员平均每周有 26% 的工作时间花在重复的手动优化上，相当于每人每周超过 10 个小时的时间耗费在修改出价系数、重新分配预算和调整曝光阈值上。

WaveGlocal 曾巧：10 个小时！换算成人力成本就很惊人了，尤其对已经在国内内卷、出海也要拼效率的中国企业来说。另一个问题是品牌适用性。在全球化投放过程中，很难保证广告不会出现在不合适的内容旁边，这个问题是否也在加剧？

DoubleVerify CEO Mark：是的。首先，内容平台每分钟都会产生大量新的内容，例如 YouTube 每分钟上传的内容超过 500 小时，平均每天新增视频 2 千万条，内容质量参差不齐，人工审核几乎不可能完成。其次，算法决定了广告呈现位置，这些算法通常是“黑箱”，广告主难以看到具体投放在哪些频道或视频。我们早期曾经帮助客户分析一个广告投放列表，发现 95% 的链接都是看不见的域名，剩下 5% 能看到的内容质量也很差。很多广告主“交钱即托管”，对投放环境完全失去控制，这就是所谓的“黑箱问题”。

另一方面，广告欺诈和无效流量也日趋严重。DoubleVerify调研报告发现，无效流量（IVT）在 2024 年比上一年增长了 101%(1)。这些无效流量不仅来自传统的僵尸网络，还包括利用人工智能生成的高级欺诈应用，模拟真实用户行为骗取广告预算。由于 AI 技术变得越来越先进，欺诈分子可以快速调整行为模式以逃避检测。这对品牌是巨大的威胁，因为投入的预算被浪费，甚至可能导致品牌出现在虚假或低质内容旁边。

WaveGlocal 曾巧：品牌适用性和效率压力同时存在，还要应对跨境文化差异和合规问题。这使得出海广告主不得不权衡“成本”和“质量”的矛盾。除此之外，您提到的“围墙花园”又带来了哪些挑战？

DoubleVerify CEO Mark：所谓拥有用户生成生态系统的“围墙花园”，不同平台之间体验差异很大，广告主难以统一管理跨平台投放。此外，围墙花园中的内容高速变化，可能同一频道里 90% 的内容适合品牌，10% 却不适合，如果人工去筛选显然不可行。研究显示(2)，近半数亚太地区消费者表示，若广告出现在令人反感的内容旁边，会降低其购买该品牌产品的意愿。这种影响在有年幼子女的家长中尤其明显，其中有 52% 表示，内容不符的广告会对他们的购买决策产生负面影响。因此，“高质量曝光”与“高效率”之间的平衡，成为全球广告主普遍面临的困境。

WaveGlocal 曾巧：如果说上一节我们描述了痛点，接下来我们想具体分析这些问题到底出在哪里？广告主在追求效果与质量时面临哪些“悖论”？

DoubleVerify CEO Mark：我们将这称为“广告主的两难”（marketer’s dilemma）。从一端来看，品牌需要更好的投放环境，希望广告出现在高质量、与品牌匹配的内容旁边，避免品牌适用性风险。要实现这一点，广告主往往需要启用各种品牌适用性策略，例如排除低质内容、按照敏感类别过滤，但这意味着可用的流量池被缩小，单次曝光成本上升。如果只关注质量，可能会错过更广泛的受众，整体覆盖率受限，效率指标（如 CPM、CPC）会被拉高。

另一方面，如果只追求成本效率，广告系统容易把出价下调至极低的媒体库存，甚至出现在不合适的频道或应用里。由于大部分程序化媒体采用竞价模式，低价往往意味着质量差，品牌可能会出现在低质甚至虚假内容中，严重损害品牌形象。

过去十年的经验告诉我们，一些频道整体看来 90% 内容符合品牌要求，但内部单个视频可能有 10% 的内容不适合，如极端暴力或不当内容，而人工很难做到精准过滤。正是在这种“质量与效率冲突”的背景下，我们开始思考如何通过技术手段打破二元悖论。

WaveGlocal 曾巧：这里背后的关键就是“数据”和“人工智能”的力量。能否详细介绍 DV 如何用数据来分析这些风险？

DoubleVerify CEO Mark：DV 是全球领先的媒体质量平台，我们每年测量超过 8 万亿次媒体交易，积累了海量的曝光、点击和互动数据，包括曝光的时长、可见性、互动度、内容文本与语音等。这些数据帮助我们了解真实的广告环境。例如，在DV最新的全球洞察报告中，DV 发现社交媒体短视频和信息流被认为是表现最好的渠道，80% 的受访广告主更偏爱短视频投放，79% 偏爱信息流广告。然而，同一份报告也揭示了隐患：近一半的用户安装了广告拦截器，机器人流量和无效流量激增。通过这些洞察，我们得以提供更加精准的风险分析。

与此同时，DV 利用自然语言处理、图像识别、音频识别等人工智能技术，对视频、图片、文字和音频进行逐帧分析，快速识别内容主题和风险级别。我们的系统支持超过 40 种语言，能够理解全球不同地区的文化语境。此外，我们还接入数百家供应商和媒体平台的数据，分析不同渠道的曝光质量与有效性，并通过算法实时发现异常模式。这些算法不仅用于检测是否存在可疑的点击和曝光，还用于预测哪些位置的广告更有可能产生关注度或转化。

WaveGlocal 曾巧：AI 技术不仅用来识别风险，还能用来“优化出价”对吗？

DoubleVerify CEO Mark：是的。DV 在 2023 年收购了法国算法竞价公司 Scibids，这家公司专门做程序化出价优化，其技术能够基于广告主的关键指标（例如转化率、访问时长等）实时调节每一次竞价，实现“千人千面”的出价策略。该技术可与主要的需求方平台（DSP）兼容，提升广告效率。我们的研究表明，与传统优化策略相比，通过算法优化可以降低 45% 的 CPM，提升 63% 的注意力指标，并将有效曝光提升近一倍。收购后，DV 把这些算法整合进自身平台，为广告主提供从投放前预测、投放中优化到投放后归因的全链路解决方案。

WaveGlocal 曾巧：分析完问题，想请您具体介绍 DV 的解决方案和产品。这几年 DV 在产品层面做了多次升级，最新推出了“Media AdVantage Platform”和“Authentic AdVantage”方案，它们分别解决什么问题？

DoubleVerify CEO Mark：为了应对上述困境，我们把 DV 的能力整合成一个统一的平台，称为 DV Media AdVantage Platform （简称 DV MAP）。MAP 代表三个核心模块：验证（Verification）、优化（Optimization）和证明（Prove）。

在验证方面，我们提供完整的品牌适用性、可见度、无效流量和受众适配检测工具，确保广告展现在合适的内容和受众群体中。在优化方面，我们利用 AI 算法实时调整出价和投放策略，在保证质量的同时压缩成本，并通过自动化工具减少人工操作。在证明方面，我们通过归因分析和增量测量帮助广告主理解广告的真正价值，形成“真实性证明”。这些模块不是孤立的，而是在一个统一的界面内彼此联通，实现数据共享和决策自动化。

WaveGlocal 曾巧：听起来 DV MAP 是一个底层的平台架构，而针对 YouTube 等围墙花园，你们推出的“Authentic AdVantage”是如何运作的？

DoubleVerify CEO Mark：Authentic AdVantage 可以理解为 DV MAP 的一个具体解决方案，尤其适用于社交媒体和视频平台。它主要集成了三项能力：

预投放控制（Pre‑Bid Controls）：我们通过收购OpenSlate公司，获得了专业的内容评分和预投放排除技术。系统会在广告竞价前快速分析视频的主题、情感和敏感程度，对不适合品牌价值观的内容进行自动排除。统计显示，启用预投放控制后，品牌遭遇风险的概率降低了 95%。

AI 出价优化（DV Scibids AI）：这是 Scibids 算法的核心，它会根据广告主设定的目标（如 CPA、CPV、ROI 等），动态调整每一次出价和预算分配，在复杂的竞价环境中找到成本效率最优解。与传统手工调整相比，AI 出价能够在保证效果的前提下将平均成本压缩 10% 甚至更多。

效果测量与归因（DV Pinnacle）：我们将所有投放数据汇总到 DV 的Pinnacle 界面，广告主可以实时看到品牌适用性得分、视听体验、曝光质量以及最终的转化情况。通过统一的界面，广告主能够更直观地理解不同策略对 ROI 的影响，并据此制定下一步决策。

联合这三项能力，Authentic AdVantage 真正解决了“质量与效率不可兼得”的悖论。在测试中，我们将 OpenSlate 的品牌适用性过滤和 Scibids 的AI竞价算法同时应用在 YouTube 视频广告上，结果发现品牌适配度提升 10%，同时每次观看成本下降了约 15%，覆盖人群增加 60%。这说明在智能算法的帮助下，广告主能够同时享受高质量曝光和成本优势，而不再需要在两者之间做出妥协。

WaveGlocal 曾巧：：这些数字非常亮眼。对于品牌来讲，具体如何使用这套系统？需要重新购买新的广告位吗？

DoubleVerify CEO Mark：完全不需要离开熟悉的平台。Authentic AdVantage 通过 API 与 Google Ads Manager、DV360 等需求方平台对接，广告主只需在 DV Pinnacle 界面登录账户，通过简单配置，即可一键调用我们的预投放过滤和出价优化策略。同时，系统会自动收集投放数据，形成实时报告，帮助广告主调优目标。例如，广告主可以同时设定“品牌适用性达成率 ≥95%”和“CPV≤X 美元”的目标，系统会自动计算出最优的投放组合，省去了繁琐的手动对比和调整。

此外，针对海外电商和游戏企业，我们的算法还可以按照用户留存等深层指标进行优化。例如，一个手游客户希望提升广告带来的玩家留存率，系统会根据用户在广告后下载、注册、付费的数据，不断调整投放策略，实现真正的效果优化。

WaveGlocal 曾巧：：技术思路听起来很完美，但真正让人信服的是实际案例。您能不能分享一些品牌使用 DV 服务后取得的成果？尤其是中国品牌出海的案例。

DoubleVerify CEO Mark：当然。我们合作了全球许多知名品牌，包括宝洁、联合利华、微软、等。通过 DV 的平台，联合利华能够监测全球几十个国家的广告投放，并在同一界面查看各市场的品牌安全、可见度和无效流量指标。同时，我们利用 Scibids AI 算法优化投放策略，降低了平均 CPM，提升了转化率。这让联合利华不仅确保了品牌形象，也节省了一大笔预算。

在中国品牌方面，有一个典型案例是某全球性银行（香港地区）使用 DV 的社交媒体预投放解决方案(3)。该品牌担心广告出现在与金融业务不相关甚至令人反感的内容旁边，影响用户信任。使用 DV 的内容识别与过滤技术后，广告避免了不合适的内容，相比传统投放提升了 20% 的潜在客户获取，获得的高质量线索提升了 10%，客户满意度明显提高。这一案例显示，对于金融等高敏感度行业，精准的内容匹配尤为关键。

还有一家消费电子品牌（考虑品牌保护，匿名）在出海投放过程中，面临广告欺诈严重和覆盖不足的问题。启用DV的Authentic AdVantage 后，系统识别并排除了大量来自虚假 App 的流量，同时通过 Scibids AI 算法调节出价，降低了 50% 的无效曝光成本。最终，他们在保持广告预算不变的情况下，达到近 75% 的覆盖增长和 50% 的 CPM 效率提升。这一成绩让品牌方意识到，AI不只是一个 buzzword，而是真正可以转化为实实在在收益的工具。

WaveGlocal 曾巧：你们近期还收购了 Attribution 平台 Rockerbox，准备完善最后一环——“证明”。能否谈谈你们在结果归因方面的想法？

DoubleVerify CEO Mark：是的。今年 DV 收购了 Rockerbox，这是一家专注于多触点归因和营销组合建模的公司。以往很多广告主只能依赖平台提供的归因数据，而这些数据往往缺乏透明度甚至存在夸大。Rockerbox 能够将广告曝光、点击和最终销售结合在一起，通过多种模型计算广告对销量的贡献。比如，可以告诉广告主某一次投放引发了多少自然搜索增长、促成了多少购买，而不是简单地看点击数。我们计划在 DV MAP 中整合 Rockerbox 的能力，使客户不仅能看到质量和成本数据，还能直接衡量广告对生意的影响。这对于中国出海品牌尤为重要，归因能够帮助企业理解广告投入与最终 GMV 之间的关系，优化营销预算。

WaveGlocal 曾巧：当下很多“AI 广告平台”宣称不需要广告主干预，一键托管即可，DV 与这些平台有什么区别？

DoubleVerify CEO Mark：最大的区别在于透明与独立。很多自动化投放工具属于媒体平台或 DSP 本身，它们的算法和数据不对广告主公开，优化策略很可能优先考虑平台自身的利益。而 DV 是独立第三方，我们不出售媒体，只专注于帮助广告主优化媒体采购。我们的算法完全透明，广告主可以了解每一次出价的依据以及为何排除某些内容。这种透明性能够保护广告主的决策权和数据安全。此外，我们的技术覆盖开放网络、CTV、社交媒体、零售媒体等多个场景，可以跨平台统一分析和优化，这也是许多围墙花园内置工具无法做到的。

WaveGlocal 曾巧：很多中国出海广告主可能还是从传统“点状优化”出发，例如先做品牌适用性、再做出价优化、再做归因，各家供应商各做一套。为什么 DV 强调必须打通三个环节？

DoubleVerify CEO Mark：原因很简单：如果只关注其中一环，很难发挥 AI 的最大价值。举个例子，某品牌只购买品牌适用性方案，那么虽然广告避免了风险内容，但成本可能上升，也不清楚广告是否真正促进了销售。如果仅使用出价优化算法，那么算法可能在一些低质媒体上获得低价，但对品牌形象造成潜在伤害。再比如归因，如果单独依赖某个平台的归因，可能会夸大该平台的贡献。因此，我们提出了“验证+优化+证明”的闭环。只有在同一套数据系统中，同时考虑品牌适合度、成本效率和商业结果，才能真正实现“质价双赢”。

WaveGlocal 曾巧：回到文章主题，您认为 DoubleVerify 为什么会获得那么多品牌的青睐？对于中国出海企业来说，最大的价值是什么？

DoubleVerify CEO Mark：首先，DV 能够帮助品牌摆脱繁琐的手动工作，把有限的团队资源从重复操作中解放出来。《2025 全球洞察报告》显示，91% 的营销人员正在使用或计划使用第三方 AI 或自动化竞价工具。DV 正是利用 AI 来自动化内容识别和出价决策，帮助团队专注于策略制定。其次，我们强调独立透明，保护品牌适用性。无论是预投放过滤、实时监测还是归因报告，广告主都可以实时查看背后的数据和算法逻辑。第三，DV 强调全球化和本地化并重，我们的模型支持 40 多种语言的语义分析，可灵活适配不同文化和法律环境，这对于中国品牌进军海外具有重要意义。最后，DV 不断引入新的技术，例如借助 Rockerbox完善归因，或将注意力指标等新维度融入优化算法，目的都是帮助广告主提升 ROI，实现“花小钱办大事”。

WaveGlocal 曾巧：那么，如果要用三个关键词来概括 DV 帮助中国品牌出海的核心价值，您会如何总结？

DoubleVerify CEO Mark：可以用三个词：可信（Authenticity）、效率（Efficiency）、增长（Growth）。所谓可信，是指我们的平台保证品牌曝光环境真实、适用性、无欺诈；效率，指通过 AI 算法压缩无效成本、提高工作效率；增长，则体现在利用验证和优化后的高质量媒体带来更好的商业结果。这三者相辅相成，构成了 DV 服务的核心价值。

WaveGlocal 曾巧：最后一个问题，您对未来 AI 在数字广告中的作用有什么期待？

DoubleVerify CEO Mark：AI 已经深刻改变了数字广告，它不仅帮助我们识别内容、优化出价，还能为广告主提供洞察和预测。随着技术进步，AI 将更好地理解用户意图和情境，让广告真正做到“为人服务”而非打扰。与此同时，AI 带来的欺诈手段也会更复杂，平台需要持续创新以保护品牌。DoubleVerify 会继续投资于 AI 技术，在保障品牌适用性的基础上，推动广告行业向更透明、更高效的方向发展。

(1)https://doubleverify.com/ai-crawlers-and-scrapers-are-contributing-to-an-increase-in-general-invalid-traffic/

(2)https://doubleverify.com/2025-dv-global-insights-apac-report-chinese/

(3)https://doubleverify.com/dvs-social-pre-bid-solution-improved-brand-suitability-for-standard-chartered-bank-hong-kong/

英文版本：

A Conversation with DoubleVerify CEO Mark Zagorski: Using AI to Break the Quality–Cost Paradox and Reshape Global Media Buying

By Zeng Qiao, Founder & CEO, Morketing Group

The internet advertising industry is entering a “complex and volatile era.”

Rapid advances in technology, the fragmentation of media, and the rise of artificial intelligence have given brands a glimpse of new efficiencies—alongside unprecedented pressure. Advertisers must hunt for growth across global markets while safeguarding brand suitability and avoiding the opacity of “black‑box” algorithms.

As a leading software platform to verify media quality, optimize ad performance and prove campaign outcomes, DoubleVerify (DV) is listed on the New York Stock Exchange. Through acquisitions and ongoing innovation, the company has evolved from a verification provider into a full‑stack platform that spans verification, optimization, and attribution.

On October 3, Morketing attended DoubleVerify’s IMPACT 2025 summit in Singapore and sat down with DV CEO Mark Zagorski. We unpacked DV’s product strategy and solutions, discussed the pain points Chinese brands may face when going global, and explored why so many international brands now view DV as a must‑have partner for overseas expansion.

1.The Global Advertiser’s Dilemma: Balancing “High‑Quality Exposure” with “High Efficiency”

Morketing (Zeng Qiao): In recent years, we’ve seen digital advertising overseas become increasingly complex, with mounting issues around brand suitability and ad fraud. What’s driving this trend?

Mark Zagorski: You’re right—the ad ecosystem has changed dramatically. First, the media landscape is highly fragmented. Channels range from search and e‑commerce to short‑form social video, and from the open web to various “walled gardens” such as social and video platforms. Fragmentation directly increases management complexity. Advertisers must deploy, monitor, and optimize across many platforms—consuming significant time and manual effort.

DV’s 2025 Global Insights Report shows marketers spend 26% of their workweek on repetitive manual optimizations—the equivalent of 10+ hours per person per week tweaking bid multipliers, reallocating budgets, and adjusting frequency and exposure thresholds.

Morketing (Zeng Qiao): Ten hours! That’s a huge labor cost—especially for Chinese companies that face fierce competition at home and must be hyper‑efficient abroad. What about brand suitability? In global campaigns it’s hard to guarantee your ads won’t appear next to unsuitable content. Is that problem getting worse?

Mark Zagorski: It is. First, content creation is explosive. For example, YouTube sees over 500 hours of video uploaded every minute—roughly 20 million new videos a day—with wildly varying quality. Human review alone is impossible. Second, algorithms determine ad placement, and those algorithms are often opaque. Advertisers may not see exactly which channels or videos their ads ran on. Early on, we analyzed a client’s placement list and found 95% of the links pointed to obscured domains; the remaining 5% were low‑quality. Many advertisers “pay and pray,” ceding control to black‑box systems. That’s the core of the opacity problem.

At the same time, DV Fraud Lab revealed that in late 2024, the internet saw a major surge in general invalid traffic (GIVT) rates — nearly doubling with an 86 percent year-over-year increase in the second half of 2024 . Fraud isn’t just botnets anymore—AI‑generated sophisticated schemes now mimic real user behavior to siphon budgets. As AI grows more powerful, fraudsters quickly adapt to evade detection. That wastes spend and can place brands next to fake or low‑quality content—an acute risk.

Morketing (Zeng Qiao): So marketers are squeezed by brand‑suitability and efficiency pressures, plus cross‑border cultural and compliance challenges. And then come the walled gardens—what unique challenges do they add?

Mark Zagorski: “Walled gardens” that have user-generated ecosystems deliver massive reach, but advertisers can’t access all the underlying data, and controls vary by platform. Managing cross‑platform buys consistently is tough. Content also changes rapidly. A channel might be 90% suitable for a brand, but 10% isn’t—and manual screening isn't feasible. Our research shows nearly half of APAC consumers say they’re less likely to purchase from brands whose ads appear alongside content they find objectionable. This effect is especially pronounced among parents with young children, 52% of whom said misaligned ads would negatively influence their purchase decisions. Balancing high‑quality exposure with high efficiency has become a universal challenge.

2.The “Paradoxes” in Pursuing Performance and Quality

Morketing (Zeng Qiao): If the first section outlined the pain points, let’s now pinpoint the root causes. What paradoxes do advertisers face when chasing both performance and quality?

Mark Zagorski: We call it the marketer’s dilemma. On one side, brands want better environments—ads adjacent to high‑quality, brand‑aligned content to reduce risk. That typically means brand suitability policies—excluding low‑quality content and filtering for sensitive categories. The tradeoff is a smaller inventory pool and higher unit costs. Focus too much on quality, and you may miss reach; CPMs and CPCs rise.

On the other side, a cost‑only mindset drives bids toward the cheapest inventory—even unsuitable channels or apps. In auction‑based programmatic markets, low price often correlates with low quality, raising the risk of adjacency issues that can harm brand equity.

From a decade of experience, we’ve learned that a channel that’s broadly 90% suitable can still have 10% problematic videos—extreme violence, inappropriate themes—hard to filter manually. That conflict between quality and efficiency pushed us to find a technical way to break the binary tradeoff.

Morketing (Zeng Qiao): Which brings us to data and AI. How does DV use data to analyze these risks?

Mark Zagorski: DV measures over 8 trillion media transactions annually, capturing massive volumes of impression, click, and engagement data—viewability, exposure duration, interaction, plus text and audio signals. This helps us map real ad environments. In our latest Global Insights Report,APAC advertisers rated short‑form social video and in‑feed ads as top‑performing channels—80% prefer social media reels; 79% prefer feed formats. But there’s a catch: nearly half of users have installed ad blockers, and bot/invalid traffic is surging. These insights sharpen our risk models.

In parallel, we apply NLP, computer vision, and audio analysis to parse video, images, text, and sound frame by frame, classifying themes and risk levels quickly. Our models support 40+ languages, capturing local nuance. We also ingest data from hundreds of supply partners and platforms to benchmark exposure quality and effectiveness, using algorithms to flag anomalies in real time. Those same models don’t just detect suspicious activity; they also predict which placements are likeliest to drive attention or conversion.

Morketing (Zeng Qiao): AI isn’t just for risk detection—it can optimize bidding, right?

Mark Zagorski: Exactly. In 2023, DV acquired Scibids, a French algorithmic bidding company specializing in programmatic optimization. Scibids dynamically tunes each bid based on the advertiser’s KPIs—conversion rate, session duration, and more—enabling granular, impression‑level bid strategies. It’s compatible with major DSPs and lifts efficiency. Our studies show algorithmic optimization can cut CPMs by 45%, increase attention metrics by 63%, and nearly double effective exposures versus traditional strategies. We’ve integrated Scibids into DV’s stack to deliver an end‑to‑end solution—from pre‑bid prediction and in‑flight optimization to post‑campaign attribution.

3.How DoubleVerify Solves It

Morketing (Zeng Qiao): You’ve made several product upgrades in recent years. Can you walk us through the Media AdVantage Platform and Authentic AdVantage—what problems do they solve?

Mark Zagorski: We unified DV’s capabilities into a single platform called the DV Media AdVantage Platform (DV MAP)—built around three modules: Verification, Optimization, and Prove.

• Verify delivers comprehensive brand suitability, viewability, IVT, and audience suitability checks, ensuring ads run in the right environments for the right audiences.

  
  

• Optimize uses AI to adjust bids and delivery in real time, maintaining quality while compressing costs, and automates workflows to cut manual effort.

  
  

• Prove provides attribution and incrementality measurement so advertisers can see the true business value of their media.

These modules are connected within one interface. Data flows across them to power automated decisions.

Morketing (Zeng Qiao): So DV MAP is the foundation. How does Authentic AdVantage work for walled gardens like YouTube?

Mark Zagorski: Think of Authentic AdVantage as a solution built on DV MAP, tailored for social and video platforms. It integrates three capabilities:

Pre‑Bid Controls: Through our OpenSlate acquisition, we brought in best‑in‑class content scoring and pre‑bid exclusion. Before bids happen, the system rapidly analyzes topics, sentiment, and sensitivity to automatically exclude content that’s off‑brand. With pre‑bid controls on, the probability of brand‑risk exposure drops by 95%.

DV Scibids AI Bidding: Scibids dynamically optimizes every bid and budget allocation to the advertiser’s goals (CPA, CPV, ROI, etc.). Compared with manual tactics, AI bidding can compress average costs by 10% or more while preserving outcomes.

Measurement & Attribution via DV Pinnacle: We unify all campaign data in Pinnacle, where marketers see brand‑suitability scores, viewing experience, exposure quality, and downstream conversions in real time. One interface makes ROI tradeoffs crystal clear.

Combined, Authentic AdVantage resolves the supposed tradeoff between quality and efficiency. In testing on YouTube, pairing OpenSlate filtering with Scibids AI yielded a 10% lift in brand suitability, ~15% lower CPV, and 60% more reach—delivering both quality and cost advantages without compromise.

Morketing (Zeng Qiao): Impressive numbers. From a workflow standpoint, do brands need to buy new inventory?

Mark Zagorski:No. Authentic AdVantage connects via API with Google Ad Manager, DV360, and other demand platforms. Advertisers simply log in to DV Pinnacle, configure a few settings, and activate our pre‑bid filters and AI bidding in one click. The system then aggregates delivery data into real‑time reporting and optimization. For example, you can set “Brand Suitability Compliance ≥ 95%” and “CPV ≤ $X” simultaneously, and the system computes the optimal mix—no more tedious manual comparisons.

For cross‑border e‑commerce and gaming, our algorithms can optimize to deeper metrics like retention. If a mobile game wants higher post‑ad retention and monetization, the system uses downstream signals—downloads, registration, in‑app purchase—to continuously recalibrate bids and placements for true outcome optimization.

4.Why DV? Closing the Loop with “Verification + Optimization + Prove”

Morketing (Zeng Qiao): Theory is great, but results convince. Can you share outcomes from brands using DV—especially Chinese brands going global?

Mark Zagorski: We work with many global leaders—Unilever, Microsoft, and more. Through DV’s platform, they can monitor campaigns across dozens of countries and review brand suitability, viewability, and IVT in a single dashboard. Adding Scibids’ optimization reduced average CPMs and improved conversion rates—protecting the brand while saving substantial budget.

Among Chinese brands, a global bank (Hong Kong) used DV’s social pre‑bid solution. Concerned about adjacency to off‑topic or objectionable content, they applied our content analysis and filtering. Result: +20% in potential‑customer acquisition and +10% in higher‑quality leads versus traditional tactics—especially meaningful in sensitive sectors like finance.

A global CPG client wanted to drive cost-efficiency without sacrificing quality. When they activated DV Authentic AdVantage, they saw a 36% lift in cost efficiencies with our algorithmic bidding optimizations. They also experienced a 10% lift in brand suitability with our pre-bid protection applied. Overall Authentic AdVantage helped them achieve a 12% more-efficient quality CPM, enabling the brand to meet their cost and suitability goals, all without negatively impacting scale.

Morketing (Zeng Qiao): You recently acquired attribution platform Rockerbox to complete the “Prove” piece. How do you think about results measurement?

Mark Zagorski: Yes. DV acquired Rockerbox, which specializes in multi‑touch attribution (MTA) and marketing‑mix modeling (MMM). Many advertisers have relied on platform‑reported attribution, which can lack transparency and sometimes overstate impact. Rockerbox links impressions, clicks, and sales to quantify contribution through multiple models—for example, how a campaign lifted organic search or drove incremental purchases, not just clicks. We’re integrating Rockerbox into DV MAP, so clients can see quality and cost and tie media directly to business outcomes. For Chinese brands going overseas, this is critical—attribution connects ad spend to GMV and informs budget allocation.

Morketing (Zeng Qiao): Plenty of “AI ad platforms” claim you can hand over the keys and walk away. What sets DV apart?

Mark Zagorski:Transparency and independence. Many automated tools are owned by media platforms or DSPs; their algorithms and data aren’t fully open, and their optimization may prioritize the platform’s interests. DV is an independent third party—we don’t sell media. Our algorithms are transparent: advertisers can understand why a bid was made—or a placement excluded. That protects decision rights and data security. Our tech spans the open web, CTV, social, and retail media, enabling cross‑platform analysis and optimization that walled‑garden tools can’t replicate.

Morketing (Zeng Qiao): Some Chinese advertisers still approach this in “point solutions”—first brand safety, then bid optimization, then attribution—each from different vendors. Why insist on connecting all three?

Mark Zagorski: Because focusing on one piece limits AI’s value. If you buy only brand suitability, you may avoid risky content—but costs can rise and you still don’t know if sales improved. If you use bid optimization alone, you might win cheap impressions on low‑quality media—hurting the brand. If you rely on a single platform’s attribution, its impact may be overstated. Our “Verify + Optimize + Prove” loop brings these together on one data spine. Only by jointly considering suitability, cost‑efficiency, and business results can you achieve quality and price wins at the same time.

5.In the AI Era: Making Advertising Serve People—Not Interrupt Them

Morketing (Zeng Qiao): Bringing it back to our theme—why do so many brands choose DV? What’s the core value for Chinese companies expanding overseas?

Mark Zagorski: First, DV frees teams from manual drudgery. The 2025 Global Insights Report shows 91% of marketers use or plan to use third‑party AI or automated bidding tools. DV uses AI to automate content classification and bid decisions, so teams can focus on strategy. Second, we emphasize independence and transparency to protect brand suitability. From pre‑bid filtering and real‑time monitoring to attribution, advertisers can see the underlying data and logic. Third, we balance global scale with local nuance—our models support 40+ languages and adapt to different cultures and regulatory frameworks, which is vital for Chinese brands abroad. Finally, we keep innovating—integrating Rockerbox to strengthen attribution and incorporating attention and other advanced signals into optimization—all to improve ROI and help brands do more with less.

Morketing (Zeng Qiao): If you had to summarize DV’s core value for Chinese brands going global in three words?

Mark Zagorski:Authenticity, Efficiency, Growth. Authenticity means real, suitable, fraud‑free environments. Efficiency means eliminating waste and raising operational productivity with AI. Growth is the outcome—verification plus optimization produce higher‑quality media that drives better business results. The three reinforce each other—that’s the essence of DV’s value.

Morketing (Zeng Qiao): Final question: What role will AI play in digital advertising going forward?

Mark Zagorski: AI has already transformed advertising. It helps us understand content, optimize bids, and generate insights and forecasts. As technology advances, AI will better understand intent and context, making advertising serve people rather than interrupt them. Of course, AI will also power more sophisticated fraud, so platforms must keep innovating to protect brands. DV will continue to invest in AI—advancing transparency and efficiency while safeguarding brand suitability.