There are a lot of capability stories out there. There aren't a lot of stories or pictures about the client's experience. We're really trying to look at how we use client experiences as a true differentiator that we showcase.
The MTA has officially kicked off a pilot program that will pipe 30-second audio ads into select subway and commuter rail stations every 10 minutes. The spots will be capped at 75 decibels, roughly the volume of a loud conversation, a vacuum or heavy traffic. The MTA hopes that these ads will generate income to keep the system running.
From the customer's perspective it felt like dealing with multiple companies wearing the same logo. Marketing sends a "We miss you!" email the day after a frustrating support call. Sales doesn't know the customer has already watched a demo. In-store purchase history is invisible to the ecommerce team. No continuity. No memory. No relationship.
In 2026, we are deliberately prioritizing user growth and teaching better. We'll focus on improving the free learner experience to grow word of mouth and feed our next user growth engines like chess, math and music, even though that moderates near-term financial growth.
And, often, the value of this data increases when it's nested within a larger company. For example, Infillion acquired Catalina this week, as it plans to infuse its DSP. In some ways, the MediaMath DSP- Catalina data combo emulates Amazon, whose DSP is superpowered by its ecommerce sales data.
Text guidelines is a campaign-level feature that allows you to refine your brand's message for assets made exclusively with text customization. These guidelines help you not only to meet business and brand requirements, but also to control the output of assets created with Google AI automatically when text customization is "ON".
The Trade Desk announced revenues of $2.9 billion during 2025, up 18% year on year, with $847 million of that total generated during the closing quarter of the year, up 14%, with both sets of numbers beating earlier expectations.
To use the new v23.1 features, you must upgrade your client libraries and client code. All the updated client libraries and code examples have been published. We are also releasing minor versions v20.2, v21.1 and v22.1 of the Google Ads API to add additional support for the EU Political Ads Regulation.
As search optimisation becomes more complex - with AI overviews, LLM citations, and zero-click results reshaping the landscape - choosing an agency with genuine consulting depth has never been more important. In 2026, the most effective consulting agencies combine deep strategic thinking with advanced technology, data-driven decision making, and a thorough understanding of how software businesses acquire and retain customers through organic search.
The 2026 Marketing Data Report basically shows a gap between AI ambition and operational reality in marketing departments that's only getting wider. Which explains why agencies continue to beef up their AI smarts - as a means of helping to guide their clients through the morass of knowledge, strategy and technology to implement gen AI.
Nearly one in the four national ad spots-about 23%-was tied directly to artificial intelligence, either as the product itself or as the technology powering the message. OpenAI and Anthropic promoted ChatGPT and Claude to mass audiences. Google and Amazon positioned Gemini and Alexa+ as embedded assistants inside everyday life. Meta showcased AI-enabled smart glasses as experiential technology.
Regardless of whether a media channel is addressable or not, it's still reaching strategic and high-value audiences. So, understanding how these different media buys interact with each other [helps buyers] make better investment decisions.
We view media quality as an engine for growth. What we hear consistently from brands is, 'Prove to us that media quality drives higher ROI and higher efficiency.' Advertisers increasingly demand verification providers demonstrate concrete performance improvements from premium placements rather than simply protecting against brand safety risks.
YouTube is now a ranking asset in AI discovery, and you can lose share even if website traffic looks stable. YouTube is highly machine-readable (transcripts, metadata, chapters) and tends to be a low-risk source for AI to summarize and cite.
As AI-generated answers increasingly influence buying decisions and brand discovery, teams face new operational questions: Where is the brand being referenced? Where is it missing? How does that representation shift over time? For agencies, the challenge expands to making AI visibility measurable and repeatable across multiple accounts and reporting cycles. "Agencies and start-ups aren't looking for another dashboard, they're looking for confidence,"
In its fourth-quarter earnings call last week, Roku announced its platform revenue - which includes ads sold both on its Roku Channel FAST service, as well as display ads on its home screen - grew 18% year-over-year to reach $1.2 billion, pushing total growth for the entire year to a similar 18% and total revenue of $4.1 billion. It's a sign the connected TV (CTV) ad market is maturing - though experts told The Daily Upside it's still far from mature.
In this tutorial Grace Leung explains how to create an AI-supported marketing team using Claude Skills in just 16 minutes. Claude Skills are structured workflows that translate standard operating procedures and brand guidelines into repeatable AI-driven processes. These workflows can handle tasks such as content creation, data analysis, and campaign management, allowing teams to maintain consistency while saving time for higher-level strategic work.
According to an SEC filing dated Feb. 18, 2026, Arohi Asset Management PTE Ltd. sold all 1,717,770 shares of DoubleVerify (NYSE: DV) during the fourth quarter. The estimated transaction value was $20.58 million based on the quarterly average price. The quarter-end value of the position fell by $20.58 million, reflecting the complete exit and changes in share price during the period.
Perplexity AI has pulled back from advertising, citing concerns that sponsored content could erode user trust. Executives confirmed the phaseout, arguing that confidence in the chatbot's answers is essential to maintain engagement and subscription revenue. Meanwhile, Meta doubled down on AI infrastructure, signing a multiyear, multibillion-dollar deal with Nvidia to purchase millions of advanced chips. The agreement reinforces Nvidia's dominance in the AI data centre market and follows Meta CEO Mark Zuckerberg's plan to nearly double the company's AI capabilities.
Ads in chatbots are becoming standard practice, and advertisers and AI companies are scrambling to determine how to make the most out of this new format. They need to deliver relevant, contextual ads that don't feel invasive or overly repetitive and fit into the chatbot's interface while still being clearly distinct from the organic chat responses. Since LLMs are a mostly untested channel, though, the ones choosing to introduce ads don't have much data to influence their strategy.
