Can ChatGPT ads prove they work?

John Ebbert

ChatGPT measurement partners

In a Friday email to ChatGPT Ads advertisers, OpenAI highlighted recent product updates, including an expansion of its measurement partnerships:

“More measurement partners: We’ve expanded our measurement ecosystem with partners including Northbeam, DV Rockerbox, and INCRMNTAL, plus mobile measurement partners Kochava, Tenjin, and Airbridge. Explore setup guides for measurement partners and mobile measurement partners.”

Among the partners, Northbeam and INCRMNTAL reinforce the importance of incrementality measurement for ChatGPT advertising, although they approach it differently.

    • Northbeam offers multi-touch attribution, media mix modeling and incrementality measurement, capabilities that can help advertisers evaluate ChatGPT’s contribution to business outcomes alongside other channels.
    • INCRMNTAL (acquired by Smartly) specializes in causal incrementality measurement, estimating how much additional business advertising generates beyond what would have happened anyway.

From tipsheet: OpenAI needs to demonstrate that ChatGPT advertising generates incremental business rather than harvesting credit for conversions that might have happened anyway. The addition of measurement partners with incrementality capabilities speaks directly to that challenge.

MARKETING

Fender, Marshall want AI to know their brands

Guitar maker Fender and amplifier and headphone brand Marshall are investing in original content and storytelling, partly to improve how their products appear in AI-powered search results, reports Digiday’s Sam Bradley.

Fender has redesigned its website to emphasize the brand’s history, musicians and craftsmanship rather than simply selling guitars. Marshall is producing original music programming, including a YouTube series featuring performances in recreated, long-closed music venues.

Read: “Brands turn to content series and Jimi Hendrix to improve AI search performance” (Oct. 15) – Digiday (subscription)

From tipsheet: AI search could give original brand content a new commercial purpose: influencing what AI assistants know, cite and recommend. Why wouldn’t you? But whether these investments measurably improve AI visibility remains an open question, as does the role paid media might play in amplifying their impact. Generative engine optimization (GEO) success seems like a constantly moving target.

TECH

Debates: Going server-side

Apple’s new domain-blocking capabilities in WebKit (AdExchanger, October 2) are adding urgency to the advertising industry’s shift toward server-side infrastructure. But moving advertising operations out of the browser raises questions about interoperability, privacy and who ultimately controls the ad stack.

In a LinkedIn article, IAB Tech Lab Sr. Director of Product Rowena Lam argues that the future of advertising is server-side. The ensuing discussion highlights both the promise and the unresolved challenges.

    • Lam makes the architectural case. Moving bidding, identity and other advertising functions to publisher-controlled edge servers could improve speed, strengthen data governance and reduce dependence on browser vendors. IAB Tech Lab’s open-source Trusted Server initiative is already being piloted to support this transition.
    • Read: “Apple’s WebKit Updates Mean Advertising’s Future is Server-Side” (Oct. 13) – Rowena Lam on LinkedIn

The ensuing discussion among commenters revealed several competing perspectives:

    • Kerv’s Joshua Koran identifies the commercial obstacle. Decentralized infrastructure still needs to support cross-publisher frequency capping, measurement and optimization. “Scale matters but so too does interoperability,” he writes. Walled gardens already offer both. Read.
    • 51 Degrees CEO James Rosewell proposes a technical response. Trusted Server implementations could maintain browser-level histories of ad exposures and make relevant signals available to advertisers. Publishers would determine what information to share and how their servers operate. Common standards and coordination mechanisms would still need to emerge. Read.
    • Former Raptive executive Don Marti challenges the privacy assumptions. Moving data processing server-side doesn’t automatically make advertising more privacy-protective. Server-to-server connections can also facilitate tracking and data sharing that consumers cannot easily observe or control. Read. Bertelsmann data executive Achim Schlosser, who is an IAB Tech Lab member, fully agrees.

From tipsheet: Server-side infrastructure could give publishers more control over how advertising operates on their properties, but publishers don’t operate in isolation. Advertisers still need reach, frequency management, measurement and optimization across thousands of properties.

The industry may be able to decentralize execution without sacrificing interoperability and the harder question may be 1) who controls the infrastructure connecting all those independent systems? and 2) whether publishers gain lasting leverage or simply trade one set of gatekeepers for another?

FINANCIALS

The quarterly earnings tidal wave

Earnings season is about to offer a fresh look at the advertising industry’s growth, AI investments and competitive dynamics.

Here’s a selection of upcoming quarterly reports from publicly traded companies with significant advertising exposure:

Additional earnings dates have yet to be announced.

From tipsheet: Beyond the revenue and profit numbers, a few developments are worth watching:

    • Meta and Google: How much are AI-driven improvements contributing to advertising performance and what are the returns on their enormous AI investments?
    • Amazon: Can its advertising business sustain its momentum as the company expands its DSP and commerce media capabilities?
    • AppLovin: Following last quarter’s slight revenue miss, is its self-serve advertising platform beginning to generate meaningful revenue beyond gaming — and particularly from ecommerce advertisers?
    • Magnite: Is CTV growth accelerating and are direct buyer connections through ClearLine and other initiatives beginning to contribute meaningfully to revenue?
    • Agentic advertising: When will agentic buying and selling begin generating material revenue? Will any company venture a forecast or is the industry still measuring progress in pilots and partnerships?

LLMs & CHATBOTS

Developments

    • Open source: “Reflection’s Beam aims to undercut Chinese open models with fewer tokens, not cheaper tokens” (Oct. 16) – VentureBeat
    • Meta and Apple: “Meta brings Muse Spark Open to Apple hardware” – Eric Seufert on Mobile Dev Memo (subscription)
    • China AI: “The Country Cramming for the AI Revolution. In China, the fear of getting left behind has created a powerful new prosperity gospel.” (Oct.15) – New York Magazine

MORE

    • Protocols: “What AdCP 3.2 Means for Marketers” (Oct. 15) – AgenticAdvertising[dot]org
    • Choose your weapon: “Reach CEO touts Google AI deal as a template, steps up pressure on OpenAI over scraping” (Oct. 15) – Digiday (subscription)
    • Addressing ads: “Google begins testing ads in AI Mode for users with Personal Intelligence enabled” (Oct. 16) – Search Engine Land
    • Q3 review: “AI Media Is Starting to Establish Its Own Identity” (Oct. 15) – Debra Aho Williamson, analyst on The AI Ad Economy Substack
    • Euro RMN: “Tesco Media expands advertising capabilities through new technology partnership with Kevel” (Oct. 12) – press release