Your Dashboard Is Missing Half the Story
AI Search Isn't Small. Your Measurement Is.
Bing just gave marketers the first real read on AI search demand.
What's the News:
Bing Webmaster Tools continues to expand its AI search reporting with a new Topics Portfolio report. Now in beta within the AI Performance tab, it groups the queries where your site is cited into broader themes, then shows how much AI search demand each theme represents, how crowded it is, and where your site has room to grow. It’s the latest in a string of AI reporting updates Microsoft has shipped this year.
Why It Matters:
AI search is still a black box: For most marketers, AI search remains hard to research and harder to measure. This year, Bing has repeatedly pushed the limits with new capabilities that help marketers better understand the opportunity.
Traffic data undersells it: AI platforms drove just 0.32% of total website traffic in 2026, which is why so many teams write AI search off. But referral clicks only count the visits AI sends, not how often it answers questions in your category or cites your competitors instead of you.
Most marketers still consider AI search “a drop in the bucket” because GA4 shows it driving less than 1% of traffic. That drastically underestimates its impact, and it’s a result of not having the right tools to size the opportunity. Bing is changing that.
Marketers Should:
Stop sizing AI search by GA4 alone. Referral traffic measures the clicks AI sends, not the demand it answers.
Open the Topics Portfolio tab in Bing Webmaster Tools. Topic Size estimates total citations across all sites for each theme, one of the only first-party estimates of AI search demand available today.
Work the Top Opportunity Queries list. Bing shows the queries where other sites are cited and yours isn’t, along with estimated citation ranges, so you know where to build next.
Use it to prioritize, not forecast. Topics are AI-classified and the numbers are approximate. Treat them as a directional map of where demand lives.
Google Updated Its AI Content Guidance. Don't Expect Your Traffic Back.
Better content can win the ranking and still lose the click.
What's the News:
Google updated its content guidance to say publishers using generative AI should manually fact-check and review that content before publishing, including titles, meta descriptions, structured data, and image alt text. It also reiterated its rules against scaled content that adds little value. What it still won’t do, as MarTech points out, is promise that following the guidance will recover lost traffic. Google warns that improvements can take time and may produce no noticeable change.
Why It Matters:
Ranking recovery isn’t traffic recovery: A page can improve in Google’s eyes and still lose the click, because AI Overviews, forums, and video now answer more of the query before anyone reaches your site. Ahrefs found AI Overviews cut the click-through rate for the top-ranking page by 58% across 300,000 keywords.
There’s no update to wait on: Google folded its helpful content signals into its core ranking systems, so there is no single update to watch to find out whether your changes worked.
Google is telling us the economics of content have changed, and most content programs haven’t caught up. When AI can publish 500 pages as easily as five, a page that restates what 50 others already say gives Google no reason to send anyone to it.
Marketers Should:
Invest in what a summary can’t replicate: original data, real testing, and direct experience.
Put a real editor on every AI-assisted page, and extend that review to titles, meta descriptions, structured data, and alt text.
Stop judging content by rankings alone. If success is defined as position, you will keep winning on paper while losing the business outcomes that matter.
Ask one question of every content program: Does it earn demand that survives a world where Google answers first?
Did Your Ads Work, or Just Take the Credit?
Google made Conversion Lift self-serve for Search and PMax. It's a step forward, not the gold standard.
What's the News:
Eligible advertisers can now run user-based Conversion Lift studies across Search and Performance Max without working through a Google representative. The studies create a holdout group that isn’t exposed to the campaigns being measured, then compare its conversion behavior against users who can see the ads. The thresholds are still meaningful: at least 1,000 observed conversions and a $5,000 minimum campaign budget.
Why It Matters:
Efficiency isn’t incrementality: Search and Performance Max are built to find conversions efficiently. But branded search, existing demand, and users already likely to convert can make campaign performance look stronger than the true business impact.
A new way to check the work: Conversion Lift gives marketers another way to test whether campaigns are creating additional conversions, not just receiving credit for them.
Marketers Should:
Keep independent incrementality testing as the gold standard. It gives the clearest view of whether advertising actually caused additional business outcomes.
Use Conversion Lift when independent testing isn’t possible. Some brands don’t have enough scale, clean enough geographies, or the right operating conditions to run a strong incrementality test. In those cases, platform-based Conversion Lift is a valuable secondary option.
Choose it over attribution. Even a platform-run lift study is a better read on what Search and PMax actually create than attribution alone.
Start where credit and impact are most likely to diverge: branded Search and Performance Max.
ChatGPT Ads Just Became a Media Plan Line Item.
Now that ChatGPT inventory runs through a DSP, testing it is a planning decision, not a platform bet.
What's the News:
ChatGPT ads are now buyable through a programmatic platform. In September, Amazon launched a pilot that lets select U.S. advertisers extend Amazon DSP campaigns into ChatGPT. It follows OpenAI naming Criteo its first ad tech partner and early talks with The Trade Desk about selling ChatGPT inventory, a sign OpenAI is plugging into existing ad infrastructure rather than building a walled garden like Meta or Google.
Why It Matters:
Easier to buy means real competition for budget: When ChatGPT ads can be bought through existing programmatic pipes, they become much easier to buy at scale and compete for the same budgets that go to search and social. The channel is already past the experiment stage: ChatGPT ads reached a $1 billion annualized run rate.
Familiar partners lower the risk: Names like Amazon, Criteo, and TTD make a brand-new ad format feel less risky to cautious advertisers. eMarketer’s advice is to test AI placements through partners that plug into existing buying tools instead of waiting for standalone AI platforms to mature.
When ChatGPT inventory shows up in the DSPs we already use, testing it becomes a media planning decision instead of a new platform onboarding.
Marketers Should:
Check your access. If you already buy through Amazon DSP, ask whether your account is eligible for the ChatGPT pilot.
Run a small, controlled test with measurement defined before launch.
Make it incrementality-based wherever possible, so you can tell whether ChatGPT adds reach and intent or just pulls budget from search and social.
We’re reinventing performance marketing because everything changed while you were reading this.
We help you navigate this new landscape with data-driven insights, platform expertise, and creative that connects, so your brand not only keeps up, but gets ahead.
Work with Us