Google Answers More, Links Less: The Week in Marketing

Three things genuinely mattered in marketing this week. Google moved further from linking to answering — with queries in AI Mode running about three times longer than traditional searches and Google testing interfaces it generates itself. B2B (business-to-business) marketers admitted in survey after survey that they don’t trust the data shaping their own budgets. And brands started buying creators on fit rather than follower count.

Put together, the week says one thing: the proxies marketers have leaned on for a decade — clicks, dashboard numbers, audience size — are drifting away from the thing they were supposed to measure.

Search is turning into an interface, not an index

Greg Jarboe’s piece in Search Engine Journal is the most actionable data point of the week: queries in Google’s AI Mode are roughly three times longer than conventional search queries. People aren’t typing “crm pricing” anymore. They’re typing something closer to a sentence they’d say to a colleague.

That changes what a good page looks like. A long, conversational query carries its own context — budget, industry, constraint, intent — and the machine reading your page is matching against all of it. Pages that bury the answer under 400 words of context-setting get skipped, because the model has already been handed the context.

The same publication’s SEO Pulse roundup covered a spam update, generative user interfaces appearing inside AI Overviews, and a visibility drop for Reddit. The Reddit item is worth sitting with: it’s a reminder that borrowed visibility on someone else’s platform is a rented asset, and rents get raised.

The part that should worry anyone with a calculator page

Matt Southern’s follow-up on Google’s generated interfaces is the quieter, sharper story. If Google can generate a working mortgage calculator, unit converter or comparison table directly in the results, then a large category of pages — the free-tool page built purely to catch search traffic — loses its reason to exist.

Plenty of Indian SaaS and fintech brands run exactly this play: an EMI calculator, a GST calculator, a salary-to-in-hand tool, all built as top-of-funnel magnets. Those pages still work today. The strategic question is whether they’re a durable asset or a depreciating one, and this week’s news pushes the answer toward depreciating.

The counterweight came from Duane Forrester in a Search Engine Land interview, whose framing was that “AI is rewriting SEO, but the mission is the same.” That’s not comfort-food optimism. It’s a scoping note: the delivery mechanism changed, the job — be the most useful, most credible source on a question — did not.

Nobody trusts the numbers on the budget slide

The second thread is measurement, and it was ugly. MarTech reported that B2B marketers broadly distrust the data used to set their budgets — which is a remarkable thing to admit about the single most consequential number in the department.

MarTech also published a structural explanation for part of it: long sales cycles break attribution by design. If a deal takes nine to eighteen months to close — normal for enterprise software, common for Indian B2B firms selling into the US — then the campaign that gets credit is whichever one happened to run last, and the campaign that actually created demand has long since fallen out of the reporting window.

The engagement illusion

The third piece completes the picture. MarTech’s “engagement illusion” argues that the metrics we celebrate — likes, opens, impressions, time-on-page — frequently measure incidental behaviour rather than actual attention.

Read the three together and a pattern appears. Distrusted budget data, attribution that can’t span a real sales cycle, and engagement metrics that don’t track attention are not three separate problems. They’re one problem: our measurement layer was built for short, clickable, single-session journeys, and buyer behaviour stopped looking like that.

Creator fit beats follower count

The fourth thread is the most immediately usable. Marketing Dive reported that audience-brand fit outperforms raw follower count for brand outcomes — a finding that’s been folk wisdom for years and now has numbers behind it.

Search Engine Land ran a useful companion on measuring the true value of creators and review content, which matters more than it sounds. Creator and review content is increasingly what AI systems read when they decide which brands to name in an answer. A well-matched creator isn’t just a media buy anymore; it’s a source that language models cite.

Practically: if your quarterly creator budget is ₹5 lakh, this week’s evidence argues for spreading it across eight genuinely well-matched mid-tier creators rather than concentrating it on one large account whose audience overlaps your buyer by 12%. And Ann Taylor’s comeback push — built partly on Substack — shows the same instinct at the channel level: go where a defined, self-selected audience already is.

The connecting shift

Machines now sit between you and your audience on both ends. They summarise your content before a human reads it, and they mediate the data you use to justify spend. In both cases, the intermediate layer — the click, the impression, the last-touch row in the dashboard — is losing its meaning as a signal.

MarTech’s look at teams getting the most from AI lands on the same point from the org-design side: the winners changed process and ownership, not just tooling. That’s consistent with Search Engine Land’s argument that visibility in large language models starts with internal communication — because getting cited requires product, support and marketing to describe the company the same way.

Thread	Underlying shift	Your move
3x longer AI queries	Context is supplied by the user, not your intro	Answer in the first 60 words
Generated interfaces	Tool pages lose their moat	Treat tool traffic as depreciating; build owned demand
Distrusted budget data	Attribution can’t span long cycles	Add a pipeline-influence view alongside last-touch
Creator fit over reach	Overlap beats volume; creators feed AI answers	Buy on audience overlap, not follower count

What this means for you

• Rewrite your top 10 pages answer-first. Put the direct answer in the opening paragraph, then explain. This is now a search requirement, not a style preference.
• Audit your tool pages this quarter. List every calculator or converter built for traffic. Ask what happens to your pipeline if each one loses 60% of its sessions in 12 months.
• Stop defending budget with last-touch alone. Bring one pipeline-influence metric to your next budget meeting — deals touched, not deals closed-by. It survives a nine-month cycle; last-touch doesn’t.
• Change one creator brief. Ask for audience-overlap evidence, not a follower screenshot. Fit is the variable that moves outcomes.
• Write one internal description of your product that sales, support and marketing all use verbatim. Inconsistent self-description is a visibility problem in AI answers.

Frequently asked questions

Why are AI Mode queries longer than normal searches?

Because people phrase them like questions to a person rather than keywords to a machine. Search Engine Journal reported that AI Mode queries run roughly three times longer than traditional queries, which means the user has already supplied the context your page used to have to guess at.

Should I delete my free calculator and tool pages?

No — not yet. They still earn traffic today. But treat them as a depreciating asset: stop building new ones purely for search capture, and shift that effort toward content and product experiences Google cannot regenerate inside a results page.

Why don’t B2B marketers trust their own budget data?

Mainly because attribution models assume short buying journeys. When a deal takes a year to close, the touchpoint that created demand is outside the reporting window, so credit lands on whatever ran most recently. Marketers know the number is wrong even when they can’t fix it.

How should I choose creators if follower count doesn’t matter?

Choose on audience overlap with your actual buyer, evidence of genuine category interest, and whether the creator’s content gets referenced elsewhere. A mid-tier creator with 70% buyer overlap beats a large account with 10% overlap on almost every brand outcome.