TL;DR
- Every 2026 guide now says the same thing: measure pipeline, shift budget to bottom of funnel, informational content has stopped paying. The first is right. The second is right. The third is wrong, and acting on it will cost you.
- Informational content did lose its traffic. Ahrefs found 99.2% of AI-Overview-triggering keywords are informational, with position-one clicks down 34.5% in April 2025 and 58% in a February 2026 re-run.
- But referring domains were the strongest measured predictor of AI citations in SE Ranking's analysis of 216,000 pages. Links come from informational content. Nobody links to a pricing page.
- Authority is the binding constraint. Only 1.74% of new pages reach Google's top 10 within a year, down from 5.7% in 2017, and 72.9% of top-10 pages are over three years old (Ahrefs, May 2025).
- So do not cut it. Re-measure it. Judge informational content on referring domains and citations, not sessions, and keep funding it as the engine that makes commercial pages competitive.
What is B2B SEO and how is it different?
B2B SEO is the practice of earning search visibility for the terms a buying committee uses across a months-long purchase decision, measured in qualified pipeline rather than in traffic.
Three structural differences from consumer SEO:
- Volumes are small. A term with 50 monthly searches can be worth more than one with 5,000, because the fifty are all buyers. Our guide to B2B keyword research gives the formula for pricing that difference.
- Several people research the same purchase. A technical evaluator, an economic buyer and an end user search differently for the same decision, and one page rarely serves all three.
- The gap between visit and revenue is months. Anything you measure weekly is a leading indicator, not a result.
One term worth defining, because it is doing all the work in 2026 commentary: search intent is what the searcher wants. Informational means they want to learn. Commercial means they are comparing options. Transactional means they are ready to buy. The rest of this article is largely about how to divide a budget across those three.
What actually changed in 2026?
The informational half of the funnel stopped producing clicks. That is the whole change, and everything else follows from it.
Ahrefs analysed which keywords trigger Google's AI Overviews and found 99.2% of them are informational in intent. It separately measured click-through on keywords with and without an AI Overview across 300,000 keywords: the top-ranking page received 34.5% fewer clicks when one was present (April 2025), and 58% fewer when the same method was re-run on December 2025 data (February 2026).
99.2% and 58%: the share of AI Overview keywords that are informational, and the position-one click reduction when one appears (Ahrefs, 2025 and 2026)
Pew Research Center measured the same effect in real browsing: users clicked a result on 8% of searches showing an AI summary against 15% without one, and clicked a source cited inside the summary on 1% of visits (July 2025).
What did not change: commercial and transactional keywords. They rarely trigger AI Overviews, so their click rates hold up. That asymmetry is the entire basis of the 2026 consensus, and the consensus draws the wrong conclusion from it.
Why is "shift budget to bottom of funnel" only half right?
Because it treats informational content as a traffic channel that has stopped working, when it is actually an authority channel that never was a traffic channel in the first place.
Every guide competing for this term now recommends moving production toward comparison, pricing, alternatives and case study content. That part is correct. Those pages convert, they escape the AI Overview discount, and most B2B sites underinvest in them.
The error is in what gets cut to pay for it.
SE Ranking analysed 216,000 pages and found referring domains were the single strongest predictor of AI citations. Referring domains are other websites linking to you. Now ask what earns those links. Nobody links to a pricing page, a comparison page or a service page. People link to research, data, guides and explanations, which is to say informational content.
And authority is the binding constraint on everything else. Ahrefs found that only 1.74% of newly published pages reach Google's top 10 within a year, down from 5.7% in 2017, and that 72.9% of top-10 pages are now more than three years old, up from 59%. New commercial pages on a low-authority domain do not rank. They sit at position 40 producing nothing, regardless of how well they convert the traffic they never receive.
So the sequence matters more than the split. Informational content earns the links. The links raise the domain's authority. The authority is what lets the commercial pages rank. Cut the first and the third quietly stops working, on a delay long enough that nobody connects the two.
The correct instruction is not "cut informational content". It is "stop measuring informational content on sessions". Judge it on referring domains earned and citations won. Those are the outputs it was always actually producing, and they are the ones that still work.
How should the keyword portfolio be split?
By what each type is for, rather than by a fixed ratio.
Published 2026 guidance commonly recommends 60% to 70% of production going to mid and bottom funnel. That is a reasonable starting point and a poor rule, because it prescribes an output ratio without saying what the remaining 30% to 40% is for.
Here is the split with a purpose attached to each part.
| Share of production | What it is for | How to judge it | |
|---|---|---|---|
| Commercial and transactional | 50% to 60% | Capturing demand from buyers already evaluating | Enquiries, pipeline, close rate |
| Informational, link-earning | 25% to 30% | Earning referring domains and citations that raise domain authority | Referring domains, citations, not sessions |
| Informational, supporting | 10% to 20% | Covering the topic so the hub reads as complete | Rankings on long-tail, internal link value |
Worked example: what the split costs and returns
Illustrative. Substitute your own figures.
A team publishing 10 articles a month at an internal cost of $600 each, so $6,000 a month.
| Allocation | Articles | Monthly cost | Judged on |
|---|---|---|---|
| Commercial and transactional | 6 | $3,600 | Enquiries |
| Informational, link-earning | 3 | $1,800 | Referring domains |
| Informational, supporting | 1 | $600 | Coverage |
The mistake the 2026 consensus produces: move to 10 commercial articles, save nothing, and lose the three pieces a quarter that were earning links. Domain authority stops rising. Eighteen months later the commercial pages stop climbing and nobody can explain why, because the cause was removed a year and a half earlier.
The cheapest correction available: keep the three link-earning pieces and change the report they appear on. They are not underperforming traffic articles. They are performing authority articles being measured with the wrong instrument.
How do you choose which keywords to fund?
By expected pipeline rather than by volume. The full method, including the formula and the AI Overview adjustment, is in our B2B keyword research guide, and this section is the summary.
Price every keyword as: volume × click rate × AI Overview adjustment × visit-to-lead rate × lead-to-close rate × deal value.
Two things that method surfaces which volume never does:
Small commercial keywords outrank large informational ones. In that guide's worked example, a 50-search commercial keyword produced more expected pipeline than a 5,000-search informational one, and the AI Overview adjustment is what flipped the result.
Your highest-intent terms are frequently invisible to tools. They show zero volume because the true figure is below the tool's reporting threshold. They come from sales call recordings, your site's internal search box and support tickets, not from a keyword tool.
Apply one exception to the pricing rule. Link-earning informational content should not be selected on expected pipeline, because pipeline is not what it produces. Select it on link potential: original data, a genuinely useful tool, a contrarian argument with evidence, or a resource in a category where nothing good exists.
What content does each funnel stage need?
| What the buyer is doing | Content that works | Primary metric | |
|---|---|---|---|
| Problem aware | Naming a problem they have | Explainers, original research, data | Referring domains |
| Solution aware | Comparing approaches | Framework guides, category comparisons | Rankings, assisted conversions |
| Vendor aware | Comparing suppliers | Comparison pages, alternatives pages, pricing | Enquiries |
| Decision | Building a case internally | Case studies, ROI calculators, implementation detail | Close rate |
One row is worth arguing about. The vendor-aware row includes alternatives and comparison pages naming competitors. Many B2B companies avoid these on the reasoning that they would rather not write about rivals.
That reasoning cedes the page to the rival. Buyers compare regardless. The only question is whether they do it on your page or on someone else's. The honest version, which is harder to write and worth more, names the cases where the competitor is the better fit. A comparison with no honest tradeoff reads as an advertisement and gets discounted by readers and by AI systems alike.
Where this is the weaker approach: if you cannot write a comparison that concedes anything, do not publish one. A dishonest comparison damages more than an absent one, and it is easy to spot.
How should the site be structured?
In hubs, with a broad pillar page and several deep cluster pages linking to each other in both directions. The full method is in our guide to pillar and cluster content.
The failure mode is sequencing, and we have measured it on ourselves. Auditing a 189-article programme before launch, 42 of 173 cluster articles, 24%, were scheduled to publish before the pillar they were supposed to support, affecting 12 of 15 hubs, with a largest gap of nine months.
24% of clusters scheduled ahead of their pillar, across a 189-article plan (Euracle content plan audit, 2026)
That happens when a calendar is sorted by keyword opportunity rather than by structure, which is the obvious way to build a content plan and the wrong one. A cluster published before its pillar has nowhere to send authority and no hub to inherit it.
The rule that prevents it: publish the pillar first, or in the same week as its first cluster, in every hub.
This article is an instance of the correction. It is the H07 pillar, it was scheduled for month 6, and it is being published first in its wave precisely because three clusters had already shipped without it.
How do you measure a B2B SEO programme?
On three separate clocks and two separate instruments, because measuring everything on sessions is what produced the mistake in section three.
Three clocks, covered fully in how long does SEO take:
| Clock | Duration | What you check |
|---|---|---|
| Time to know | 30 to 60 days | Impressions and indexed pages in Search Console |
| Time to rank | 3 to 12 months | Position for the target keyword |
| Time to revenue | Time to rank plus your sales cycle | Enquiries attributed to organic |
Two instruments, which is the part most programmes get wrong:
- Commercial and transactional pages are judged on enquiries, pipeline and close rate.
- Link-earning informational pages are judged on referring domains gained and citations won. Not sessions. Not conversions. If you report them on the same dashboard as your pricing page, they will look like failures and get cut.
And one measurement most B2B teams do not have at all. A share of your buying committee now researches inside AI answers before visiting any website. Build a fixed set of 20 to 50 buyer questions, run them monthly across the assistants your buyers use, and record which brands appear. That method is set out in our generative engine optimization guide.
What do most B2B teams get wrong?
Cutting informational content to fund commercial content. The central error of the 2026 consensus and the reason for this article. Re-measure it instead.
Measuring everything on sessions. It makes link-earning content look like a failure and commercial content look like a disappointment, when they are producing exactly what they should.
Chasing volume. A 50-search commercial term can outproduce a 5,000-search informational one, and the AI Overview adjustment usually decides it.
Publishing clusters before pillars. We did it on 24% of a 189-article plan. It is invisible on a dashboard and it stops the compounding.
Refusing to publish comparison content. Buyers compare anyway. The only choice is whose page they do it on.
Judging at six weeks. Rankings are a lagging indicator. Impressions are the leading one, and they appear far sooner.
Who pays for this: the marketing lead who approved a twelve-month budget, because every one of these failures is invisible on a monthly report until the compounding that justified the investment simply never begins.
How does Euracle run B2B SEO?
With two separate scorecards, because a single one produces the mistake this article is about.
The Eureka Method, Euracle's discovery sprint, runs four phases.
Discover prices the keyword set on expected pipeline using the client's own conversion and deal-value figures, and separately identifies the link-earning opportunities, which are selected on a different basis entirely.
Design orders the calendar by structure before opportunity, which is the correction that came out of the audit in section seven, and sets the two scorecards in writing.
Deploy ships each pillar with its first clusters and adds links in both directions on publication day.
Scale reports commercial pages on enquiries and link-earning pages on referring domains, monthly, to the same person.
The stack is conventional: Ahrefs and SEMrush for keyword and backlink data, Google Search Console and GA4 for first-party measurement, and n8n with the Claude API to run the monthly assistant prompt set on a schedule.
Two structural commitments come from how Euracle is set up. Senior practitioners only: the people in the pitch do the work, which matters most in the judgement about which informational pieces are worth funding for links. And one contract across six disciplines, so a finding that the constraint is a missing pricing page or a positioning problem rather than a content volume problem does not require a second vendor.
If you want the portfolio split and the two scorecards set on your own numbers before a calendar is built, that is Euracle's SEO service. The split is hardest for B2B SaaS companies, where a large existing content library was built entirely on the traffic assumption this article argues against.
FAQ
Conclusion
You can now act on the 2026 consensus without following it off a cliff. Shift production toward commercial and transactional pages, because they convert and they escape the AI Overview discount. But do not fund that shift by cutting informational content, because that content was never a traffic channel and cutting it removes the referring domains that let your commercial pages rank at all. Keep roughly a quarter of production on link-earning pieces, measure them on referring domains rather than sessions, publish pillars before clusters, and judge the whole programme on leading indicators at day sixty rather than rankings at day thirty. If you want the split and the two scorecards set on your own numbers first, talk to Euracle about SEO.
Sources
- Ahrefs, AI Overviews Reduce Clicks by 34.5%, April 2025, and its February 2026 update reporting 58% on December 2025 data. The 99.2% informational-intent finding. https://ahrefs.com/blog/ai-overviews-reduce-clicks/
- Ahrefs, How Long Does It Take to Rank in Google?, Patrick Stox, May 2025. The 1.74% figure and the 72.9% age distribution.
- Pew Research Center, 22 July 2025. Click rates of 8% with an AI summary against 15% without, and 1% on cited sources. https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/
- SE Ranking, analysis of 216,000 pages finding referring domains the strongest predictor of AI citations.
- Published 2026 guidance recommending 60% to 70% of production in mid and bottom funnel.
- Euracle content plan audit, 2026. 189 planned articles, 173 clusters, 42 scheduled before their pillar, 12 of 15 hubs affected, 9-month maximum gap.
