Publishing volume stopped working around the time assistants started answering the question outright. We produce fewer, better pieces — planned from real prompts, structured to be quoted, and built to earn citations rather than sessions.
The volume playbook — publish consistently, target keywords, build topical authority through sheer output — worked when the reward for ranking was a click. Increasingly the reward for ranking is having your paragraph summarised above the results while the reader never arrives.
That doesn't make content worthless. It changes what good looks like. A piece that gets quoted and attributed in an AI answer puts your brand in front of a buyer at the exact moment of research, which is arguably better placement than a blue link. But it only happens if the piece is structured so a clean passage can be lifted from it.
Most content isn't. We audit blogs constantly and the pattern is consistent: dozens of decent articles where the actual answer sits in paragraph six, depends on context established earlier, and is phrased in language nobody would ever search. Fixing that is usually worth more than another year of publishing.
Almost every engagement starts with a triage decision, and the answer is usually less new content than clients expect.
Best for: Any site with a real content archive. Most have 30–60 pieces worth restructuring immediately.
Best for: Brands whose prompt benchmarking shows whole clusters with no corresponding coverage at all.
Four properties. Content quality in the conventional sense is necessary but nowhere near sufficient.
The answer appears in the first sentence or two of its section, not after three paragraphs of context-setting the model has to wade through.
Each section survives being quoted alone. Passages that only make sense in sequence get skipped, because extracting them would distort the meaning.
Real numbers, named examples, documented methodology. Generic claims that every competitor also makes give a model no reason to pick you as the source.
A named author with real credentials, marked up properly. Anonymous content is discounted heavily, particularly in commercial and sensitive categories.
Strategy, briefs, production or editing, and the post-publish testing most agencies skip entirely.
Scoring every existing piece for extraction readiness, ranking value, and cannibalisation, then sorting into restructure, retire, or leave alone.
Building the plan from benchmarked conversational questions rather than keyword volume, since the two overlap far less than you'd expect.
Every piece briefed with the question it answers, where the answer must sit, required headings, terms to define, and schema to apply.
We write, or we brief and edit while your team writes. Where you have genuine in-house expertise, the second usually produces better content.
Building named author entities with real credentials in markup, so content is attributed to a credible source rather than read as anonymous.
Testing after publication whether pieces actually get pulled into answers, and restructuring the ones that don't rather than moving on.
Which clusters matter is entirely dependent on how your buyers research.
Comparison, alternatives, integration, and pricing content dominate. These get asked constantly and most vendors write them badly.
Stricter source standards apply. Credentialed authorship and documented editorial process largely determine whether you're eligible for citation.
Buying guides, sizing, materials, and comparison content. Models cite sources that help someone decide, not sources that only list products.
Methodology documentation and original research. Service categories are won on demonstrated expertise, not published volume.
Planning is front-loaded, then production and testing run continuously against the plan.
We score your existing content for extraction readiness and benchmark the real questions your category gets asked conversationally.
We identify which clusters you cover, which you cover badly, which competitors own, and which are honestly not worth contesting.
Archive restructuring first because it's cheapest, then new pieces produced against explicit structural briefs.
Monthly extraction testing on published pieces, with the ones that aren't being pulled getting another structural pass.
Leading with the answer, defining terms, and using headings that match real questions makes content easier to skim, easier to rank, and easier to extract. There's no tradeoff here — the same discipline serves all three.
The questions clients ask us most before starting. If yours isn't here, ask us directly on a consultation call.
Two things. We plan from benchmarked conversational prompts rather than keyword volume, and we hold every piece to explicit structural requirements so a clean answer can be extracted from it.
We also usually recommend publishing less than you currently do, and spending the difference on restructuring what you have.
Less than most agencies will sell you. For most B2B and SaaS clients, two to four genuinely substantive pieces a month plus an archive restructuring programme outperforms eight thin ones.
The constraint is rarely volume. It's whether each piece contains something specific enough that a model has a reason to cite you rather than anyone else.
Either, and the honest answer is that briefs plus your in-house experts often beats us writing everything — because specificity is what gets cited and your team has it.
Where you don't have writing capacity we produce, working from interviews with your subject-matter experts rather than desk research.
It isn't automatically penalised, but it tends to be structurally competent and factually generic, which is exactly what doesn't get cited. Models favour sources with specific, verifiable, original information.
We use AI tooling in production like any competent team. The differentiating substance — your data, your numbers, your methodology — has to come from you.
It gets audited and triaged. Typically a third is worth restructuring, a third can be left alone, and a third is thin enough that consolidating or retiring it improves the site overall.
Pruning is genuinely part of the work. A large archive of unextractable content dilutes topical clarity rather than building it.
Extraction and citation, alongside classic signals. We track which specific pieces get pulled into AI answers, which win snippets, and which rank — then report all three rather than defending one number.
Sessions still matter, but a piece cited in AI answers without a click is doing real work, and reporting only on traffic misses it.
Restructured archive pieces typically move in four to eight weeks. New pieces take longer to earn authority — usually two to four months for competitive clusters.
We sequence archive work first specifically so there's visible movement while newer content matures.
It's normally delivered inside a retainer, typically $4,000–$12,000 monthly depending on production volume and whether we're writing or briefing.
Strategy and archive audit can be scoped as a standalone fixed-fee project. Detail is on our pricing page.
We'll score every piece on your blog for extraction readiness and show you how many are one restructure away from being cited.
No commitment · 45 minutes · Immediate value