Search & AI Visibility
Find out why nobody is finding it, before rewriting it.
Almost every site that is not ranking has already been rewritten once, on the theory that the content was the problem. Usually it was not. The pattern we find is mechanical: pages a crawler cannot render, four hundred products sharing one title, a canonical tag pointing at a staging origin, a migration that dropped its redirects. We measure before recommending, we say which findings are worth money and which are hygiene, and the output is a prioritized list with the fixes applied, not a monthly report.
- LCP budget the templates we touch are held to
- <2.5s
- Orphaned pages left in the templates we own
- 0
- Vitals scored on real users, never a lab number
- Field
Sounds like
You might recognize one of these.
We rebuilt the site and traffic fell off a cliff the week it launched.
Google Search Console says the pages are discovered but not indexed, and we don’t know what that means.
We rank for our own name and nothing else.
Impressions are up, clicks are flat, and nobody can tell us why.
What this includes
The work, specifically.
Not every engagement needs all of it. This is the range we cover and what each part is actually for.
Crawl, render and index diagnosis
What a crawler actually receives, which is not always what your browser draws. Client-rendered content, blocked resources, soft 404s, redirect chains, parameter explosions, and the difference between discovered, crawled and indexed in Search Console, which is where most of the answer lives.
Duplication and canonical integrity
Templated titles and descriptions, near-identical pages competing with each other, staging and print origins that got indexed, and canonicals that point somewhere the page does not agree with. This is the single most common finding on a site that was migrated.
Query, intent and content mapping
Which queries actually exist, what the person typing them wants, and whether you have a page that answers it. Relevance and depth decide more here than any technical item on this list, which is why this is a mapping exercise against real demand rather than a keyword count: one page per intent, no two pages competing for the same one, and an honest note where the demand is not there.
Information architecture and internal linking
The pages that matter should be reachable in a few clicks from somewhere with authority. We map what links to what, find the orphans and the dead ends, and fix the structure rather than adding a keyword to the ones that are already fine.
Structured data that matches the page
Schema.org markup derived from what the page actually shows, not asserted alongside it. We also say what is retired: FAQ and HowTo rich results are gone from general results, and markup that only ever existed to chase them is code to delete, not to expand.
Core Web Vitals as an engineering budget
LCP, INP and CLS measured on field data rather than a lab score, tied to the specific template and third-party script causing them, and enforced in CI so the fix does not decay. Page experience is a real but small factor, and we will tell you when it is not your problem.
AI answer surfaces and citation
AI Overviews and assistant answers broke the chain from ranking to click-through that most traffic alarms still assume, so a fall in sessions against flat impressions is now a normal reading rather than a penalty. We instrument for it, and we make the site legible to a retriever: clean entity markup, machine-readable corpora, and claims that carry a source a model can check.
Local and map pack
For a business with an address or a service area: Google Business Profile completeness, category and service accuracy, citation consistency, and the on-site signals the map pack reads. Frequently the highest-return work available to a regional firm, and frequently the least interesting to the agency selling content.
What you get
Deliverables, not documents.
- Findings report, scored by impact and effort, with the evidence behind each item
- Applied fixes for everything mechanical, as reviewed pull requests against your repository
- Crawl, render and index baseline you can re-run
- Structured data emitted from content, validated against the rendered page
- Core Web Vitals field baseline and a CI budget on the templates that matter
- Redirect map and a launch checklist, for anything about to migrate
- Measurement that survives AI answer surfaces: what to watch, and what no longer means what it used to
Shapes
How this usually runs.
Diagnostic
1–2 weeksA full technical, content and authority pass, ending in a prioritized findings report and an estimate for the repairs. If the answer is that the site is fine and the problem is that nobody is looking for what it sells, that is what the report says.
Remediation
3–8 weeksWe implement the findings. Template and rendering changes, canonical and redirect repair, structured data, performance budgets, and the internal linking work, shipped as reviewed changes to your codebase.
Migration cover
Around a launchThe engagement to buy before a redesign or a replatform rather than after. Redirect mapping, parity checking against the old URL set, and a monitored first month, because the traffic loss people call us about is nearly always something that happened at cutover.
Tooling
What we build it with.
No tool here was picked because it was new. Where we do reach for something novel, it is in one place, for a stated reason, and it is written down.
- Diagnosis
- Performance
- Markup & feeds
- Measurement
Questions
Search visibility, honestly.
No, and neither can anyone else. What we can do is remove the mechanical reasons a page cannot rank, make the site legible to crawlers and retrievers, and tell you honestly whether the demand you are chasing exists. Anyone quoting you a position for a keyword is selling either luck or something that will earn a manual action.
Not the kind you are probably asking about. Paid links, private blog networks and reciprocal schemes are against Google’s own spam policies and the downside is uncapped. We work on the things that legitimately earn references: publishable material, digital PR, and fixing the reasons the links you already have point at pages that no longer resolve.
Usually not. Rank it against the likelier causes first: a migration that lost redirects, a core update, a seasonal pattern, or AI answer surfaces absorbing clicks that used to arrive on informational queries. Impressions holding steady while clicks fall is the signature of the last one, and it is not something a rewrite fixes. We separate those before anyone starts editing pages.
We will map what needs to exist and edit what is there, and we will point out where AI-generated bulk is already hurting you. Publishing volume without review is the most reliable way to look like the thing the spam policies were written for, and it is why a lot of sites that ran that play in the last two years are now in a hole.
Now. A redesign is the single highest-risk event in a site’s search history, and the cheapest version of this work is the one that maps the redirects before the URLs change rather than reconstructing them from a traffic graph afterward.
Sources
- 1.Google Search Essentials (opens in a new tab), Google,
- 2.Web Vitals (opens in a new tab), Google (web.dev),
Next step
Tell us what’s breaking.
Forty-five minutes, no charge, no deck. We’ll tell you what we’d do, what it would likely cost, and whether what you already have can be made to work.