Writing

Shorter notes on SEO and AI search.

Each post answers one question, usually one that came up in day-to-day work. The longer guides are organised by topic.

21 October 202628 min read
The best local SEO tools in 2026, tested and compared
Eighteen local SEO tools compared on grid tracking, listings, reviews, AI search and what the same setup really costs, with prices checked on 21 October 2026.
Local
20 October 20269 min read
Set your holiday hours now: what Maps and AI assistants do with them
Google may warn searchers that your hours might differ on holidays you have not set, and Gemini reads opening hours from the same map data. Set them on every platform this week, while there is still time for them to publish.
Local
19 October 202610 min read
Staging sites, test pages and other things that get indexed by accident
Staging copies, preview deployments, test pages and stray PDFs end up in search indexes and in front of AI crawlers more often than anyone intends. This is how they get found and how to get them out for good.
Technical
18 October 202610 min read
Does schema markup help you get cited by AI assistants?
It depends on how the assistant reaches your page. Search indexes parse markup, live fetchers mostly throw it away, and nobody outside the AI labs knows what training does with it.
Entities
17 October 20269 min read
Review incentives and review gating: what the rules allow
Asking customers for reviews is allowed almost everywhere. The lines are drawn around rewards, staff reviews and choosing who gets asked, and UK law, US law and the review platforms each draw them in slightly different places.
Local
16 October 20269 min read
"Crawled - currently not indexed": what it usually means and what to do
Google fetched these pages and decided not to index them for now. Most of them do not need fixing, and the ones that do are easiest to find by working through the report one template at a time.
Technical
15 October 202610 min read
When an AI assistant gets your business wrong, and how to find where the error came from
A wrong fact reaches an answer from the model's training, from a page it read, or from a listing it pulled. Each route has a different fix, so the first job is working out which one you are on.
AI search
14 October 20269 min read
How to move premises without losing your place in the map pack
Keep the same Google Business Profile, expect to verify it again, and plan for rankings to shift with the new address. This is the order I would run a move in, from six weeks before to two months after.
Local
12 October 20269 min read
Anthropic's crawlers: ClaudeBot, Claude-SearchBot and Claude-User explained
What each of Anthropic's three bots does, what blocking each one costs you, the retired tokens still sitting in old robots.txt files, and how to check the traffic is really Anthropic.
AI crawlers
12 October 202610 min read
Applebot and Applebot-Extended: Apple's crawlers explained
What Applebot feeds across Spotlight, Siri, Safari and Apple Intelligence, how Applebot-Extended controls training without crawling, and the robots.txt, meta tag and log checks for each goal.
AI crawlers
12 October 202611 min read
Google's crawlers and AI: Googlebot, Google-Extended and the rest
Which Google crawler feeds AI Overviews, AI Mode, Gemini and Vertex AI, what each robots.txt token really controls, and which control to use for each business goal.
AI crawlers
12 October 20269 min read
Meta's crawlers: Meta-ExternalAgent, Meta-ExternalFetcher and link previews
The five crawlers Meta documents, which one builds the previews on Facebook, Instagram and Messenger, which ones feed Meta AI, and the robots.txt lines that block training without breaking sharing.
AI crawlers
12 October 20269 min read
Bingbot and Copilot: how Microsoft's crawler feeds AI answers
Microsoft has no separate AI crawler to block. Bingbot builds the index behind Bing, Copilot and Microsoft grounding, and the AI controls live in meta tags. Here is what each one does, how to verify Bingbot, and how to see your Copilot citations.
AI crawlers
12 October 202613 min read
OpenAI's crawlers: GPTBot, OAI-SearchBot and ChatGPT-User explained
What each OpenAI user agent does, the robots.txt lines for the common goals, how to verify the traffic, and how ChatGPT's signed cloud browser identifies itself.
AI crawlers
12 October 202610 min read
The other AI crawlers in your logs: CCBot, Amazonbot, Bytespider and more
What Common Crawl, Amazon, ByteDance, Mistral, DuckDuckGo and You.com document about their crawlers, what each is for, and which ones I would block.
AI crawlers
12 October 20269 min read
Perplexity's crawlers: PerplexityBot and Perplexity-User explained
What each Perplexity user agent does, which one follows robots.txt, how to verify them against the published IP lists, and what the 2025 dispute with Cloudflare means for your settings.
AI crawlers
8 October 20269 min read
Why latency matters for AI search
Assistants answer from what they already know whenever they can, because searching costs time and money. When they do fetch a page, they will not wait long for it.
Technical
6 October 20269 min read
Migrating to or from Shopify without losing search traffic
Shopify fixes your URL prefixes, limits what its redirects can catch and adds its own canonicals and hreflang, so a move in either direction needs a plan built around those rules.
Migrations
1 October 202611 min read
How to review machine-translated and localised content before it goes live
What Google's spam policies say about translated content at scale, what to localise beyond the words, and a tiered review model that puts human attention where each page is worth it.
International
24 September 202611 min read
Product data that AI shopping features and agents can use
Where Google's AI Mode, ChatGPT and shopping agents get product data, which fields carry the weight, and a checklist for making the feed, the markup and the page agree.
Ecommerce
21 September 20268 min read
Judge backlinks by traffic and leads, not DR
Links still count for rankings and for AI answers. The useful test of a link is whether it sends people who become customers, and whether the page it sits on has readers of its own.
Links
17 September 202610 min read
How to check what crawlers see on a JavaScript site
Googlebot renders JavaScript later, most AI crawlers never do, and a few commands show you exactly which parts of your SaaS site each of them can read.
SaaS
10 September 202611 min read
When your CDN blocks AI crawlers, whatever robots.txt says
How bot management at the edge decides what AI crawlers and Googlebot receive, how to check a crawler is real, and how to test what each one gets.
AI crawlers
4 September 20269 min read
Topical authority is consistency
A site earns authority on a subject by covering it, in the same place and the same way, for a long time. The same expectation is why migrations hurt.
Content
3 September 202612 min read
Location pages that are not doorway pages
Where the search engines draw the line on doorways, what a location page needs to earn its place, how LocalBusiness markup ties it to the Business Profile, and how to audit a set of hundreds.
Local
29 August 202615 min read
How to measure visibility in AI search without fooling yourself
Build a small prompt set from your services and seed keywords, measure the customer asking cold, record every answer as structured data, and read pooled rates, not positions.
AI search
26 August 20267 min read
How to set up robots.txt for AI crawlers
Three worked configurations for three business goals, the parsing rules that catch people out, and how to confirm in your logs that it works.
AI crawlers
25 August 20269 min read
How an assistant decides which businesses to name
A working model of what happens between the question and the three names that come back.
AI search
19 August 20267 min read
How to measure SaaS SEO in signups and pipeline
Setting up key events, carrying organic landing pages into the CRM, separating branded from non-branded demand, and knowing where Search Console data stops.
SaaS
13 August 20268 min read
Who llms.txt actually helps
It does little for training and nothing measurable for being cited in AI answers. It earns its place with agents working inside JavaScript-heavy apps and developer documentation.
AI search
12 August 20267 min read
What AI search actually changes for a local business
Less than the headlines suggest, and more than most owners have planned for.
AI search
5 August 20266 min read
Choosing a URL structure for a new market: three worked examples
How to decide between a country domain, a subdomain and a subfolder, worked through for a shop, a software company and a regulated business.
International
1 August 20268 min read
Product variants: URLs, canonicals, ProductGroup markup and Merchant Center
How to decide whether variants share one page or get their own, and how to make the canonicals, the structured data and the Merchant Center feed all describe the same setup.
Ecommerce
29 July 20268 min read
What to check in the first month after a site migration
A day one, week one and month one schedule for checking a migration, with the Search Console reports and log signals that separate normal movement from a fault.
Migrations
22 July 20266 min read
The first five things I check in a local SEO audit
Before the crawl, before the spreadsheet, before anyone mentions keywords.
Local
15 July 20267 min read
How to audit hreflang and find the mistakes that quietly break it
The hreflang errors that let the wrong country version rank, why nothing warns you about them, and a step-by-step audit to find them.
International
11 July 20269 min read
How to decide which faceted navigation pages to index
A worked method for sorting filter combinations into the few that get their own indexable page and the many that crawlers should never see, using search demand, product depth and duplication.
Ecommerce
8 July 20268 min read
How to write comparison and alternatives pages that buyers trust
A working structure for "vs" and "alternatives" pages, how to keep competitor facts current, what Google's reviews guidance implies, and where the legal lines sit.
SaaS
1 July 20268 min read
How to build a redirect map for a site migration
A step-by-step method for listing every old URL, matching each one to the right new address, and testing the whole map before launch day.
Migrations
30 June 20265 min read
Your Google Business Profile is doing more work than your website
For a lot of local businesses, the profile is the first impression, the shop window, and the enquiry form all at once.
Local
9 June 20267 min read
Why group reporting hides the branches that are failing
A blended average across forty sites is arithmetic. It is not information about any of them.
Local
14 May 20269 min read
How the map pack is put together
Relevance, distance and prominence, and what each of those actually means when you sit down to work on it.
Local
2 April 20266 min read
Most local businesses do not need more content
The blog is rarely the bottleneck, and publishing into an unfixed foundation is an expensive way to feel productive.
Local
18 February 20267 min read
Service area businesses, and when to be one
What the setting actually does, who it is for, and the mistake that gets profiles suspended.
Local

Shorter notes on SEO and AI search.

The best local SEO tools in 2026, tested and compared

Set your holiday hours now: what Maps and AI assistants do with them

Staging sites, test pages and other things that get indexed by accident

Does schema markup help you get cited by AI assistants?

Review incentives and review gating: what the rules allow

"Crawled - currently not indexed": what it usually means and what to do

When an AI assistant gets your business wrong, and how to find where the error came from

How to move premises without losing your place in the map pack

Anthropic's crawlers: ClaudeBot, Claude-SearchBot and Claude-User explained

Applebot and Applebot-Extended: Apple's crawlers explained

Google's crawlers and AI: Googlebot, Google-Extended and the rest

Meta's crawlers: Meta-ExternalAgent, Meta-ExternalFetcher and link previews

Bingbot and Copilot: how Microsoft's crawler feeds AI answers

OpenAI's crawlers: GPTBot, OAI-SearchBot and ChatGPT-User explained

The other AI crawlers in your logs: CCBot, Amazonbot, Bytespider and more

Perplexity's crawlers: PerplexityBot and Perplexity-User explained

Why latency matters for AI search

Migrating to or from Shopify without losing search traffic

How to review machine-translated and localised content before it goes live

Product data that AI shopping features and agents can use

Judge backlinks by traffic and leads, not DR

How to check what crawlers see on a JavaScript site

When your CDN blocks AI crawlers, whatever robots.txt says

Topical authority is consistency

Location pages that are not doorway pages

How to measure visibility in AI search without fooling yourself

How to set up robots.txt for AI crawlers

How an assistant decides which businesses to name

How to measure SaaS SEO in signups and pipeline

Who llms.txt actually helps

What AI search actually changes for a local business

Choosing a URL structure for a new market: three worked examples

Product variants: URLs, canonicals, ProductGroup markup and Merchant Center

What to check in the first month after a site migration

The first five things I check in a local SEO audit

How to audit hreflang and find the mistakes that quietly break it

How to decide which faceted navigation pages to index

How to write comparison and alternatives pages that buyers trust

How to build a redirect map for a site migration

Your Google Business Profile is doing more work than your website

Why group reporting hides the branches that are failing

How the map pack is put together

Most local businesses do not need more content

Service area businesses, and when to be one