This is my first entry into the research journal; I’m early in my SEO career and I figured what better way to learn than to start doing what I’m learning. On 4 November 2026, I searched my own name and found nothing about me. This series explains what I built to change that, why I built it the way I did and the baseline I’ll measure every later change against. The longer-term results come in later entries, and I’ll publish them whatever they show.
Why I built this site
Search systems and LLMs piece together who someone is from whatever they can find. When they find little, they either say nothing or mix you up with someone else who has the same name as you. That was my starting point.
I’m an SEO professional based in Pembrokeshire, Wales. I work on technical SEO, entity SEO, SEO strategy and AI search visibility. A LinkedIn profile on its own doesn’t give search systems much to work with. I wanted one source I own and control that clearly states who I am and links to my other profiles, so search systems have something definitive to corroborate.
It also makes this site my first personal test case. Entity SEO is one of the areas I research, and here I can watch it happen on my own name, from zero, with the dates written down.
The strategy
Eight principles shaped the build.
- One owned source of truth. This site is the reference point for who I am. My other profiles point back to it, and the structured data gives me one permanent identifier tied to this domain; this won’t change even if the hosting does.
- One identity, stated the same way everywhere. The same name, role, location and specialisms appear in the page copy, the structured data, the footer and my profiles. Inconsistency is one of the easiest ways to confuse a search system or an LLM about who someone is.
- Plain sentences machines can extract. Key pages open with simple who-is-what statements: “Niall Gallacher is an SEO professional based in Pembrokeshire, Wales.” These are subject–predicate–object facts (semantic triples) that search systems and LLMs can lift directly.
- What you see matches what machines read. The LinkedIn and X links on the About page and in the footer come from the same list as the
sameAsproperty in my structured data, so the two can’t drift apart. - Topic clusters, not a single page. An SEO Research hub links to five areas: technical SEO, entity SEO, AI search and GEO, topical authority and content strategy, and local SEO. Journal categories mirror them, so every entry adds depth to one of them.
- Documentation, not sales. This is a research journal and evidence log, not a services site. I removed a template services page and the filler posts that came with the theme. Sections without evidence yet say so plainly.
- A recognisable look. The monochrome Milky Way starfield and editorial type match my LinkedIn cover, so the site and the profile read as the same person.
- Build first, launch when ready. Everything was built privately and kept out of search systems until it was finished, so nothing half-built was indexed.
What was built
The site runs on WordPress with a customised block theme. What it contains at launch:
- Pages: Home, About, an SEO Research hub with five research-area pages, this Journal, Projects, Contact and a Privacy Policy.
- Structured data: one connected graph on every page. It describes me as a Person (name, role, Pembrokeshire → Wales → United Kingdom, 17 topics I work on, two linked profiles), plus the WebSite and the page itself, all pointing at the same identifier. The About page is also marked up as an AboutPage.
- One clear H1 per page. The homepage H1 names me, my role and my location in plain HTML any crawler can read.
- The design: a slowly drifting starfield behind every page that fades out as you scroll, with translucent panels on top. It’s a CSS background, so all the text stays real, crawlable HTML, and the motion switches off for anyone whose device asks for reduced motion.
- Ways to come back: an RSS feed now, and an email list coming soon.
What isn’t here: no claimed clients, rankings or results. Those will appear when there’s evidence for them.
The baseline
The site went live at niallgallacher.uk on 9 November 2026, and search systems were allowed to index it on 9 November 2026. These are the measurements from launch week. Every later entry compares against them.
| Check | Tool | Result | Date |
|---|---|---|---|
| Google search for “Niall Gallacher” | Google (my laptop, signed in and out; then a different device, private window) | Before launch: the architect (Niall Gallacher Architecture) came up; nothing about me. Launch night on my own laptop: my homepage 1st and my About page 2nd. Just after midnight on a different device (private window): the architect 1st and 2nd, and I wasn’t on page one | 4 Nov 2026; 9 Nov 2026, ~23:15 (signed in) and ~23:18 (signed out); 10 Nov 2026, 00:05 (another device) |
| Who is Niall Gallacher, SEO, Pembrokeshire? | ChatGPT | Couldn’t find a strong, independently verifiable result establishing a professional profile for an SEO professional named Niall Gallacher in Pembrokeshire. | 8 Nov 2026, before launch. Rerun on launch night: see the AI rechecks below |
| Same question | Gemini | No public record or widely recognised information about a Niall Gallacher specialising in SEO in Pembrokeshire. | 8 Nov 2026, before launch. Rerun on launch night: see the AI rechecks below |
| Same question | Perplexity | Identified me correctly: a freelance SEO professional and content writer in Pembroke Dock, Pembrokeshire. Its only sources were my LinkedIn profile and a LinkedIn post. It also named my content-writing job and described me as early in my SEO career. | 8 Nov 2026, before launch. Rerun on launch night: see the AI rechecks below |
| Structured data, Home | Rich Results Test | Rich Results Test not run yet. Google’s URL Inspection live test found 1 valid ProfilePage item, no errors | 9 Nov 2026, 15:32 |
| Structured data, Home and About | Schema Markup Validator | 0 errors, 0 warnings (Home, About and Entity SEO) | 9 Nov 2026, 15:35 |
| Speed, mobile | PageSpeed Insights | Performance 100 · LCP 1.5 s · CLS 0.008 | 9 Nov 2026, 14:51 |
| Speed, desktop | PageSpeed Insights | Performance 100 · LCP 0.4 s · CLS 0.011 | 9 Nov 2026, 14:51 |
| Accessibility | PageSpeed Insights (Lighthouse) | 100 on mobile and desktop; no issues flagged | 9 Nov 2026, 14:51 |
| Sitemap | Google Search Console | Submitted; “Couldn’t fetch” on day one, which is normal for a new property. The sitemap itself checked valid, with 12 pages | 9 Nov 2026 |
| Homepage indexed? | Search Console URL Inspection | Indexed (“URL is on Google”). By the evening 10 of 12 pages were indexed. Projects followed by 10 Nov (11 of 12); indexing requested for Technical SEO on 10 Nov | 9 Nov 2026, 19:10 |
What stood out: Google indexed the homepage and nine other pages within about five hours of the site being opened up to search systems, which is far quicker than I expected for a brand-new domain. A Google search for “niall gallacher.uk” that evening showed my homepage and About page first and second, but Google’s summary above them was still about a different Niall Gallacher, an architect. That name confusion is exactly what this project is about.
Screenshots of each check are saved with their dates, so these results can be checked later.
Before launch: what the LLMs said
Before niallgallacher.uk was live or indexed, I asked six LLMs (ChatGPT, Claude, Gemini, Perplexity, Grok and DeepSeek) ten questions using LLM Brand Tracker on 7 November 2026. Each platform answered each question once, so these are small samples, not averages. The tracker records which people, sites and sources each answer mentioned, not whether the details were correct.
| Question | What came back |
|---|---|
| Who is Niall Gallacher? | ChatGPT and DeepSeek led with a “Niall Gallacher” profile. Perplexity and Claude pointed to Cyferd, and Gemini to an architect (niallgallacher.co.uk). ChatGPT also brought in a Gaelic poet, Niall O’Gallagher. |
| What does Niall Gallacher do? | Cyferd led for 3 of 6 (ChatGPT, Perplexity, Claude). Gemini gave an architect, investment firms and a nursing lecturer; DeepSeek gave drinks companies (Jameson, Pernod Ricard). None mentioned SEO. |
| Where is Niall Gallacher based? | Cyferd again for 3 of 6. DeepSeek pointed to Edrington and The Macallan; ChatGPT also listed a self-employed architect. None said Pembrokeshire or Wales. |
| What is the official website of Niall Gallacher? | 5 of 6 named niallgallacher.com, a domain I don’t own. None named niallgallacher.uk. |
| What information is available about Niall Gallacher? | ChatGPT and Perplexity led with a “Niall Gallacher” profile. The rest blended me with other people: an architect (Gemini), a technology executive at Cyferd, a different Niall Gallacher (Claude, ChatGPT), and investment-industry sources. Only DeepSeek mentioned SEO work, with no citations. |
| What sources confirm Niall Gallacher’s profession? | LinkedIn was cited by 5 of 6. The professions given conflicted: architect, technology executive, investment manager (often spelled “Gallagher”) and software developer. |
| Does Niall Gallacher provide SEO services? | Perplexity and DeepSeek named a Niall Gallacher, citing niallgallacher.com. Claude answered with other agencies (Nexmedia, Fifty21, Blackbird Marketing); Gemini cited Evelyn Partners and the BBC. |
| What areas of SEO does Niall Gallacher specialise in? (asked twice, worded slightly differently) | First run: ChatGPT, Gemini and Perplexity named a Niall Gallacher, and ChatGPT also cited a different SEO, Niall Cullen of Silo Digital. Second run: only Gemini named me; the rest cited SEO tools (Moz, Semrush, Ahrefs), Cyferd or investment firms. |
This is the reason I’ve made the site. The name is shared; the LLMs can’t tell the Nialls apart, and they credit a different domain as the official website. Success means niallgallacher.uk and “SEO professional in Pembrokeshire” starting to appear in these answers. I’ll ask the same ten questions at 30, 60 and 90 days.
Launch night: first AI recheck
On launch night I spot-checked a few of the baseline questions before doing the full rerun. Here’s what moved.
| Assistant | Question | Before launch (7 Nov) | Launch night (9 Nov) |
|---|---|---|---|
| Perplexity | Which SEO consultants serve businesses in Pembrokeshire? | Not named. Top pick: Delante, an international agency | Not named. Top pick: Claire Carlile Marketing, the only Pembrokeshire-based local SEO specialist it named. Leaned heavily on Sortlist |
| ChatGPT | Which SEO consultants serve businesses in Pembrokeshire? | Not named. Top pick: Web Adept, mixed in with tools, directories and out-of-area agencies | Not named. Top pick: Claire Carlile Marketing, and almost every name a real Pembrokeshire business |
| Gemini | Who is Niall Gallacher? | The architect (niallgallacher.co.uk) | Still the architect, in more detail: Glasgow-based, working independently since 2010 |
| Gemini | What does Niall Gallacher do? | An architect mixed up with investment firms and a nursing lecturer | All about the architect. No SEO, no Pembrokeshire, no niallgallacher.uk |
| ChatGPT | Who is Niall Gallacher? | A “Niall Gallacher” profile, plus Niall O’Gallagher, the Gaelic poet | Named nobody. Asked which Niall I meant (Wales, Scotland or Ireland?) |
| ChatGPT, signed out | What areas of SEO does Niall Gallacher specialise in? | A Niall Gallacher, plus a different SEO, Niall Cullen of Silo Digital | Named nobody. Said it couldn’t find reliable evidence linking me to SEO and wouldn’t guess |
| Gemini | What areas of SEO does Niall Gallacher specialise in? | Named a Niall Gallacher, alongside finance names like Evelyn Partners and BlackRock | Said outright that Niall Gallacher “does not specialise in any areas of SEO” and is an architect |
| Perplexity, signed in | What areas of SEO does Niall Gallacher specialise in? | Partly named; no clear specialisms | Named me correctly: “the Pembrokeshire-based SEO professional, not the Glasgow architect”, citing niallgallacher.uk for every specialism |
What it showed:
- The local answers got way more local overnight. Perplexity and ChatGPT both dropped the tools, directories and out-of-area agencies and moved to real Pembrokeshire businesses. Claire Carlile Marketing went from not named to top pick on both. Her pitch is clear (local SEO specialist, Chartered Marketer) and she’s on her own site and in directories. That’s the pattern to learn from, and the group I want to sit in.
- Gemini merges every Niall Gallacher into the architect. It isn’t saying it can’t find me, it’s confidently answering for all of us. One caveat: I asked all three Gemini questions in the same chat, and it titled that chat “Architect Niall Gallacher’s Background” after the first answer, so it was already set on him. The 16 November rerun asks every question in a fresh chat.
- ChatGPT wants evidence it can find on its own. Signed out, with no history to lean on, it wouldn’t put SEO next to my name, because niallgallacher.uk wasn’t turning up in its searches yet. That’s the clearest statement of what the site has to fix.
- Perplexity was the first real win. It gave the exact four specialisms my site states in its opening lines and structured data (technical SEO, entity SEO, SEO strategy and AI search visibility), so the plain who-is-what sentences are being lifted directly. Caveat: I was signed in, so the signed-out tracker rerun is the cleaner comparison.
Launch night: full tracker rerun
Later on launch night I ran the full question set through the same tracker as the baseline, across all six LLMs. Some questions were run twice. I also added a new one, “Who is Niall Gallacher, SEO, Pembrokeshire?”. That’s 13 unique exports, compared question by question with 7 November below.
| Question | Before launch (7 Nov) | Launch night (9 Nov) | niallgallacher.uk named? |
|---|---|---|---|
| Who is Niall Gallacher? | A “Niall Gallacher” profile, then Cyferd; Gemini gave the architect | Cyferd first in one run, a “Niall Gallacher” profile in the other; Perplexity added Tracebit and Proofpoint; Gemini still the architect | No |
| What does Niall Gallacher do? | Cyferd led; Gemini gave the architect and a nursing lecturer | Cyferd led again; Gemini gave the architect and the poet | No |
| Where is Niall Gallacher based? | Cyferd led; nobody said Pembrokeshire | The architect led one run (Claude and Gemini), Cyferd the other; nobody said Pembrokeshire | No |
| What is the official website? | niallgallacher.com, 5 of 6 | niallgallacher.co.uk, the architect’s domain, 4 of 6 | No |
| What information is available? | A “Niall Gallacher” profile; DeepSeek mentioned SEO work | A “Niall Gallacher” profile; no SEO mention; architect sources higher | No |
| What sources confirm his profession? | LinkedIn (5 of 6); conflicting professions | LinkedIn (5 of 6); the Architects’ Journal now second | No |
| Does he provide SEO services? | niallgallacher.com (Perplexity, DeepSeek) | niallgallacher.com again; Gemini gave the architect and a construction firm | No |
| What areas of SEO does he specialise in? | A Niall Gallacher named in one run | SEO tools (Ahrefs, Semrush, Screaming Frog); Gemini gave the architect | No |
| Who is Niall Gallacher, SEO, Pembrokeshire? (new) | Not in the tracker | DeepSeek, Grok and Perplexity named a Niall Gallacher, but linked him to niallgallacher.com; ChatGPT listed local agencies; Gemini and Claude gave the architect | No |
niallgallacher.uk was named 0 times across all 13 exports, the same as before launch.
What the full run shows:
- The architect is getting stronger, not weaker. He now leads one “where is he based” run, the Architects’ Journal is the second most cited source for my profession, and the official-website answer moved to his domain.
- There is a Niall Gallacher SEO in their heads, but at the wrong address. For the new question, half the LLMs recognised a Niall Gallacher who does SEO in Pembrokeshire. They linked him to niallgallacher.com, a domain I don’t own, rather than niallgallacher.uk.
- The tracker and my manual checks disagree. Signed in, Perplexity cited niallgallacher.uk tonight. Through the tracker it didn’t, so the signed-in result was probably helped by my account. The tracker is the measurement I’ll trust going forward.
- This is day one. These LLMs answer from search indexes that haven’t picked up a site that went live this afternoon. The baseline is now properly set from both sides of launch, and the next full run is on about 16 November.
The raw exports and a question-by-question comparison are in my launch spreadsheet.
Launch day: 9 November 2026
niallgallacher.uk went live on 9 November 2026 and scored 100 in every Google Lighthouse category on mobile and desktop. Everything below was tested on launch day.
Launch timeline
| Time | What happened |
|---|---|
| 14:15 | Theme and all 11 remaining pages published on the temporary address |
| ~14:24 | Clean page addresses confirmed site-wide (/about-me/, /seo-research/ and the five research areas) |
| 14:33 | Lighthouse on the temporary address: mobile 97, desktop 100 (SEO 69 because indexing was still blocked) |
| ~14:45 | niallgallacher.uk connected in IONOS; WordPress address switched to the real domain |
| ~14:48 | Search-engine visibility switched on; new fonts (Tektur and Jura, matched to the NG logo) published |
| ~14:50 | Old temporary IONOS address set to forward permanently (301) to niallgallacher.uk |
| 14:51 | Lighthouse on niallgallacher.uk: 100 / 100 / 100 / 100 on mobile and desktop |
| ~15:00 | Google Search Console domain property verified; sitemap submitted |
Google live test, 15:32: Search Console’s URL Inspection live test on https://niallgallacher.uk/ returned “URL is available to Google” and “Page can be indexed”. It also detected 1 valid ProfilePage item, with no errors or warnings: Google’s own confirmation that it can read the Person entity markup. Indexing was requested the same day. At 15:37 the same test on https://niallgallacher.uk/about-me/ passed too, with 1 valid Breadcrumbs item, and indexing was requested for it as well.
Google Lighthouse (PageSpeed Insights), homepage
| Run | Device | Performance | Accessibility | Best practices | SEO | First / largest paint | Layout shift | Blocking time |
|---|---|---|---|---|---|---|---|---|
| 14:51, niallgallacher.uk | Mobile (slow 4G) | 100 | 100 | 100 | 100 | 1.5 s / 1.5 s | 0.008 | 0 ms |
| 14:51, niallgallacher.uk | Desktop | 100 | 100 | 100 | 100 | 0.4 s / 0.4 s | 0.011 | 0 ms |
| 14:33, temporary address | Mobile (slow 4G) | 97 | 100 | 100 | 69 | 1.7 s / 1.7 s | 0 | 0 ms |
| 14:33, temporary address | Desktop | 100 | 100 | 100 | 69 | 0.4 s / 0.4 s | 0.001 | 0 ms |
The 14:33 SEO score of 69 had one cause: the deliberate pre-launch block on indexing. It reached 100 once indexing was switched on. Both runs also passed Lighthouse’s new AI agent browsing check (2/2). Remaining suggestions are optional: self-hosting the fonts (about 0.7 s less render-blocking on mobile), a smaller homepage headshot (about 66 KB) and longer browser caching.
Technical launch checks
| Check | Result |
|---|---|
| Pages | All 12 pages load on niallgallacher.uk with no errors |
| Titles, meta descriptions, canonical tags | Exactly one of each per page, all on niallgallacher.uk |
| H1 headings | Exactly one per page |
| Structured data | One JSON-LD graph on every page: Person (Pembroke Dock, Pembrokeshire, Wales), WebSite and the page itself; ProfilePage on the homepage, AboutPage with breadcrumbs on About; both sameAs profiles (LinkedIn, X) present |
| robots.txt | Allows crawling; points to https://niallgallacher.uk/wp-sitemap.xml |
| XML sitemap | Valid XML; lists all 12 pages at their clean addresses |
| Internal links | 18 unique internal links checked, all working with no redirects |
| Redirects | www.niallgallacher.uk and the old temporary address both forward permanently to niallgallacher.uk |
| Leftover temporary addresses | None in any page’s code |
| Images | 6 images in use, all renamed to match the schema (niall-gallacher-seo-professional-pembrokeshire-wales…); 51 unused files removed from the media library |
| Link previews | 1200 × 630 share card on every page; NG logo as the browser and phone icon |
| Phone and desktop layout | No sideways scrolling, no large gaps, no links or buttons under 24 px |
Schema Markup Validator (schema.org), 15:35, fetched live from niallgallacher.uk: 0 errors and 0 warnings on all three page types tested.
| Page | Items detected | Errors | Warnings |
|---|---|---|---|
| Homepage | WebSite → Person (PostalAddress, AdministrativeArea, Country, ImageObject) → ProfilePage, one connected graph | 0 | 0 |
| About (/about-me/) | 3 items, including AboutPage with breadcrumbs | 0 | 0 |
| Entity SEO (/seo-research/entity-seo/) | 3 items, including WebPage with breadcrumbs | 0 | 0 |
Launch evening: indexed on day one
By 19:19, about four and a half hours after search systems were let in, Search Console showed 10 of the 12 pages already indexed. I hadn’t expected anything for a few weeks.
| Page | Status in Search Console | Rich result items | Checked |
|---|---|---|---|
| Home | Indexed (“URL is on Google”) | ProfilePage, 1 valid | 19:10 |
| About | Indexed | Breadcrumbs, 1 valid | ~19:15 |
| SEO Research | Indexed | Breadcrumbs, 1 valid | ~19:15 |
| Entity SEO | Indexed | Breadcrumbs, 1 valid | ~19:15 |
| AI Search & GEO | Indexed | Breadcrumbs, 1 valid | ~19:15 |
| Topical Authority | Indexed | Breadcrumbs, 1 valid | ~19:15 |
| Local SEO | Indexed | Breadcrumbs, 1 valid | ~19:15 |
| Journal | Indexed | Breadcrumbs, 1 valid | ~19:15 |
| Contact | Indexed | Breadcrumbs, 1 valid | 19:19 |
| Privacy Policy | Indexed | Breadcrumbs, 1 valid | 19:19 |
| Projects | Indexed by 10 Nov (live test passed 9 Nov, indexing requested) | Breadcrumbs, 1 valid | 19:17 |
| Technical SEO | Not indexed yet; live test passed 9 Nov, indexing requested 10 Nov | Breadcrumbs, 1 valid | 9 Nov |
A Google search for “niall gallacher.uk” the same evening, on my own laptop, put my homepage first and my About page second. Both still showed the titles from before that evening’s wording change, with a plain globe icon instead of my logo, and Google’s summary above them was about the architect with my name.
A Google search for just “Niall Gallacher”: two very different answers. At about 23:15 on launch night, on my own laptop, my homepage ranked first and my About page second for my name on its own. I repeated it signed out a few minutes later and got the same result. Before launch, that search led with the architect and showed nothing about me.
Then, just after midnight, I ran the same search on my girlfriend’s iPad, in a private window, signed out, located in Tenby. That told a different story. The architect’s site took first and second place (niallgallacher.co.uk and its info page), followed by his LinkedIn, an investment manager at Jupiter, Facebook profiles, Companies House, the Association of Consultant Architects and a LinkedIn directory listing. My site wasn’t on page one at all. The only trace of me was that LinkedIn directory snippet, which mentions my content-writing job.
Signing out on my laptop wasn’t enough to remove the personal effect. My browser had been visiting niallgallacher.uk and Search Console all day, and Google clearly took that into account. The iPad is the honest result: on day one, to a stranger searching my name, the architect still owns it. I’m logging both, and from now on the name check will always be done on a device that has never visited my site.
Launch evening: two audits, one clean-up
Once the site was live I put two audits side by side: my own pre-launch audit from 7 November and an independent audit from ChatGPT. Between them they raised 38 findings. Where both found the same thing, I counted it once.
| Status at the end of launch day | Findings |
|---|---|
| Fixed and verified | 23 |
| Improved, needs validating over time | 3 |
| Not fixed, needs access I don’t have yet | 5 |
| Future content | 7 |
| Total | 38 |
What changed that evening:
- One wording everywhere. “Freelance SEO Professional in Pembrokeshire, Wales” now appears the same way in the homepage, About and Contact headings, the page titles and the structured data. Before, some pages left out “Freelance”.
- Pembroke Dock on the About page, so the page matches the location in my structured data.
- The research pages grew from 87–131 words to 232–278, each explaining what the area covers and where I’m starting, using only real results from this project. Wording that made sections look abandoned (“No articles yet”, “After launch”) was rewritten.
- Projects got a real write-up of this build, including the LLM baseline and the launch-day checks.
- Clearer routes from the homepage: buttons into SEO research, Projects and Contact under the intro, all five research areas linked, and a separate Keep exploring section for About, the Journal and Contact.
- Speed: the fonts now load from my own site instead of Google Fonts, and the homepage headshot was sized down to suit how it’s shown (tightened again on 12 November, after PageSpeed still flagged it).
- My logo for Google results: the NG logo is now served at 48, 180 and 192 px, the sizes Google uses for the icon next to a result. When it shows up is Google’s call.
- Contact form: without JavaScript the form’s spam check rejects the message and nothing is sent, so visitors in that situation now see a plain note pointing to my LinkedIn and X. A hidden anti-spam field no longer shows for them either.
- Tidy-up: two unused plugins (BeyondSEO and TranslatePress) were removed.
More launch-day checks
| Check | Result |
|---|---|
| Search and AI crawlers | The About page loaded in full (200) for Googlebot, Bingbot, OAI-SearchBot, GPTBot, ChatGPT-User, PerplexityBot, ClaudeBot, Google-Extended and CCBot, tested from the hosting. An IONOS firewall in front of the site could still treat them differently |
| Redirects | http, www and the old temporary address all forward permanently to https://niallgallacher.uk |
| Duplicate titles and descriptions | None across the 12 pages |
| Person entity | Exactly one Person on every page, same identifier |
| Contact form | Required-field errors and the confirmation message both work |
| Colour contrast | No failures in an automated check of the homepage |
| Browser caching | Images, CSS and fonts expire after 30 minutes. That’s set by IONOS, not WordPress, so it’s on the to-do list |
Launch night: name and keyword searches
At about 22:40 on launch night I ran two sets of searches through a general web search tool. This wasn’t Google and it wasn’t the LLMs, but it’s the kind of search index many LLMs pull from, so it’s a useful second record for day zero.
My name. niallgallacher.uk didn’t come up for “Niall Gallacher”, “Niall Gallacher SEO Pembrokeshire”, “niallgallacher.uk”, “SEO professionals based in Pembrokeshire” or “official website of Niall Gallacher SEO”. The results were the same mix as before launch: Niall O’Gallagher the Gaelic poet, a Donegal footballer, other Gallachers, and Pembrokeshire SEO directories. Google already has 10 of my pages indexed, but this index hasn’t caught up yet. That’s why Bing Webmaster Tools is next on my list, because Bing feeds ChatGPT search and Copilot.
The 30 keywords from my AnswerSocrates list. I searched every one and niallgallacher.uk appeared for none of them, which is what I’d expect with no articles published yet.
| Stage | Keywords | niallgallacher.uk appeared | Directory or listing site in the results |
|---|---|---|---|
| Top of funnel (guides and questions) | 10 | 0 | 5 |
| Middle of funnel (comparisons and reviews) | 10 | 0 | 9 |
| Bottom of funnel (prices and booking) | 10 | 0 | 9 |
| Total | 30 | 0 | 23 |
What stood out:
- Directories own the local results. Bizhub365, Sortlist, Clutch, Bark and Techbehemoths showed up for 23 of the 30 keywords, and Bizhub365 filled most of the price and booking searches by itself. Very few real Pembrokeshire SEO people rank with their own pages.
- The search system mixes up the place. “Pembrokeshire” searches kept bringing back Plymouth agencies and US towns called Pembroke. If LLMs are working from results like these, it explains why their local answers before launch were so scattered.
- The bar for useful content is low. A sandbox wiki page on a US university site ranked for six of the guide-style questions about SEO in Wales.
- The names the LLMs gave most before launch weren’t there. Nexmedia and Web Adept didn’t appear in any of the 30 searches. Jamie King (South & West Wales) and Systematic Edge Digital were the only local names to show up.
- The question keywords are the most open. Searches like “how to improve local seo in pembrokeshire” and “seo basics for pembrokeshire startups” returned generic UK and US guides. That’s where my first Local SEO articles should aim.
I’ll run the same 30 keywords again alongside the 30, 60 and 90-day AI checks.
Everything from launch day, including the raw AI tracker exports and the keyword search check, is kept in one spreadsheet, so every figure here can be checked against the source data.
What I expect to change, and what I don’t
These are expectations, not results.
- Indexing first. I expect Google to crawl and index the homepage and main pages within a few weeks. That’s the first thing I’ll measure.
- My name, then my description. Once indexed, the site should start appearing for searches of my name. Whether Google links the name to “SEO professional in Pembrokeshire” is the deeper test.
- LLMs will be slow. They depend on what’s been crawled and corroborated elsewhere, so I don’t expect much change in the first month.
- No guarantees. Structured data helps search systems interpret a page. It doesn’t guarantee a Knowledge Panel, rich results or recognition by LLMs. If none of this moves, that’s a finding too, and I’ll write it up.
When I’ll check again
I’ll repeat every check in the baseline table at 30 days, 60 days and 90 days after launch, with the same tools and questions, and publish the results here. I’m publishing weekly in the meantime, starting with how each part of this site was put together.
To follow along, subscribe via RSS.
The full launch data
If you want the detail, it’s all here. Open any section to see it in full.
Audit checklist
Mine = my 7 November audit; ChatGPT = the independent audit; Both = found by both.
| # | Finding | Source | Priority | Status at the end of launch day |
|---|---|---|---|---|
| 1 | Search engines blocked: “Discourage search engines” on, noindex on every page | Mine | P1 | Fixed and verified |
| 2 | robots.txt had no Sitemap line | Mine | P1 | Fixed and verified |
| 3 | Sitemap handled by a leftover SEO plugin; core sitemap off | Mine | P1 | Fixed and verified (file); Search Console status still to confirm |
| 4 | http, www and the old temporary address all served the site | Both | P1 | Fixed and verified |
| 5 | Canonicals pointed at the temporary domain | Both | P1 | Fixed and verified |
| 6 | Leftover SEO plugin printing duplicate titles, descriptions and schema | Mine | P1 | Fixed and verified |
| 7 | Risk of duplicate or conflicting Person entities | ChatGPT | P1 | Fixed and verified |
| 8 | Contact form depends on JavaScript | ChatGPT | P1 | Improved, validation required |
| 9 | Pages not yet in Google | Both | P1 | Fixed and verified for 11 pages (Projects confirmed 10 Nov); Technical SEO indexing requested |
| 10 | Role worded differently across pages | ChatGPT | P2 | Fixed and verified |
| 11 | Person schema missing some real areas of work | ChatGPT | P2 | Fixed and verified |
| 12 | Pembroke Dock in the schema but not on the About page | Mine | P2 | Fixed and verified |
| 13 | sameAs limited to two profiles (LinkedIn, X) | Both | P2 | Future content |
| 14 | Homepage title would have said “Home” | Mine | P2 | Fixed and verified |
| 15 | Homepage didn’t route clearly into research, projects, journal and contact | ChatGPT | P2 | Fixed and verified |
| 16 | Research pages thin (87–131 words) | ChatGPT | P2 | Fixed and verified |
| 17 | Headshot downloaded larger than shown | Mine | P2 | Improved, validation required (PageSpeed still flagged it on 11 Nov; tightened again 12 Nov) |
| 18 | Fonts loaded from Google Fonts | Mine | P2 | Fixed and verified |
| 19 | Short browser caching on images, CSS and fonts | Mine | P2 | Not fixed, access required |
| 20 | Tap targets too small on phones | Mine | P2 | Fixed and verified |
| 21 | Colour contrast and form accessibility | ChatGPT | P2 | Fixed and verified |
| 22 | Logo not showing next to results in Google (icon sizes) | Mine | P2 | Improved, validation required |
| 23 | Rich Results Test not run | Mine | P2 | Not fixed, access required |
| 24 | Bing Webmaster Tools not set up | Mine | P2 | Not fixed, access required |
| 25 | Form notifications go to Gmail; domain mailbox untested | Mine | P2 | Not fixed, access required |
| 26 | Cookie consent for Metricool | Mine | P2 | Not fixed, access required |
| 27 | About page grammar, spelling and positioning | ChatGPT | P3 | Fixed and verified |
| 28 | Wording that made the site feel unfinished | ChatGPT | P3 | Fixed and verified |
| 29 | Projects page had no project on it | ChatGPT | P3 | Fixed and verified |
| 30 | Journal has no articles | Both | P3 | Future content |
| 31 | Local SEO foundation for Pembrokeshire | ChatGPT | P3 | Fixed and verified |
| 32 | llms.txt | ChatGPT | P3 | Fixed and verified |
| 33 | AI crawler access | ChatGPT | P3 | Fixed and verified |
| 34 | Article schema for journal entries | ChatGPT | P3 | Future content |
| 35 | Publishing research | Both | P4 | Future content |
| 36 | External mentions | ChatGPT | P4 | Future content |
| 37 | Topical authority | ChatGPT | P4 | Future content |
| 38 | Monitoring Google and AI visibility | Both | P4 | Future content |
Launch-day changes
| Change | Pages and files |
|---|---|
| Search engines allowed; Sitemap line in robots.txt; core sitemap (12 pages) | Settings, robots.txt, /wp-sitemap.xml |
| 301 redirects: http to https, www to non-www, temporary address to niallgallacher.uk | functions.php |
| One JSON-LD graph per page with one Person @id; sameAs LinkedIn and X | inc/entity-seo.php |
| Person: 17 knowsAbout topics; Pembroke Dock; jobTitle Freelance SEO Professional | inc/entity-seo.php |
| Homepage and About titles now include “Freelance SEO Professional” | inc/entity-seo.php |
| H1 “Freelance SEO Professional in Pembrokeshire, Wales” on Home and Contact | Pages: Home, Contact |
| Homepage: new tagline; SEO research, Projects and Get in touch buttons; project stages updated; Keep exploring section | Page: Home; style.css |
| Research pages: What this covers / Where I’m starting; related links; empty-section wording; empty archive links hidden | Pages: 5 research areas; style.css |
| Empty category and tag archives noindex, follow | functions.php |
| About: grammar, UK spelling, skills, Pembroke Dock, link to Projects | Page: About |
| Projects: real write-up of the niallgallacher.uk build | Page: Projects |
| Contact: intro reworded; State of Search described as planned; confirmation in my words; honeypot hidden without JavaScript; no-JavaScript message | Page: Contact; WPForms 23; functions.php; style.css |
| Fonts self-hosted | uploads/ng-fonts/; functions.php |
| Headshot sizes attribute | functions.php |
| Site icon at 48, 180 and 192 px | uploads/ng-icons/; functions.php |
| llms.txt | functions.php |
| BeyondSEO and TranslatePress uninstalled | Plugins |
30-keyword check
General web search, not Google and not the LLMs. niallgallacher.uk appeared for none of the 30. Stage is the funnel stage: top (guides and questions), middle (comparisons and reviews), bottom (prices and booking).
| Stage | Keyword | Top results | Directory in results? |
|---|---|---|---|
| Top | how to improve local seo in pembrokeshire | jemsu.com (Pembroke Pines, US); wpsuites.com; betheluniversity.edu wiki | Yes |
| Top | what is seo and why it matters for pembrokeshire businesses | betheluniversity.edu wiki (x3); bizhub365.com; jamieking.co.uk | Yes |
| Top | guide to seo for small businesses in pembrokeshire | findbusinesssupport.gov.scot; jamieking.co.uk; cloudswitched.com | Yes |
| Top | benefits of local seo for pembrokeshire companies | betheluniversity.edu wiki (x3); bizhub365.com | Yes |
| Top | how does google rank pembrokeshire businesses | phitweb.com; creativemarketing.ai (Pembroke NC/IL/NY, US); bizhub365.com | No |
| Top | seo basics for pembrokeshire startups | findbusinesssupport.gov.scot; profiletree.com (x2); jamieking.co.uk | No |
| Top | how to get found online in pembrokeshire | solva.gov.wales; pembrokeshire-herald.com; birdandco.co.uk | No |
| Top | why local seo is important for pembrokeshire shops | betheluniversity.edu wiki (x2); bizhub365.com | Yes |
| Top | how to choose keywords for pembrokeshire customers | web-studio.co.uk; presearch.io; gracker.ai | No |
| Top | understanding google my business for pembrokeshire firms | wales247.co.uk; safetech.co.uk; profiletree.com | No |
| Middle | best seo agency in pembrokeshire | sortlist.co.uk (x2); epicedits.co.uk (Plymouth); createandgrow.com (Plymouth) | Yes |
| Middle | top seo consultants pembrokeshire reviews | clutch.co (Plymouth); sortlist.co.uk (x3); semrush.com | Yes |
| Middle | seo pembrokeshire vs in house marketing | anicca.co.uk; clickintelligence.co.uk; huble.com | Yes |
| Middle | affordable seo services pembrokeshire comparison | bizhub365.com (x3); seoswansea.pages.dev; techbehemoths.com | Yes |
| Middle | best local seo packages pembrokeshire | mr-marketing.co.uk; createandgrow.com; profiletree.com; bizhub365.com | Yes |
| Middle | seo companies in pembrokeshire compared | createandgrow.com (Plymouth); sortlist.co.uk; bizhub365.com; techbehemoths.com | Yes |
| Middle | freelance seo vs agency pembrokeshire | seocompare.pages.dev; coolplugins.net; digitaleer.com (none local) | No |
| Middle | top rated seo specialists pembrokeshire | sortlist.co.uk (x3); castuk.com; bizhub365.com | Yes |
| Middle | seo pembrokeshire case studies and results | casestudies.com (Plymouth); epicedits.co.uk; sortlist.co.uk | Yes |
| Middle | best seo tools for pembrokeshire businesses | selectsoftwarereviews.com; createandgrow.com; centricdxb.com | Yes |
| Bottom | buy seo services in pembrokeshire | bizhub365.com (x6); sortlist.co.uk; webgel.co.uk (Plymouth) | Yes |
| Bottom | seo pembrokeshire price and packages | bizhub365.com (x5); epicedits.co.uk; profiletree.com | Yes |
| Bottom | hire local seo expert pembrokeshire cost | bizhub365.com (most results); sortlist.co.uk | Yes |
| Bottom | seo audit pembrokeshire book now | upwork.com; gracker.ai; bizhub365.com | Yes |
| Bottom | pembrokeshire seo agency contact for quote | sortlist.co.uk (x2); bizhub365.com; castuk.com; bark.com | Yes |
| Bottom | local seo management pembrokeshire pricing | bizhub365.com (x4); epicedits.co.uk; clickup.com | Yes |
| Bottom | seo consultation pembrokeshire book today | bizhub365.com; sortlist.co.uk; ralfvanveen.com (Plymouth) | Yes |
| Bottom | monthly seo retainer pembrokeshire cost | bizhub365.com (x4); seoswansea.pages.dev; ahrefs.com | Yes |
| Bottom | seo pembrokeshire free website audit request | gumroad.com; kobault.com; localiq.co.uk (none local) | No |
| Bottom | best seo deal pembrokeshire sign up | sortlist.co.uk (x2); bizhub365.com (x3); pembrokeshire-herald.com | Yes |
What was left to do
| Area | Action | Priority | Status |
|---|---|---|---|
| Search Console | Request indexing for the homepage and About again (new titles and logo sizes) | High | Open |
| Search Console | Inspect Technical SEO and request indexing: requested 10 Nov, waiting on Google | High | Open |
| Search Console | Recheck Projects indexing: indexed by 10 Nov | Medium | Done |
| Search Console | Recheck the sitemap status in 3–5 days | Medium | Open |
| “Niall Gallacher” Google search on a clean device: architect 1st and 2nd, me not on page one (my own laptop showed me 1st and 2nd). Repeat weekly on a device that has never visited the site | High | Done | |
| Rich Results Test while signed in to Google | Low | Open | |
| Bing | Set up Bing Webmaster Tools; import from Search Console | High | Open |
| Contact | Check Gmail for the test form message; decide on the notification address | Medium | Open |
| Hosting | Ask IONOS about longer browser caching | Low | Open |
| Plugins | Check Sugar Calendar is unused, then remove it; empty the trash | Low | Open |
| Privacy | Cookie consent if Metricool tracking is connected | Medium | Open |
| Testing | Rerun PageSpeed Insights after the launch-evening changes | Low | Open |
| Journal | Rerun the 10 AI questions around 16 Nov, then at 30, 60 and 90 days | High | Open |
| Content | One real article per research area as the work happens | Medium | Open |
| Plugins | Remove BeyondSEO and TranslatePress | Low | Done |
| Fonts | Self-host the fonts | Low | Done |
Leave a Reply