Category: SEO

  • The SEO audit process from kickoff to handover

    The SEO audit process from kickoff to handover

    The verdict

    A useful SEO audit is a diagnosis, not an inventory. It starts with what changed and what the business needs, works through crawling, rendering, indexing and ranking in that order, and ends with a short list of fixes ranked by what they are costing you. If nothing is worth fixing, the audit should say so.

    Most people who buy an SEO audit have bought one before. What they got was a long PDF, a health score, and several hundred issues sorted by severity colour. Some of it was fixed. Nothing much changed.

    That is not a failure of effort. It is a failure of method. A crawl tool can list everything that differs from a checklist. It cannot tell you which of those differences is the reason you are not ranking. That judgement is the audit. Everything else is preparation for it.

    This is how I run one, from the first call to the handover.

    Kickoff: what changed, and what matters

    Before any tool runs, I want three answers. What does the business actually sell, and which searches bring in buyers? When did things go wrong, or have they never gone right? And what changed on the site, in the team or in the market around that time?

    Those answers shape everything after. A site that lost half its traffic in March needs a different audit from a site that has never ranked. Google’s own advice on hiring an SEO makes the same point from the buyer’s side: a good SEO “should ask questions about what makes your business unique, who your competitors are, how search results can help, and how your customers find you.” If an audit starts without those questions, it is a crawl report with a cover page. I wrote more on the difference in an SEO audit is not a crawl report.

    Access: the data that decides it

    • Search Console. The only source of what Google itself recorded: queries, pages, indexing status, manual actions.
    • Analytics. Which landing pages bring traffic that converts, so the audit weights pages by value, not by page count.
    • The CMS, read-only if you prefer. Templates and settings explain problems a crawler only sees the symptoms of.
    • Business Profile, for local businesses. Half of local visibility lives there, not on the website.
    • A change history. Redesigns, plugin changes, content removals, agency handovers. Dates matter more than detail.

    Diagnosis: four stages, in order

    I work through the site the way Google does. Can it find the important pages? Can it see their content once it renders them? Does it choose to index them? And once indexed, why do they rank where they rank? I explain that sequence in crawling, rendering, indexing, ranking: which stage is broken, and the order matters because a problem at an early stage hides everything after it.

    The crawl happens here, but it is evidence, not the finding. A crawler flags every page without a meta description. The audit decides whether any of those pages matter.

    Ranking the findings by cost

    An audit with forty findings of equal weight is not finished. The useful version has a short list at the top: these three things are costing you rankings on the searches that bring you customers, here is the evidence, here is the fix. Then a second tier. Then a list of things that are technically true and not worth your money right now.

    That last list matters as much as the first. Every fix costs developer time, and the most expensive audit is one that sends a team after problems that were never the cause. I explain how I decide in if you only do three things, which three?

    Why the tools cannot do this part

    Google is unusually direct about third-party tools. Its hiring guidance says Google “doesn’t evaluate or endorse third-party SEO tools, and these tools don’t have access to Google’s internal ranking data.” It also tells site owners to check a tool’s recommendations “against official guidance from Google Search” before making significant changes.

    That is the gap an audit fills. Tools are good at finding differences. They are poor at knowing which differences Google cares about on your site, for your queries, against your competitors.

    It also explains why so many sites never recover their investment. Ahrefs’ study of 14 billion pages found 96.55% get no traffic from Google at all. Most of those pages are not broken in any way a crawler would flag. They are simply not what anyone is searching for, or not better than what already ranks. An audit that only looks at technical health will never find that.

    The deliverable

    A prioritised fix list, not a 90-page PDF. Each finding names the pages affected, the evidence, the fix, and who has to do it. Developer items are written so a developer can act on them without a meeting. Content items say which page, what is missing, and why it matters for that search.

    And if the site is in reasonable shape and the real constraint is something SEO cannot fix, such as a market that is too small or a business with no reviews, the audit says that. An honest “nothing here is costing you much” is worth more than a manufactured list.

    Handover and after

    The audit ends with a walkthrough, so the people doing the work understand why each item is there. Then the fixes ship, and the useful check is a few months later: did the pages that were fixed move, and did they move on the searches that matter? A fixable indexing problem can turn around in weeks. A history of bad links takes months. The audit should tell you which kind of problem you have, so you know what to expect.

    The short version

    1. An audit is a diagnosis. The crawl is only evidence.
    2. Start with what changed and what sells. The answers shape everything after.
    3. Get Search Console, analytics and the change history. They decide the findings.
    4. Work through crawl, render, index, rank in order. Early problems hide later ones.
    5. Rank findings by cost. Name what is not worth fixing too.
    6. Say so when nothing is wrong. That is a finding as well.

    If a site should be ranking and it isn’t, that’s the work I do. An SEO audit is a fixed-scope diagnosis that ends in a short list of what is actually costing you.

  • Structured data in 2026: which types still produce something visible

    Structured data in 2026: which types still produce something visible

    The verdict

    For most service businesses, structured data now produces almost nothing you can see in search. FAQ and HowTo results are gone, and your own reviews cannot earn stars. What still shows is Product markup for stores, Organization details, breadcrumbs and a few niche types. I add it last, after the problems that actually cost traffic.

    Structured data used to be an easy win to sell. Add FAQ markup, get a bigger result, take more of the page. Agencies built whole retainers around it. Most of that ended in 2023, and the rest has been closing steadily since.

    So the honest question for a business site in 2026 is not “have we got schema?” It is “which of the types we have still does anything?” For a law firm, a clinic or a contractor, the answer is a short list.

    What Google has taken away

    The removals have come in waves, and Google has documented each one on its Search documentation updates page:

    • HowTo, 2023. Removed from results on desktop and mobile in September 2023.
    • FAQ, 2023 to 2026. Limited to “well-known, authoritative government and health websites” in August 2023, as Google announced at the time. Then dropped entirely: the feature stopped appearing from 7 May 2026 and the documentation was deleted.
    • A batch of niche types, 2025. Course info, estimated salary, learning video, special announcement and vehicle listing were phased out, and their documentation was removed in September 2025.
    • Practice problems, 2026. Removed in January 2026.

    Google’s stated reason for the recent removals is its “ongoing efforts to simplify the Google Search results page.” Whatever the reason, the direction is clear. Fewer types produce a visible result every year, and nothing suggests that is about to reverse.

    Your own reviews will not earn stars

    This is the one I get asked about most, because review stars are the result people picture when they hear “rich snippets.” For a business marking up its own testimonials, they are not available. Google’s review snippet documentation says that if the business being reviewed “controls the reviews about itself, their pages that use LocalBusiness or any other type of Organization structured data are ineligible for star review feature.”

    That rules out almost every service business marking up the testimonials on its own site. I explain the rest of that rule in why your own reviews do not belong in your schema.

    What still produces something visible

    For the kinds of sites I work on, this is the realistic list:

    • Product, for stores. Price, availability, shipping and return details can all appear in results and in Shopping surfaces. For ecommerce this is the one type that clearly pays for itself.
    • Organization. Logo, name, contact details and identifiers that Google can use in knowledge panels and brand results.
    • Breadcrumb. Can change how the URL line reads in a result. Small, cheap, and harmless.
    • Article and Profile page. Help Google connect content to an author or organisation. Useful on a site where who wrote the page matters.
    • Event, Job posting, Recipe and the other vertical types. Only relevant if that is what the business does.

    LocalBusiness markup is still documented, and Google says it can feed details into the knowledge panel. In practice, for a business with a verified Business Profile, the profile is where Google gets those details. The markup is a consistency check, not a ranking lever.

    What structured data cannot do

    It does not make a page rank. It does not get a page into AI answers. Google’s own guidance on AI features says there is “no special schema.org structured data that you need to add,” and I go through what it is still useful for in schema will not get you into AI answers.

    It also has to describe what is actually on the page. Markup that claims content the visitor cannot see breaks Google’s structured data policies and is a common way for a well-meant plugin to cause trouble. Many sites already output more schema than their owners realise, which is why I check what your website builder already says about you in JSON-LD before adding anything.

    Why I add it last

    Not because it is worthless. Because on a business site it almost never explains why the site is not ranking. A site with orphaned service pages, thin location pages and a Business Profile in the wrong category will not be rescued by valid Organization markup. When those problems are fixed, structured data is an hour of tidy work that makes the site’s details consistent.

    Ecommerce is the exception. For a store, Product markup is close to essential, because it affects how products appear in results that compete directly for the click.

    How I check what a site has

    • Read the JSON-LD on three templates. Home, a service or product page, and a blog post. Most sites have three or four sources of markup and no single owner.
    • Look for duplicates. Two Organization blocks with different names or addresses is worse than none.
    • Remove dead types. FAQ markup no longer produces anything. It is not harmful, but it is clutter someone will keep maintaining.
    • Check Search Console’s enhancement reports. They show which types Google detected and whether it found errors.

    The short version

    1. FAQ and HowTo results are gone. FAQ ended completely in May 2026.
    2. Your own reviews cannot earn stars. Google calls them self-serving.
    3. Product markup matters for stores. It is the type that clearly pays.
    4. Organization and breadcrumbs are tidy, cheap work. Not a ranking lever.
    5. No schema gets you into AI answers. Google says so directly.
    6. Do it last on a service site. It rarely explains why a site is not ranking.

    If a site should be ranking and it isn’t, that’s the work I do. Technical SEO puts structured data where it belongs in the order of work, which is usually near the end.

  • HTTP status codes for site owners: what each one tells Google

    HTTP status codes for site owners: what each one tells Google

    The verdict

    A status code is an instruction to Google, not an error message for you. A 301 says move everything to the new URL, a 302 says keep the old one for now, a 404 or 410 says drop the page, and a 5xx says slow down and come back. Most problems I find are pages sending the wrong instruction, not broken servers.

    Every time Google requests a URL, the server answers with a three-digit number before it sends a single word of the page. That number decides what Google does next. It can tell Google to index the page, to look somewhere else, to forget the page exists, or to back off because the server is struggling.

    Site owners usually meet status codes as entries in a crawl report coloured red or amber. That framing makes them look like faults to clear. They are better understood as messages, and the useful question is whether each page is sending the message you meant to send.

    What Google does with each family

    Google publishes exactly how its crawlers handle the common codes in its documentation on how HTTP status codes affect Google’s crawlers. Stripped down to what matters for a business site:

    • 2xx: considered, not guaranteed. A 200 means Google takes the content into indexing. Google is explicit that a success code “doesn’t guarantee indexing.” If the page looks like an error or is empty, Search Console reports it as a soft 404 regardless of the code.
    • 3xx: go elsewhere. Google follows up to 10 redirect hops and ignores any content on the redirecting URL. A 301 or 308 is “a strong signal that the redirect target should be processed.” A 302 or 307 is a weak one.
    • 4xx: forget this page. Every 4xx except 429 is treated the same way. Already indexed URLs are removed, and Google crawls them less over time.
    • 429 and 5xx: slow down. Google reads these as an overloaded server and reduces its crawl rate. Indexed pages survive for a while and are eventually dropped if the errors persist.

    Permanent or temporary is a choice about which URL shows

    Google’s redirects documentation puts the difference in one line each. Permanent redirects “show the new redirect target in search results.” Temporary redirects “show the source page in search results.”

    That is the whole decision. If the old URL is never coming back, use a 301. If you are running a short promotion, sending traffic to a holding page during maintenance, or testing something you might undo, a 302 keeps the original URL in the index. The common mistake runs the other way: a site moves a page for good with a 302 because that is what the plugin defaulted to, and the old URL lingers in results for months.

    Google also ranks the ways of redirecting by how reliably it reads them. Server-side redirects come first. JavaScript redirects should only be used “if you can’t do server-side or meta refresh redirects.” If a developer suggests a JavaScript redirect because it is easier to deploy, that is a reason to push back.

    404 or 410 does not matter much to Google

    This one surprises people. Google’s documentation says all 4xx errors except 429 “are treated the same.” Both a 404 and a 410 tell Google the content does not exist, and both get the URL dropped. I would not spend developer time converting one to the other.

    What does matter is that a page that no longer exists actually returns one of them. The most expensive status code problem I find is the reverse: a missing page that returns 200 with “sorry, nothing here” in the body. That is a soft 404, and I cover why Google is usually right to call it one.

    5xx errors cost crawling before they cost rankings

    A server error does not knock pages out of the index straight away. Google slows down, and its documentation says the slowdown “is proportionate to the number of individual URLs that are returning a server error.” It speeds up again gradually once the errors stop.

    For a planned outage, a 503 is the right answer. It says “temporarily unavailable” in a way Google understands. A maintenance page served with a 200 tells Google your homepage now consists of the words “back soon.”

    The status code problems I actually find

    Rarely a broken server. Usually one of these:

    • Internal links pointing at redirects. Each one is a wasted hop and a sign the navigation was never updated. I found forty on my own site and wrote up the redirects nobody sees.
    • Sitemaps listing URLs that do not return 200. A sitemap should list pages you want indexed. Redirects and 404s in it tell Google the list cannot be trusted.
    • Not-found templates returning 200. Common on page builders and headless builds. Every mistyped URL becomes an indexable thin page.
    • A robots.txt that errors. Google reads a 4xx robots.txt as permission to crawl everything, and a 5xx as a reason to pause crawling the site until it can read the file again. HTTP Archive found robots.txt returning a 404 on 13% of sites in its 2025 crawl, which is usually harmless, but worth knowing about.

    How I check them

    A crawl of the site gives the status code of every internal link target. Search Console’s Page indexing report shows the codes Google itself received, which is the version that counts. When the two disagree, something is serving Google a different response from the one your crawler gets, and that is worth chasing before anything else.

    The short version

    1. Status codes are instructions. Ask whether each page sends the one you meant.
    2. 200 means considered, not indexed. Google still decides.
    3. 301 for permanent, 302 for temporary. The choice decides which URL shows in results.
    4. 404 and 410 are treated the same. Make sure dead pages return one of them.
    5. 5xx slows crawling first. Use a 503 for planned downtime.
    6. Most problems are links and templates. Not servers.

    If a site should be ranking and it isn’t, that’s the work I do. Technical SEO covers the status codes your pages send Google, and the ones they should be sending.

  • Crawling, rendering, indexing, ranking: which of the four stages is broken

    Crawling, rendering, indexing, ranking: which of the four stages is broken

    The verdict

    A page that gets no search traffic has failed at one of four stages: Google never crawled it, could not render it, chose not to index it, or indexed it and ranks it nowhere useful. Each stage leaves a different mark in Search Console. Find the stage before fixing anything, because the fixes do not overlap.

    Most “why isn’t this page ranking” questions are really a different question: where did the page drop out? A page that was never crawled needs links. A page that was crawled and not indexed needs a reason to exist. A page that is indexed and sits on page six needs a better answer than the pages above it. Those are three different jobs, and doing the wrong one wastes months.

    This is the order I work through on every page that should be getting traffic and isn’t.

    Google says three stages. I check four

    Google’s own guide to how Search works describes three stages, crawling, indexing and serving, and adds that “not all pages make it through each stage.” It also says plainly that Google “doesn’t guarantee that it will crawl, index, or serve your page,” even when the page follows every rule.

    I split rendering out as its own stage because that is where JavaScript sites fail without anyone noticing. Google’s guide folds rendering into the crawl: “Google renders the page and runs any JavaScript it finds using a recent version of Chrome.” When that goes wrong, the page is crawled, sometimes even indexed, and the content that should rank is simply not there.

    Stage one: was the page ever crawled?

    Google finds URLs in three ways: links from pages it already knows, sitemaps, and URLs it has seen before. A page with no internal links pointing at it, missing from the sitemap, depends on luck.

    The sign in Search Console is the URL Inspection tool saying the URL is not on Google, or the Page indexing report listing it as “Discovered, currently not indexed”, which means Google knows the address and has not fetched it yet. The fix is almost always internal links. I once audited a site where ninety per cent of pages were unreachable from the navigation, and no amount of content work would have helped until that changed.

    Crawl capacity is rarely the real constraint below a few hundred thousand URLs. I have written about why almost nobody has a crawl budget problem, and it is still the first wrong diagnosis I see in other people’s audits.

    Stage two: did Google see the content?

    Rendering problems hide well. The page loads fine in your browser, and the HTML Google receives is a shell that fills itself in with JavaScript. If a script fails, times out, or waits for a click, the text that should rank never reaches Google.

    The test takes two minutes. Run the URL through URL Inspection, open the tested page, and look at the rendered HTML and the screenshot. If the main copy, the product grid or the internal links are missing there, the page has a rendering problem, whatever else is wrong with it. I go through what this looks like on real builds in what Googlebot actually sees on a JavaScript site.

    Stage three: did Google decide to index it?

    Crawled is not indexed. Google’s guide describes how, during indexing, it groups similar pages and picks “the one that’s most representative of the group.” Every page that loses that vote is a duplicate in Google’s eyes, whatever you think of it.

    Status codes matter here too. Google’s documentation on how status codes affect its crawlers is blunt: “an HTTP 2xx (success) status code doesn’t guarantee indexing.” A 200 response gets the page considered. It does not get the page in.

    The Page indexing report shows the reason for each excluded URL. The two that matter most:

    • Crawled, currently not indexed. Google fetched the page and decided it was not worth a place. This is usually about the page, not the site’s technology, and I cover it in detail in what crawled, currently not indexed actually means.
    • Duplicate, Google chose a different canonical. Google grouped the page with another URL and picked the other one. Sometimes Google is right. Sometimes your own internal links and canonicals are telling it to.
    • Excluded by noindex. Less common than people fear. HTTP Archive’s 2025 crawl found a noindex directive on only 2.4% of mobile pages, but when it is on a page that should rank, it is usually a plugin or a staging setting nobody switched off.

    Stage four: indexed, so why no traffic?

    If URL Inspection says the page is on Google, the technical part is done. What is left is ranking, and ranking is about relevance and competition.

    Open the Performance report filtered to that one page. Three patterns cover most of what I see:

    • No impressions at all. The page is indexed but matches nothing people search for, or matches it so weakly it never appears. Look at what the page is actually about against the words people use.
    • Impressions at position 30 and beyond. Google considers the page relevant and not good enough. Compare it honestly with the pages in the top five.
    • Impressions in the top ten, almost no clicks. The page ranks and the result does not get chosen. I wrote up a case with 27,903 impressions and 106 clicks where this was the whole problem.

    The order matters more than the checks

    Every one of these checks is simple. The mistake is doing them out of order. Rewriting copy on a page Google has never crawled changes nothing. Building links to a page that renders empty sends authority to a blank. Fixing canonicals on a page that ranks fine and gets no clicks solves a problem the page does not have.

    So I start at stage one and stop at the first stage that fails. Whatever is wrong further down can wait, because Google cannot see it yet.

    The short version

    1. Four stages: crawl, render, index, rank. Google describes three. Rendering deserves its own check.
    2. Not crawled means not linked. Internal links fix this far more often than sitemaps do.
    3. Check the rendered HTML. If the content is missing there, nothing else matters yet.
    4. A 200 status does not mean indexed. Google still decides whether the page earns a place.
    5. Indexed with no traffic is a ranking problem. Read the page’s own Performance data.
    6. Stop at the first stage that fails. Fixes further down are wasted until it is solved.

    If a site should be ranking and it isn’t, that’s the work I do. Technical SEO starts by finding which of these four stages your pages are failing at.

  • Reciprocal links are not banned. Excessive link exchanges are

    Reciprocal links are not banned. Excessive link exchanges are

    The verdict

    Wikipedia says Google stopped crediting reciprocal links back in 2005. Google’s own spam policy is narrower: it names “excessive” link exchanges and partner pages built “exclusively for the sake of cross-linking”. Two businesses that genuinely work together linking to each other is normal. A partners page of forty strangers is the pattern Google describes.

    A supplier lists your business as a stockist. You list them as a supplier. That is a reciprocal link, and almost every small business has a few. A common question is whether they need to be removed. Usually they do not.

    What they should remove is something else, and the difference is one word in Google’s policy.

    Two descriptions of the same thing

    The Wikipedia article on link building says reciprocal linking “is no longer an important part of the search engine optimization process”, and that Google stopped giving them credit with an update in 2005.

    Google’s own spam policies put it more precisely. Among the examples of link spam they list: “Excessive link exchanges (“Link to me and I’ll link to you”) or partner pages exclusively for the sake of cross-linking.”

    The words doing the work are “excessive” and “exclusively”. A mutual link between two businesses that actually deal with each other is neither. A page whose only reason to exist is to trade links is both.

    What I leave alone

    • Suppliers, stockists and partners you actually work with. The link describes a real relationship and customers use it.
    • Associations and memberships. A trade body lists you, you show their badge and link back. That is normal.
    • Referral relationships. A chiropractor and a physiotherapist who refer patients to each other and say so on their sites.

    What I would remove

    • A partners or resources page of unrelated sites. If you would not send a customer to half the links on it, it is a link exchange page.
    • Exchanges arranged by email offers. “We’ll link to you if you link to us” from a site you have never heard of is the pattern the policy names.
    • Exchange networks. Rings of sites linking to each other in rotation. They are easy to see from outside, which is the problem.

    The test is the same one I use for any link: would you mention it to a customer? A supplier you stock, yes. A page of forty sites you have never dealt with, no.

    What removing them does

    Probably very little to rankings, in either direction. Google already discounts most of this. The reason to remove it is that it costs you a page of your own site linking out to things you would not recommend, and it tells anyone reading it, customer or reviewer, that the site was built for search engines.

    Link counts are the habit this whole cluster argues against. The case for reading the profile by quality instead is in why more pages and more links lost to fewer of both.

    The short version

    1. Reciprocal links between real partners are normal.
    2. Google names excessive exchanges and partner pages built only for cross-linking.
    3. Keep suppliers, associations and referral partners.
    4. Remove link-exchange pages and email-offer swaps.
    5. Would you mention it to a customer? The same test as every other link.
    6. Removing them is about the site, not the ranking.

    If a site should be ranking and it isn’t, that’s the work I do. Link building covers what to keep, what to remove and what to earn instead.

  • Your own practitioners can filter your firm out of the map pack

    Your own practitioners can filter your firm out of the map pack

    The verdict

    When several businesses in the same category share an address, Google tends to show one of them in the map pack and filter the rest. In a law firm or a clinic, the listings competing for that one slot are often your own practitioners. Decide which profile should win, give it the firm’s category, and do not create practitioner profiles by default.

    A firm with four lawyers creates a Business Profile for the firm and one for each lawyer, because somebody said practitioner profiles are free visibility. Same address, same building, often the same category. Then the firm’s own listing drops out of the map pack for its main search, and nobody can work out why.

    Usually nothing is wrong with the firm’s profile. It is being filtered, and the listings filtering it are its own.

    What the filter is

    When several listings in the same category sit at the same address, Google generally shows one and hides the others for a given search. The local search community gave the changes that shaped this behaviour nicknames. The Wikipedia article on local search optimisation records “Possum”, rolled out on 1 September 2016, as the point where listings with “the same address and same categories” began to be filtered “though they belong to different businesses”, and gives the example: “if several dentists share the same address, Google will only show one of them.”

    It records “Hawk”, on 22 August 2017, as narrowing that filter so it seemed to apply only to listings in the same building or very close by. These are community names for observed behaviour, not Google’s own, but the behaviour is easy to see for yourself: search the firm’s main service and watch which of your listings appears.

    What Google’s own rules say about practitioners

    Google’s guidelines for representing your business name “doctors, dentists, lawyers, financial planners, and insurance or real estate agents” as individual practitioners. A practitioner should only have their own profile if they work in a public-facing role and “can be contacted directly at the verified location during stated hours.” Support staff should not have one.

    Where there are several practitioners at one location, the organisation keeps its own separate profile, and each practitioner’s profile carries only their name, not the firm’s. Where a practitioner is the only public-facing one at a branded location, Google recommends a single shared profile in the format “[brand/company]: [practitioner name]”.

    Nothing in the guidelines tells you that practitioner profiles help the firm rank. They are permitted, not recommended.

    How I decide

    • Pick the profile that should win the main search. For most small firms and clinics that is the organisation, because it collects the reviews, the links and the website.
    • Give that profile the firm’s main category. If a practitioner profile shares the same primary category at the same address, the two are competing for one slot.
    • Only keep practitioner profiles that earn their place. A partner people search for by name, with their own reviews, is worth a profile. An associate added to fill the map is a listing that can only cost the firm.
    • Never use practitioner profiles to cover more specialisms. Google says a practitioner “shouldn’t have multiple Business Profiles to cover all of their specializations.”

    That first rule is a judgement, and I hold it for small practices. A large firm with well-known named partners may genuinely be better served by their individual profiles. The test is which listing people are actually searching for.

    The neighbour version

    The same filter applies between unrelated businesses. Two clinics in one medical building, or two firms on one floor, compete for one slot on searches where both are relevant. Distance cannot separate them, so relevance and prominence decide which is shown: the primary category, the reviews, and how clearly the website and the profile describe one business.

    That last point is why this sits next to the identity question. One business needs one consistent identity across the site and the profile, set out in why a business needs one identity everywhere Google looks.

    The short version

    1. Same address and same category usually means one listing shown.
    2. In a firm, your own practitioners can be the competition.
    3. Google permits practitioner profiles. It does not recommend them. Public-facing, reachable at the location, name only.
    4. Decide which profile should win the main search. Usually the organisation, for a small practice.
    5. Keep practitioner profiles that people search for by name.
    6. Neighbours are filtered the same way. Relevance and prominence decide who is shown.

    If a site should be ranking and it isn’t, that’s the work I do. Local SEO covers the profiles, the categories and which listing should be the one people see.

  • Your Business Profile already lists the searches that found you. Use them

    Your Business Profile already lists the searches that found you. Use them

    The verdict

    Your Business Profile’s performance report lists the search terms people used to find it, refreshed monthly. Most owners never open it. It is the cheapest local keyword research there is: it shows which services people look for, whether your categories match how they search, and which services deserve a page of their own.

    Every Business Profile has a performance report, and inside it is a list most owners have never scrolled to: the actual searches people typed when the profile appeared. It costs nothing, nobody has to estimate anything, and it describes your customers in their own words.

    What it is

    Google’s performance report documentation describes the searches section as “The search terms people used to find your business.” It is updated at the start of each month and can take up to five days to appear. A query shows up whenever the profile appeared for it, and you cannot edit or manage the list.

    The same report counts views as unique visitors, one per person per day, alongside calls, directions and website clicks. Those are the outcomes. The search terms are the explanation.

    What I look for in it

    • Services people search for that have no page. If a service appears in the terms month after month and the website has no page about it, that is the most direct case for a new service page you will ever get.
    • Services missing from the profile itself. The services list on the profile should be built from every service page on the site. Terms describing something the profile does not list are a gap on both sides.
    • Whether the primary category matches. If most discovery searches describe something other than the primary category, revisit the category. Google asks for categories that complete the sentence “this business IS a”, and the search terms show how customers complete it.
    • Brand against discovery. Searches for the business’s own name are people who already know it. Searches for the service are people finding it. A rising share of the second kind is a profile doing its job.
    • Terms you do not want. If the profile keeps appearing for a service you do not offer, something on it, often the description or a secondary category, is saying you do.

    The category point is a judgement I hold firmly: a primary category is a commercial decision about which searches you want to win, not a description of everything the business does. The search terms are the best evidence for making it.

    What it cannot tell you

    Where you rank. It lists searches the profile appeared for, not positions. For that you need a grid tracker, and I have written up how to read a grid rank tracker separately.

    It is also monthly and a little late, so it suits a quarterly review rather than a weekly one. Read three months together before deciding anything.

    The basics of checking where a profile stands, and which other performance numbers matter, are in how to check where your Business Profile ranks.

    The short version

    1. The report lists the searches that found you. Updated monthly, up to five days late.
    2. Services searched with no page are the clearest gap.
    3. Build the profile’s services list from the site.
    4. Let the terms test the primary category. “This business IS a” is Google’s test.
    5. Watch discovery searches grow against brand searches.
    6. It shows searches, not rankings. Read three months at a time.

    If a site should be ranking and it isn’t, that’s the work I do. Local SEO covers the profile and the pages behind it as one job, starting with what customers are already searching for.

  • A grid rank tracker draws a map, not a score. How I read one

    A grid rank tracker draws a map, not a score. How I read one

    The verdict

    A grid tracker searches your keyword from dozens of points and colours each one by where you rank. The single average rank the tool puts at the top hides the useful part. What I read is the shape: how far the top-three zone reaches, which side is weak, who holds the points you lose, and whether the zone grows month on month.

    A grid rank tracker is how I track local work for clients. It searches a keyword from a pattern of points across a map and shows where the business ranks at each one. The result looks like a heat map, strong near the address and fading with distance.

    It is the most useful local report there is, and the number most tools put at the top of it is the least useful thing on the screen.

    Why a grid at all

    Because one of Google’s three local ranking factors is distance, which its local ranking documentation defines as “how far each business is from the customer who’s searching.” There is no single ranking to report. There is a ranking at every point a customer might stand.

    Settings decide whether any of it is comparable

    • Fit the grid to the real service area. Too wide and every point near the edge shows the business missing, which is expected and tells you nothing. The grid should cover where customers actually are.
    • Centre it on the business. And keep it there. A grid that moves between scans is a different test.
    • Use searches customers type. Three to five of them: the main service, a secondary one, and the way people actually phrase it.
    • Same keywords, same time of day, every month. Results drift. Consistency is what makes this month comparable with last.

    The number at the top, and why I look past it

    Most tools summarise the scan as an average rank across all the points. It looks precise. It is heavily shaped by the points at the edge where the business does not appear at all, which tools usually score as a fixed low number. Shrink the grid and the average improves without anything changing on the ground.

    The figure I find more honest is the share of points where the business is in the top three. Tools name these figures differently, so check what yours is actually calculating before you report it to anyone.

    What I actually read

    • How far the top-three zone reaches. In each direction. This is the business’s real catchment in the map pack.
    • Which side is weak. A zone that stops sharply on one side usually has a reason there, often a competitor sitting closer to those customers.
    • Who holds the points you lose. Most tools show the top results at each point. If the winners are simply nearer, that is distance, and no profile work beats it. If they are further away and still ahead, they are winning on relevance or prominence, and that is something you can work on.
    • Whether the zone grows. Month on month, same settings. A slowly widening top-three area is what working on the profile looks like.

    Prominence is the factor with a shelf life. BrightLocal’s 2026 survey of 1,002 US consumers found 74% look for reviews written in the last three months, which is one reason a profile that has gone quiet can lose ground at the edges first.

    When the map improves and the phone does not

    Then the problem is probably not ranking. Look at what the profile and the website say to the person who found them: reviews, photos, hours, and where the website link lands. A grid tells you who can see the business. It cannot tell you why they did not call.

    The basics of checking a profile’s position, and the performance data to read alongside the grid, are in how to check where your Business Profile ranks.

    The short version

    1. One local rank does not exist. Distance makes it different at every point.
    2. Fit the grid to the real service area. Keep keywords, centre and timing fixed.
    3. Look past the average rank. Edge points and grid size distort it.
    4. Share of top-three points is more honest. Check what your tool calculates.
    5. Read reach, weak side, and who beats you. Nearer competitors are distance. Further ones are fixable.
    6. A better map with no more calls is not a ranking problem.

    If a site should be ranking and it isn’t, that’s the work I do. Local SEO includes grid tracking from where your customers actually search, and the work to move it.

  • Asking for a bought link to be removed: the request that gets answered

    Asking for a bought link to be removed: the request that gets answered

    The verdict

    Google’s own advice is to remove as many bad links as you can before disavowing anything. Removal is real; disavowal is a request. The emails that get answered are short, name the exact page, and come from the business that paid. Keep a dated log of every request, because a reconsideration request needs it.

    If a business paid for links and they are still live, the strongest move is usually to have them taken down at the source. A disavow file tells Google to ignore a link. Removal means it is gone.

    Google says the same thing, in this order. Its disavow documentation opens its advice with: “First and foremost, we recommend that you remove as many spammy or low-quality links from the web as possible.”

    Who to ask

    • The vendor that placed them. If an agency or a link seller arranged the placements, they usually hold the contacts, and sometimes the logins. Start there.
    • The sites themselves. For anything the vendor will not or cannot remove.
    • Directories with an account. Many let you log in and delete the listing yourself. No email needed.

    Do not email sites you never dealt with. Scrapers and random low-quality pages linking to you are not something you arranged, and Google is generally good at ignoring them. A removal campaign aimed at them is a lot of email for nothing.

    The email that works

    • From the business, not an agency. A request from the company’s own domain about its own link reads as legitimate. A bulk request from a third party reads as spam.
    • One page, named exactly. The full URL of the page carrying the link, and the link itself.
    • Offer the alternative. If they would rather keep the article, ask them to mark the link rel="sponsored". Google’s spam policies accept paid links qualified that way, so it fixes the problem without deleting their page.
    • No threats. Legal language gets ignored or escalated. A polite, specific request gets acted on.
    • One follow-up. After a couple of weeks. Then stop and move it to the disavow list.

    Some sites ask for a fee to remove a link. My view is that I would not pay it. Put that URL in the disavow file instead, with a comment saying so.

    Keep a log

    A simple sheet: the linking page, the date you asked, who you asked, any reply, and the outcome. Two reasons it matters.

    The unanswered rows become the disavow list, already annotated. And if the links led to a manual action, Google’s manual actions documentation asks a reconsideration request to explain the issue, describe the steps taken and document the outcome. The log is that document.

    The file itself is set out in how to build a disavow file that only does what you meant. Whether you need either is the question in when a disavow file is worth using at all.

    The short version

    1. Remove first, disavow second. Google’s order, and the right one.
    2. Start with whoever placed the links.
    3. Ignore links you never arranged. Google already discounts most of them.
    4. One page per email, from the business. Offer rel=”sponsored” as the alternative.
    5. Do not pay removal fees. Disavow that URL instead.
    6. Log every request. It becomes the disavow list and the reconsideration record.

    If a site should be ranking and it isn’t, that’s the work I do. Link building covers cleaning up what was bought and earning what replaces it.

  • How to build a disavow file that only does what you meant

    How to build a disavow file that only does what you meant

    The verdict

    If you have a real reason to disavow, the file is the easy part, and still the place it most often goes wrong. It is a plain text list, one URL or domain per line, and every upload replaces the one before. The two mistakes that do damage are domain lines nobody checked and a new file quietly wiping out an old one.

    I almost never recommend a disavow file now. When one is justified, it is because there is a manual action for unnatural links, or because a business knows it bought links that are still live and cannot get them removed. In both cases the file has to be right, because there is no feedback afterwards to tell you it was wrong.

    The format

    Google’s disavow documentation is short and strict.

    • A plain text file. Encoded in UTF-8 or 7-bit ASCII, with a name ending in .txt.
    • One entry per line. Either a full URL, or a whole domain written with the prefix domain:.
    • Comments allowed. Any line starting with # is ignored by Google and is for you.
    • Limits. 100,000 lines, including blank and comment lines, and 2MB.

    A small one looks like this:

    # Paid placements from a 2021 campaign. Removal requested, no reply.
    domain:example-directory.com
    https://example.com/2021/best-plumbers-in-town/
    https://example.net/sponsored/roof-repair-guide/

    The two decisions that actually matter

    URL or domain. A domain: line throws away every link from that site, including any genuine one. Use it only when the whole site exists to sell links. Where a site gave you one paid link and one real mention, list the paid URL and leave the domain alone. It is one of the commonest ways a disavow file removes something that was helping.

    Replace, not add. Google is explicit: “Uploading a new list for a property will replace any existing list for that property.” If somebody uploaded a file two years ago, a new upload deletes it. Download the existing list first, merge, then upload the merged version. If the new file has errors, Google keeps the old one, which is the only safety net there is.

    Three practical details

    • Domain properties are not supported. The tool works on a URL-prefix property. If Search Console only has a domain property, add the https version of the site as a URL-prefix property first.
    • It takes weeks. Google says it “can take a few weeks” to incorporate a list. Nothing confirms when it has.
    • Google’s own warning applies. It calls the tool “an advanced feature” that “can potentially harm your site’s performance” if used incorrectly.

    Use the comments as a record

    A line with no comment is a decision nobody can explain later. I write the reason and the date above each group: which campaign, when removal was requested, whether anyone replied. If the file is part of answering a manual action, that record is most of what the reconsideration request needs, because Google asks you to explain the issue, the steps taken and the outcome.

    Removal comes before any of this. I have written up how to ask for a bought link to be taken down separately. And whether you should be building a file at all is the earlier question, set out in when a disavow file is worth using at all.

    The short version

    1. Plain text, one entry per line. UTF-8, .txt, comments start with #.
    2. 100,000 lines and 2MB at most.
    3. Use domain: only for sites that exist to sell links. Otherwise list the URL.
    4. Every upload replaces the last. Download, merge, then upload.
    5. URL-prefix property only, and weeks to take effect.
    6. Comment every group with a reason and a date. It doubles as the reconsideration record.

    If a site should be ranking and it isn’t, that’s the work I do. Link building covers the off-site half, including when to clean up and when to leave a profile alone.