Tag: ai-overviews

  • Keyword research that starts with intent, not the volume column

    Keyword research that starts with intent, not the volume column

    You open a keyword tool. You sort by search volume. You start at the top.

    Two months later you’ve got a handful of posts chasing big, fuzzy terms, and almost none of them have brought in a single lead.

    It’s a really common trap, and it catches smart teams too. The problem isn’t effort. It’s that the volume column answers the wrong question first.

    So in this guide, I’ll show you the keyword research process I actually use. Start with the person and the job they’re trying to get done. Map the intent. Only then check demand. You’ll get a step-by-step workflow, free tools, a scoring table you can copy, and a look at why all this matters even more now that Google AI Overviews, AI Mode, and ChatGPT answer questions before anyone clicks.

    Why the volume column misleads you

    Search volume feels objective. It’s a number, and numbers feel safe.

    But volume only tells you how many times a phrase was typed. It doesn’t tell you who typed it, what they wanted, or whether your page could ever be the best answer.

    That’s a lot of missing information.

    Most searches live in the long tail

    Here’s the part most people miss. Ahrefs reports that keywords with fewer than 10 searches a month make up almost 93% of its U.S. keyword database.

    Ninety-three percent. Read that again.

    Google has also said for years that a big chunk of what people type is brand new. In its 2019 post on BERT, Google wrote that 15 percent of the queries it sees every day are ones it hasn’t seen before.

    So if you only work from what a tool can show you, you’re working from a partial map.

    Your tools hide the smallest, most specific queries

    Even Google’s own tool has blind spots. The Keyword Planner help page notes that keywords with very low search volume aren’t discoverable or forecastable.

    That’s fine for planning ads. For SEO, though, those “invisible” queries are often the ones sitting closest to a buying decision.

    AI answers change which keywords are worth a page

    Now for the new wrinkle.

    Pew Research Center analyzed the browsing data of 900 U.S. adults and found that Google users clicked a traditional result in 8% of visits when an AI summary appeared, versus 15% when one didn’t.

    The same study found AI summaries showed up for just 8% of one- or two-word searches, but for 53% of searches with 10 or more words. And 60% of queries starting with “who,” “what,” “when,” or “why” triggered one.

    Put simply: the long, question-shaped searches that intent-based research digs up are exactly where AI answers now live. You still want those topics. You just have to plan the page so it’s worth citing and worth clicking.

    What “search intent” really means

    Search intent is the job behind the search. What’s the person trying to get done when they type it?

    Most SEOs sort intent into four buckets. Here’s a quick reference you can keep open while you work.

    IntentWhat the searcher wantsExample queryPage that usually fits
    InformationalTo learn or understand somethinghow does schema markup workGuide, explainer, how-to
    CommercialTo compare options before buyingbest SEO audit tools for small sitesComparison, review, “best of” list
    TransactionalTo act or buy nowhire technical SEO consultantService page, product page, pricing
    NavigationalTo reach a specific brand or pageseopya contactHome, contact, or brand page

    Real searches often blend two of these. “Technical SEO audit cost” is part research, part shopping. That’s fine. What matters is that you decide which job your page will do best.

    My intent-first keyword research process

    Here’s the whole workflow, start to finish. It works for a five-page local business site, and it works for a big content program.

    Step 1: Write down the jobs your customers hire you for

    Before you open any tool, list the problems your best customers bring to you.

    Not keywords. Problems. Something like:

    • “Our traffic dropped after a Google update and we don’t know why.”
    • “We rank on Google, but we never show up in ChatGPT answers.”
    • “Our Amazon listings get views but don’t convert.”

    Pull these from sales calls, support tickets, onboarding forms, and the questions you get asked over and over. Shoot for 10 to 20 jobs.

    Step 2: Collect the questions people actually ask

    Now turn each job into the words people use. This is where you go wide.

    Good sources, most of them free:

    • Google’s own results. Search your seed term and write down the People Also Ask questions, related searches, and autocomplete suggestions.
    • Google Search Console. Your Performance report shows real searches that already bring up your pages.
    • Google Trends. See whether interest is rising, falling, or seasonal.
    • Question tools like AlsoAsked and AnswerThePublic, which show how questions branch off each other.
    • Communities. Reddit threads, Quora, industry Slack groups, and product reviews give you the exact phrasing people use when they’re frustrated.
    • Your own inbox. Real customer emails. Honestly, the most underrated keyword source there is.

    Step 3: Mine Search Console for long, conversational queries

    If your site already gets some traffic, this step is a gold mine.

    1. Open Search Console and go to Performance, then Search results.
    2. Set the date range to the last 3 months.
    3. Click + Add filter, choose Query, and switch the match type to Custom (regex).
    4. Paste a question pattern like ^(how|why|what|which|where|when|can|is|are|does|should) (the same pattern Ahrefs suggests in its long-tail guide).
    5. Sort by impressions and look for queries where you’re getting impressions but sitting past page one.

    Those are questions Google already connects with your site. A focused section or a dedicated page can often push them up.

    Step 4: Label every query with its intent

    Go back to your list and tag each query with one of the four intents from the table.

    The fastest, most honest way to do it? Look at the live results. Search the query and ask yourself:

    • Are the top results guides, comparison posts, product pages, or local listings?
    • Is there an AI Overview? A map pack? Videos? A shopping carousel?
    • Would the page you’re planning look like it belongs on this results page?

    If Google shows ten service pages and you were planning a 3,000-word guide, the results are telling you something. Listen to them.

    Step 5: Group queries into topics, not single keywords

    This is where most keyword research goes sideways. People build one page per phrase.

    Don’t.

    Instead, cluster queries that share the same intent and would be answered by the same page. “How long does SEO take,” “SEO timeline,” and “when will I see SEO results” belong on one page, not three.

    The quick rule I use: if the same pages rank for two queries, they probably belong to one topic.

    This matters for AI search too. Google says AI Mode uses a “query fan-out” technique that breaks a question into subtopics and runs many searches at once. A page that covers a whole topic cluster gives those sub-queries more to find on one URL.

    Step 6: Only now, check demand and difficulty

    Now you open the volume column. But you use it as a tiebreaker, not a starting point.

    For each topic cluster, add up the estimated volume of the queries in it. Then look at how tough the current results are. Huge brands? Or forums and thin pages you could beat?

    Tools like Ahrefs, Semrush, and Google’s Keyword Planner all give you estimates. Treat them like estimates. Different tools often disagree, so compare the trend and the rough size rather than chasing a precise number.

    Step 7: Score and prioritize your topics

    Last step. Score every topic so you know what to write first. Here’s the simple scoring table I use. Copy it into a spreadsheet.

    FactorQuestion to askScore 1 to 3
    Business fitDoes this topic lead naturally to what we sell?3 = directly, 1 = loosely
    Intent matchCan we build the type of page the results reward?3 = yes, 1 = not really
    ExperienceDo we have first-hand expertise or examples to add?3 = deep, 1 = none
    DifficultyCan we realistically compete with the current results?3 = yes, 1 = unlikely
    DemandIs there meaningful or growing interest?3 = strong, 1 = minimal

    Add up the scores and start with the highest totals. You’ll notice business fit and experience often outweigh raw demand.

    That’s the whole point.

    A worked example: an independent bike shop

    Let’s make this concrete with a made-up example. Say you run an independent bike repair shop.

    The volume-first approach would go after something like “bikes.” It’s huge, it’s vague, and it’s owned by national retailers and brand sites. Good luck.

    The intent-first approach starts with jobs:

    • “My gears keep slipping and I don’t know why.”
    • “I need a tune-up before a charity ride.”
    • “Is it worth fixing my old road bike or should I replace it?”

    Those turn into topic clusters like:

    1. Gear and shifting problems (informational): a troubleshooting guide on slipping gears, noisy chains, and when to bring the bike in.
    2. Bike tune-up (transactional, local): a service page with what’s included, pricing, turnaround time, and booking.
    3. Repair or replace (commercial): an honest guide with a simple decision checklist.

    Each cluster gets one strong page, and each page links to the next step. The guide links to the tune-up page. The tune-up page links to booking.

    None of these topics would win a volume sort. All three are a lot closer to revenue.

    How to turn your research into pages that win clicks and citations

    Great research is wasted if the page doesn’t match it. Here’s how I turn a topic cluster into an actual page.

    Answer the main question early

    Put a clear, direct answer to the core question in the first few paragraphs. Then go deeper.

    The reader who wants the short version gets it. And the page becomes easier for Google’s AI features and other answer engines to quote accurately.

    Use the real questions as subheads

    The questions you collected in Step 2 make great H2s and H3s. They mirror how people search and how AI systems break questions apart.

    Then make your title tag match the main intent too. My guide to writing title tags Google will actually use walks through that part.

    Add what only you can add

    Google’s guidance on helpful, people-first content asks whether content provides “original information, reporting, research, or analysis” and “insightful analysis or interesting information that is beyond the obvious.”

    That’s where your experience comes in. Real photos. Your process. The mistakes you see clients make. Specific numbers you can stand behind. I go deeper on this in E-E-A-T in practice.

    Connect the cluster with internal links

    Every page in a cluster should link to the others, and to the page where the reader takes the next step. That’s how topical depth becomes visible to people and crawlers.

    Want the full framework? Read internal linking as a system, not an afterthought.

    Make it easy for AI systems to cite you

    The intent-led pages you build here are also the ones most likely to get pulled into AI answers. Google’s documentation on AI features and your website says there are no extra technical requirements to appear in AI Overviews or AI Mode beyond being indexed and eligible for a snippet.

    So the fundamentals carry you most of the way: crawlable pages, clear answers, accurate facts, and a recognizable brand. My guide on how to rank better in LLMs and AI search covers the rest.

    Keyword research mistakes to avoid

    These are the shortcuts I see most often, and what I’d do instead.

    • Starting with the biggest number. Start with the job. Use volume to break ties.
    • One page per keyword. Build one strong page per topic cluster and intent.
    • Ignoring zero-volume queries. If customers ask it, it’s real demand, even when a tool says zero.
    • Guessing intent. Look at the live results before you write a word.
    • Treating tool numbers as facts. They’re estimates. Compare trends and rough size.
    • Researching once and never again. Revisit your list every quarter. Search behavior keeps shifting, especially with AI features in the mix.

    Your intent-first keyword research checklist

    Run through this before you green-light any new page:

    • We know which customer job this page serves.
    • We collected real questions from at least three sources.
    • We checked the live results and labeled the intent.
    • The page covers a topic cluster, not a single phrase.
    • We scored the topic on business fit, intent match, experience, difficulty, and demand.
    • The main question gets a direct answer near the top.
    • The page links to related posts and to a clear next step.

    How this shows up when we work together

    At Seopya, keyword research is never a spreadsheet we hand over and walk away from. We start from your customers’ jobs, map intent against the live results in Google and in AI answers, and build a content plan that ties back to the pages where you actually make money.

    If you’d like to see how that fits with technical work, content, and AI search visibility, take a look at our SEO and AI search services.

    Key takeaways

    • Search volume tells you how often something is typed, not why. Start with intent.
    • Most searches live in the long tail. Ahrefs puts keywords under 10 monthly searches at almost 93% of its U.S. database.
    • Long, question-style searches are where AI summaries show up most, according to Pew’s 2025 analysis. Plan those pages to be cited and clicked.
    • Cluster queries into topics and build one strong page per cluster.
    • Score topics on business fit, intent match, experience, difficulty, and demand, roughly in that order of importance.

    Frequently asked questions

    Is search volume still useful?

    Yes, as a tiebreaker. Use it to compare topics you’ve already decided fit your business and intent, and to spot rising or seasonal interest. Just don’t let it pick your topics for you.

    How many keywords should one page target?

    Think in topics, not counts. One page should go after one main intent and the cluster of closely related searches that share it. If two queries need different kinds of pages, split them.

    What free tools are best for intent-based keyword research?

    Start with Google itself (People Also Ask, related searches, autocomplete), Google Search Console, and Google Trends. Add a question tool like AlsoAsked, and read the community threads where your customers hang out.

    Does keyword research still matter with AI search?

    More than ever. AI answers lean on pages that answer specific questions clearly. Pew found that longer, question-style searches are far more likely to trigger an AI summary, and those are exactly the searches intent-first research uncovers.

    How often should I redo my keyword research?

    Review your priority topics every quarter, and check Search Console monthly for new queries. Big shifts, like a core update or a new AI feature, are a good excuse to look sooner.

    Where to start this week

    Keyword research isn’t a hunt for the biggest number. It’s finding the questions your customers are already asking, then building the page that answers them best.

    If you only do one thing this week, do Step 1. Grab a doc and write down ten problems your customers have brought you in the last few months. You’ll probably find your next three blog posts hiding in that list.

    If you’d like help building an intent-first content plan for your business, take a look at our SEO and AI search services or contact me and we’ll talk it through.

  • How to rank better in LLMs and AI search

    How to rank better in LLMs and AI search

    For about a decade, SEO meant one thing for most of us: getting found in a list of blue links.

    That work still matters. But here’s what changed. A lot of people now get an answer first, from Google AI Overviews or AI Mode, from ChatGPT, from Gemini, from Perplexity. Only after that do they decide which source to open, if they open one at all.

    So you can rank fine in classic search and still be missing from the answer. Frustrating, right?

    Let’s clear something up first. “Ranking” in a large language model isn’t a position you can buy. It’s closer to this question: when a system needs a source, is your page easy to find, easy to read, and specific enough to quote without twisting what you said?

    Nobody can guarantee you a citation. Anyone who says they can is selling something. What you can do is remove the reasons you’d get skipped.

    That’s what this post is about. I’ll cover entity clarity, writing that’s worth citing, honest structured data, brand mentions, crawler access for AI bots, and a format that works for people and machines.

    What “ranking” in LLMs can reasonably mean

    Google is the clearest official voice on its own AI features. Its documentation on AI features and your website says there are no extra technical requirements to show up in AI Overviews or AI Mode. No special schema either. A page just has to be indexed and eligible to show in Google Search with a snippet.

    The same document repeats the usual SEO basics:

    • Allow crawling.
    • Use internal links.
    • Put important information in text.
    • Keep structured data consistent with the visible page.
    • Keep your Business Profile or Merchant Center info current, if those apply.

    And inclusion still isn’t guaranteed.

    ChatGPT, Gemini (the app), and Perplexity are different animals. Depending on the mode and the question, they may pull live sources, lean on training data, or do both. What you can influence is more down to earth: Can the right crawler reach your page? Is the page specific and clearly about a named entity? Can your claims be quoted accurately? Is your brand referred to the same way around the web?

    If someone promises you a reliable “LLM rank,” don’t take it on faith. I wouldn’t.

    Make the entity obvious

    Systems retrieve and cite more reliably when they can tell who you are, what the page is about, and how it fits with the rest of your site. That’s traditional entity SEO. And it’s become more important, not less.

    Here’s what that looks like day to day. Use one consistent name for your organization, your authors, and the products or services you write about. Your homepage, About page, bylines, and contact info should all agree. If you work in a specific place, say so in plain words, not only in the footer. Put the subject in the title, the H1, and the first couple of sentences.

    Ambiguity costs you. If a system can’t tell whether “Mercury” is a company, a planet, or a metal, it’ll go find a less confusing source.

    Thin, interchangeable posts don’t travel well in AI search either. When I read a page that could’ve been written about any brand in the category, I always ask: what would a model even hold on to here? A page that describes an actual method or a first-hand observation gives it something real. That’s the page that gets pulled.

    Write so a system can cite you without distorting you

    Answer engines like passages they can lift without having to fix the meaning. That changes how you write. What you target is only half of it.

    • Lead with the answer. Open important sections with a direct answer, then add the conditions and detail. A heading that’s a real question, followed by a plain answer, is much easier to retrieve than a slogan.
    • Make claims concrete and bounded. “Meaningful SEO trust often takes six to eight months of persistent work” is something a reader can use. “We deliver explosive growth” isn’t.
    • Credit your sources. If a fact isn’t yours, link to the official or named source. If you can’t name a source, don’t make up a statistic. Ever.
    • Keep the text complete. Google’s AI features guidance says plainly that important content should be available as text. Don’t bury the answer in an image or a tab that never renders as HTML.
    • Date your updates. When the substance changes, update the page and say what changed.

    Bonus: the same habits help your classic snippets and title tags. Two birds.

    Structured data, used honestly

    Structured data doesn’t come with a secret AI-ranking bonus. Google says you don’t need special schema to show up in AI Overviews or AI Mode, and that any structured data you do use should match the visible text on the page.

    So use schema when it accurately describes something already on the page: an Organization, a Person, an Article, a LocalBusiness, a Product, or an FAQ people can actually see.

    Don’t mark up reviews you don’t have. Don’t credit authors who didn’t write the piece. Don’t add FAQ markup for questions that never appear on the page. Inaccurate markup isn’t a shortcut. It’s a trust problem waiting to happen.

    When your Organization name, logo, and official profiles line up in the markup and on the page, you’ve given systems a cleaner identity to attach mentions to. That’s useful. It still doesn’t replace being mentioned and linked as a real source.

    Brand mentions, and the sites that already cite you

    Classic SEO taught us to care about links. They still matter.

    But unlinked mentions and consistent naming help too. They help a model connect a claim to a brand, especially when those mentions show up on sources the system already treats as reliable.

    Can you fake this? Not usefully. A burst of identical bios on low-quality directories won’t make you the go-to source for a technical question.

    What helps, slowly, is the same work that has always built authority. Publish pages other practitioners would actually cite. Be accurate when someone quotes you. Show up in places your buyers already trust.

    If your public footprint is thin, start by making the official record consistent. Then write a few genuinely specific pages instead of a big stack of interchangeable ones.

    Give AI crawlers a clear, deliberate path

    Retrieval requires access. This is the most technical part of the whole job, and it’s the part I see sites get wrong by accident more than any other.

    OpenAI documents its crawlers separately. OAI-SearchBot is used to surface websites in ChatGPT’s search features. Sites that opt out won’t show up in ChatGPT search answers, though they can still appear as navigational links. GPTBot is the one used to crawl content that may go into training generative AI foundation models.

    Those two controls are independent. Keep that in mind.

    ChatGPT-User is a fetcher that kicks in when a user asks for something. It’s not an automatic crawler in the same sense. OpenAI also notes a robots.txt change can take about a day to be reflected, and firewalls can block crawlers even when robots.txt lets them in.

    PerplexityBot is used to surface and link websites in Perplexity results, and according to Perplexity, it isn’t used to crawl content for foundation models. Perplexity-User handles user-initiated fetches and generally ignores robots.txt, since a person asked for the page.

    Google’s Search AI features rely on ordinary Search eligibility. That means Googlebot has to be able to crawl and index the page. Google also documents other product tokens, including Google-Extended, in its crawler overview. Those tokens don’t replace being indexable in Search.

    Here’s the audit I run:

    1. Read your robots.txt. Make sure you haven’t disallowed OAI-SearchBot, PerplexityBot, or Googlebot on the URLs you want visible, unless that’s a deliberate policy.
    2. Check your CDN, WAF, and bot-challenge rules. A 403, a JavaScript challenge, or a login wall can shut out a crawler that robots.txt allows. When a provider publishes IP ranges, use those official lists, not a one-time sample from your logs.
    3. Compare what the crawler sees with what a user sees. The crawler should get the same main content, not a blank shell that only fills in after a client-side app loads.
    4. Decide training access separately from search access. And write the decision down.

    That third one catches more sites than you’d think.

    What not to do

    Don’t buy “submit your site to 50 LLMs” services. Don’t paste a hidden block of machine-only answers at the bottom of every page. There’s no evidence it helps, and it tends to create thin, duplicate content.

    Don’t confuse blocking GPTBot with blocking all visibility in ChatGPT. If you want to appear in ChatGPT search, the crawler that matters is OAI-SearchBot.

    Don’t invent statistics, fake expert quotes, or schema that doesn’t match the page.

    And don’t assume you need a special file or a “GEO score” for Google AI Overviews or AI Mode. Google has said the opposite.

    How this work shows up in a real engagement

    When I work on AI search visibility at Seopya, I start with the same foundation I’d use for classic SEO. Are the important pages indexable? Internally linked? Written with real specificity?

    Then I check crawler access, entity consistency, and whether the pages are actually citable. It all lives inside our SEO and AI search services.

    Here’s the blunt version: if the site can’t be crawled or the content is interchangeable, no amount of “LLM optimization” will make it a preferred source.

    Key takeaways

    • You can’t buy a reliable spot inside ChatGPT, Gemini, Perplexity, or Google AI Overviews. You can make your pages easier to retrieve and safer to quote.
    • For Google AI Overviews and AI Mode, the official requirement is ordinary Search eligibility. No special schema.
    • Be consistent about who you are, write bounded claims, keep structured data in line with visible text, and allow the crawlers that power search visibility.
    • Mentions and citations around the web still help, as long as they’re real.

    FAQ

    Is GEO different from SEO?

    Generative engine optimization is a handy name for the work of becoming a source in AI answers. It doesn’t replace technical SEO, on-page clarity, or earned authority. For its own AI features, Google treats SEO fundamentals as the relevant work.

    Should I block AI crawlers?

    That’s a policy decision, not a default. If you want to be eligible for ChatGPT search or Perplexity results, allow the search crawlers those companies document. Training crawlers can be a separate call. Make the decision on purpose, then check that your security tools actually honor it.

    Does structured data get me cited by ChatGPT or Perplexity?

    There’s no official structured-data requirement that guarantees a citation in those products. Accurate markup can make your entity easier to interpret, especially for Google. It won’t make up for a page a crawler can’t reach, or claims too vague to quote.

    The next step

    Pick five pages that should already be sources. Check each one: Is it indexable? Internally linked? Clear about the entity? Does it give a direct answer? Is the schema consistent (if you use any)? Can the crawlers you want to allow actually reach it?

    Fix those five before you publish ten more.

    If you’d rather have help, get in touch. I can tell you what an AI search pass on your site would involve.

  • Local SEO in the AI era: being found beyond the 3-pack

    Local SEO in the AI era: being found beyond the 3-pack

    Local search used to be pretty simple. A map pack, a few organic results underneath, done.

    Not anymore.

    People still search for a business “near me.” But now they also ask Google AI Overviews and AI Mode, ChatGPT, Gemini, and Perplexity which plumber to call, whether a restaurant is open late, or what they should know before walking in. And whatever answer they get is only as good as the public record those systems can reach about you.

    Here’s the good news. The fundamentals haven’t been replaced. NAP consistency, a complete Google Business Profile, and local pages that describe real services in a real place still matter. What’s changed is how many surfaces read that same record, and how much it costs you to be vague or inconsistent.

    So in this post, I’ll walk you through the local work that still has to be done well, and how to make it useful in a world of AI answers. We’ll cover NAP hygiene, your Business Profile, local pages and structured data, and the mentions that help a system tell your business apart from a similarly named one down the street.

    The local record is now read by more than the 3-pack

    Google has described local results for a long time as a mix of relevance, distance, and prominence.

    You can’t do much about distance. (Moving your shop isn’t an SEO tactic.) What you can control is whether your information is accurate, whether your categories and services match what people actually ask for, and whether the rest of the web refers to your business the same way.

    AI features sit right on top of that same record. Google’s documentation on AI features and your website says there are no extra technical requirements for AI Overviews or AI Mode. It also includes a very practical reminder: keep your Business Profile information up to date.

    A page still has to be indexed and eligible to show in Search with a snippet before it can appear as a supporting link. And for a lot of local queries, the Business Profile itself is part of what shoppers and systems are looking at.

    Other answer engines will pull from your website, your profile, directories, and reviews in ways you can’t fully see. So what’s the right response? It isn’t some special “local GEO” file. It’s making the official record consistent and your on-site pages specific enough to quote.

    NAP consistency, without the folklore

    NAP stands for name, address, and phone number.

    Consistency means the same public business name, the same storefront or service address, and the same phone number, written the same way, on your website, your Google Business Profile, and every listing you actually control.

    I’ll be honest. This is boring work. It’s also still the foundation, and I check it first on every local project.

    • Pick one public name and stick with it. If the sign on the door, your homepage, and your Business Profile all say slightly different things, you’re asking every system to guess. They won’t always guess right.
    • Put your full address and phone number in HTML text. On a contact or location page. Not only inside an image or a widget.
    • Use one address format. And if you’re a service-area business, don’t invent a street address. Follow Google’s rules on hiding or showing your address when you don’t serve customers at a storefront.
    • Update everything together. When you move or change your number, fix the site, the profile, and your listings in the same sitting.
    • Skip the bulk citation packages. A smaller set of accurate listings on real local and industry sites beats hundreds of messy, low-quality ones.

    Got more than one location? Give each one its own clear NAP and its own profile. And don’t mix addresses on the location pages.

    Google Business Profile: the details that still do the work

    For most local businesses, the Google Business Profile is the most important listing you have. It’s also one of the sources Google specifically tells you to keep current if you care about Search, AI features included.

    When I go through a profile, this is what I’m checking:

    • Primary and additional categories. Pick the category that best describes what you are. Add others only when they’re accurate. Categories aren’t a keyword field to cram things into.
    • Services and products. List what you really offer, in the words customers use. A half-empty list hides you from relevant searches and gives an assistant less to go on.
    • Hours, attributes, and special hours. Showing “closed” when you’re open (or forgetting holiday hours) costs you visits. It also chips away at trust.
    • Photos and description. Show the real place and the real work. Write a description that matches your website, not a paragraph of keywords.
    • Posts, Q&A, and reviews. Use posts and Q&A when you’ve got something real to say. Reply to genuine reviews. Don’t try to manufacture them.

    Google lays out its Business Profile policies in the Google Business Profile help center. The job here is accuracy and completeness. There’s no trick that “gets you into the 3-pack,” and pack placement isn’t something I can (or will) promise anyone.

    Local pages, structured data, and pages an AI system can cite

    Your Business Profile doesn’t replace a good local page on your own site.

    The page is where you explain who you serve, what you do, where you work, and what a customer should do next. It’s also the page you can make eligible for regular search results and for supporting links in AI features. You own it. Use it.

    A few things I look for:

    • Write for a real person in that area. Name the area naturally, describe your services in specific terms, and include your NAP in text. Please don’t spin up a “unique” city page for every nearby ZIP code if you’ve got nothing unique to say about it.
    • Answer the questions people ask before they call. What’s included, how appointments work, which areas you cover, and any limits they should know about.
    • Use real photos. Your actual location or your actual work, with honest alt text.
    • Keep the important stuff in visible text. If a system can’t read your hours, service area, or phone number from the HTML, it can’t cite them reliably.

    If you use structured data, be honest with it. LocalBusiness markup should match the visible name, address, phone number, hours, and URL. Don’t mark up reviews you don’t have. Don’t list a different address than the one on the page. Markup won’t make up for an incomplete profile or a thin, templated city page.

    One more thing. Your location page should be reachable from the main navigation or a locations hub, and from related service pages. A local page that’s only linked from the footer is easy to ignore, for people and for crawlers.

    Mentions, reviews, and being identifiable

    Local prominence has always been about more than your own site. Mentions in local publications. Membership in real associations. Sponsorships. Reviews from people who actually used you.

    In an AI-driven web, there’s an extra benefit: consistent naming across all those sources helps a model tell your business apart from another one with a similar name. If you’ve ever been confused with a competitor, you know why this matters.

    You can’t fake this with identical directory bios. What you can do:

    • Keep your official name consistent everywhere you control the listing.
    • Ask happy customers to leave a genuine review on the platform they already use, without incentives that break the platform’s rules.
    • Go after mentions that would make sense even if search didn’t exist. A local organization, a trade group, a real news item.
    • Make your About and contact pages match your profile, so a system checking more than one source finds the same business every time.

    It’s slow work. A new city-page template won’t replace it.

    What not to do

    Don’t build dozens of near-duplicate “city pages” that swap the place name and leave everything else identical. That isn’t local content. It’s also a weak source for any answer engine.

    Don’t stuff categories, reviews, or map embeds onto a page to look more local than you are. Don’t create a Business Profile for a location that doesn’t exist, or set up a service area in a way that breaks Google’s rules on how businesses can represent themselves.

    Don’t buy reviews or citations in bulk. Looking busy for a few months isn’t worth the long-term trust hit.

    And don’t assume you need a special schema or a “local AI” plugin to show up in AI Overviews. Google has said the opposite: keep your profile current and keep your site eligible for Search.

    How this work shows up in a real engagement

    When I take on local SEO at Seopya, I start with the record. That means the NAP on your website, your Business Profile, and the listings you already have.

    I fix the disagreements first. Then I improve the profile and the location pages so a person (and a retrieval system) can actually use them. All of this is part of our SEO and AI search services.

    What I won’t do is promise you a 3-pack spot or an AI Overview citation. What I will give you is a straight account of what’s inconsistent, what’s thin, and what needs to get more specific.

    Key takeaways

    • Local SEO still starts with an accurate, consistent public record: NAP, Business Profile, and specific local pages.
    • Google asks you to keep your Business Profile current, and local pages still need to be indexable to show up as supporting links in AI features.
    • Fill in categories, services, hours, and attributes accurately. They’re not keyword fields.
    • Use LocalBusiness structured data only when it matches what’s on the page.
    • Real reviews and genuine local mentions help. Bulk citations and cookie-cutter city pages don’t.

    FAQ

    Does NAP still matter if most people find me on Google Maps?

    Yes. Your profile, your site, and your listings should still agree with each other. Inconsistencies confuse customers, and they give every system a weaker record to work from.

    Can I optimize my local pages specifically for AI Overviews?

    Not with a special trick. Make your pages specific, accurate, and indexable. Keep the Business Profile current. Write answers a person would actually use. Google doesn’t describe any separate local optimization for AI Overviews or AI Mode.

    Should a service-area business still have a location page?

    Yes, if you’ve got something real to say about the service and the area you cover. Don’t invent a storefront address you don’t have, and don’t publish a separate page for every nearby town unless the content really is different. One clear service-area page plus an accurate profile usually beats twenty thin city pages.

    The next step

    Grab a notepad. Write down the exact public name, address, and phone number you want used everywhere. Now compare that to your homepage, your contact page, your Business Profile, and your top listings.

    Fix whatever doesn’t match. Fill in the profile. Then rewrite your local page so a customer (or an AI system) can actually use it.

    If you’d rather have help with that, get in touch. I’m happy to talk through what local SEO would look like for your business.

  • What a real SEO site health audit actually covers

    What a real SEO site health audit actually covers

    Here’s a pattern I see all the time. A business publishes a blog post every week for a year. They pick up a few links. They tweak their homepage copy twice. And traffic just… sits there.

    Usually the problem isn’t the content.

    It’s something underneath it. Pages Google never indexed. A redirect chain that bleeds every internal link. A template that crawls along on phones. Three versions of the same URL quietly fighting each other. You can publish on top of that for a long time and never see anything compound.

    That’s why every serious project at Seopya starts with a site health audit. I don’t mean a one-click “score out of 100.” I also don’t mean a fifty-page PDF of warnings that nobody opens twice. I mean a structured look at whether search engines can find, crawl, render, and trust the site you already have.

    And yes, this matters for AI search now too. Google’s documentation on AI features says a page has to be indexed and eligible for a snippet before it can show up as a supporting link in AI Overviews or AI Mode. Other answer engines can only cite pages they can actually reach and read.

    So let me walk you through what a real audit covers, in the order I actually do it, and how to turn the findings into a short list of fixes instead of a pile of noise.

    What a site health audit is for

    A site health audit answers one question: what is stopping good pages from being found and understood?

    Everything else is secondary.

    Google breaks its own process into three stages (crawling, indexing, and serving results) in its guide to how Search works. I follow that same path in an audit, because a problem early in the chain caps everything after it. A beautifully written page that can’t be crawled won’t rank. It won’t get quoted in an AI answer either. It’s invisible, no matter how good the writing is.

    When I’m done, you should walk away with three things:

    • A clear picture of which pages search engines can reach, which ones they index, and which ones they ignore.
    • A short, prioritized list of technical problems, ranked by how much they hurt important pages (not by how many URLs a tool happened to flag).
    • A baseline you can measure the next six to eight months of work against.

    That’s it. If an audit doesn’t give you those, it was a report, not an audit.

    Step 1: Crawl the site the way a search engine would

    I always start with a full crawl. Screaming Frog, Sitebulb, whatever you like. The important part is setting it up to behave like Googlebot Smartphone.

    Why smartphone? Because Google uses mobile-first indexing, so the mobile version of your page is the one that counts. If your mobile template hides content, links, or structured data that the desktop version shows, that gap is a finding all by itself. You’d be surprised how often it shows up.

    What I look for in the crawl

    • Status codes. Broken internal links (4xx), server errors (5xx), and redirect chains or loops. Every internal link should go straight to a final 200 URL. No detours.
    • Orphan pages. Pages that sit in the sitemap or analytics but that no internal link reaches. If you don’t link to a page, you’re telling crawlers it doesn’t matter. They tend to believe you.
    • Crawl depth. How many clicks it takes to get from the homepage to your important service and product pages. Money pages buried five clicks deep are common, and they’re one of the easier things to fix.
    • Rendering. Whether key content and links actually exist in the rendered HTML. This one bites hardest on JavaScript-heavy themes and page builders.

    Robots.txt and crawl directives

    Next, I open robots.txt and read it line by line. It’s a short file. One wrong Disallow can still hide an entire section of your site.

    Keep one distinction in mind, which Google spells out in its robots.txt introduction: robots.txt controls crawling, not indexing. If you want a page out of the index, you use a noindex directive, as covered in Google’s guidance on blocking indexing. And the page has to stay crawlable so Google can actually see that directive.

    Blocking a URL in robots.txt and adding noindex to it at the same time? That’s one of the most common contradictions I run into. Google can’t read the noindex on a page it isn’t allowed to fetch.

    Step 2: Check what’s actually indexed

    A crawl shows you what exists. Google Search Console shows you what Google decided to do with it.

    The Page indexing report lists your indexed pages and sorts the rest by reason: “Crawled – currently not indexed,” “Discovered – currently not indexed,” “Duplicate without user-selected canonical,” “Excluded by ‘noindex’ tag,” and so on.

    Don’t panic when you see a big excluded number.

    The skill here is reading the report, not reacting to it. Plenty of URLs are excluded on purpose: filtered pages, tag archives, old parameters. What you care about is whether any important page landed in the wrong bucket. So I export the report, line it up against the crawl, and flag every page that should be indexed but isn’t.

    Canonicals and duplicates

    Duplicate URLs split your signals. The usual suspects:

    • HTTP and HTTPS versions
    • www and non-www
    • Trailing slash and no trailing slash
    • Uppercase paths, tracking parameters, and print or pagination versions

    Google’s documentation on consolidating duplicate URLs lays out your options: redirects, rel="canonical", and consistent internal linking.

    In the audit, I check that every important page has a self-referencing canonical, that canonicals point to indexable 200 URLs, and that the canonical Google picked (you can see it in the URL Inspection tool) matches the one you declared. When Google disagrees with you, it’s worth finding out why.

    XML sitemaps

    Your sitemap should be a clean list of the URLs you want indexed. Nothing redirected. Nothing noindexed. Nothing returning an error. Google’s sitemaps overview covers the format and the limits.

    Here’s a trick I use on every audit. I compare three lists: the sitemap, the crawl, and the indexing report. Where they disagree is usually where the problem lives.

    Step 3: On-page hygiene across templates

    This isn’t copywriting yet. That comes later.

    At this stage I’m doing a hygiene check across templates, because one broken template can repeat the same mistake on hundreds of pages. Fix the template and you fix them all.

    • Titles and meta descriptions: missing, duplicated, truncated, or identical across a whole section.
    • Headings: one clear H1 per page that matches what the page is for, with a sensible H2/H3 outline underneath.
    • Thin and near-duplicate pages: location or service pages that differ by a single word. (You know the ones.)
    • Images: missing ALT text, huge files, and filenames like IMG_4021.jpg.
    • Structured data: whether the markup is valid, matches what’s visible on the page, and describes the organization, author, and page type accurately. Google’s introduction to structured data is clear that markup should reflect what’s actually on the page.

    That last one is where old-school SEO and AI-era SEO meet.

    Clear titles, honest headings, consistent business details, and accurate markup help Google understand a page. That same clarity makes it easier for answer engines to pull out a correct answer and credit the right source. Same work, two payoffs.

    Step 4: Performance and page experience

    Speed isn’t a ranking trick. Think of it as a floor you need to stand on.

    Google reports Core Web Vitals on three metrics:

    • Largest Contentful Paint (LCP) for loading
    • Interaction to Next Paint (INP) for responsiveness
    • Cumulative Layout Shift (CLS) for visual stability

    INP replaced First Input Delay as a Core Web Vital in March 2024, as the Chrome team announced on web.dev. The published “good” thresholds are an LCP of 2.5 seconds or less, an INP of 200 milliseconds or less, and a CLS of 0.1 or less, measured at the 75th percentile of real visits.

    When I audit performance, I start with field data: the Core Web Vitals report in Search Console and Chrome UX Report data. That’s what real visitors experience. Then I use lab tools like Lighthouse to figure out why a template is slow. It’s usually some mix of giant hero images, render-blocking scripts, bloated page-builder output, and third-party tags nobody remembers adding.

    Google’s own page on Core Web Vitals and Search treats these metrics as part of page experience. I think that’s the right way to see them. They matter for users. They don’t replace relevance.

    Step 5: Turn findings into a priority list

    This is where most audits fall apart.

    A crawler will happily hand you thousands of “issues.” Your job is to figure out which ten actually matter. I score every finding on two questions:

    How many important pages does it affect? And how directly does it block crawling, indexing, or understanding?

    Then I sort them into this order:

    1. Blockers first: anything that stops important pages from being crawled or indexed. Robots.txt mistakes, stray noindex tags, broken canonicals, server errors.
    2. Signal leaks next: redirect chains, duplicate URL versions, orphaned money pages, and internal links pointing at redirects.
    3. Template fixes: one change that cleans up hundreds of titles, headings, or image attributes at once.
    4. Performance: the slowest templates with the most traffic, fixed at the source.
    5. Polish: the long tail of minor warnings, handled as normal maintenance.

    Every item gets an owner, a rough estimate, and a way to confirm it’s fixed. Usually that’s a re-crawl or a validation request in Search Console.

    That list is the deliverable. Not the raw export.

    What not to do

    A few traps I’d steer you away from:

    • Don’t chase a tool’s health score. A 98 out of 100 can hide a noindexed service page. A 70 can belong to a site that’s doing just fine.
    • Don’t start with content or outreach. New pages and new links just pour into the same leaks. Fix the plumbing first.
    • Don’t “fix” everything at once. If you change URLs, redirects, and templates in one big release, you’ll never know what helped and what hurt.
    • Don’t treat the audit as one and done. Sites change every time a plugin updates or someone adds a page. A lighter health check every quarter catches regressions before they get expensive.

    How this shows up in a real engagement

    At Seopya, the audit is the first phase of our SEO and AI search services. It happens before we write a word of new content or send a single outreach email.

    You get the prioritized fix list. We work through the blockers together. Only then do we move into content strategy, link building, and AI search visibility, because all of that depends on a site search engines can read cleanly. Building on a cracked foundation just makes a bigger crack.

    FAQ

    How long does a site health audit take?

    It depends on how big and complicated the site is. A small brochure site can be reviewed in days. A large ecommerce site with faceted navigation and international versions takes longer. Honestly, speed matters less than whether the findings get prioritized and verified.

    Can I run an audit myself with free tools?

    Yes, a basic one. Google Search Console, PageSpeed Insights, and the free tier of a desktop crawler will surface most blockers on a small site.

    The hard part is interpretation. Knowing which warnings are intentional and which ones are actually costing you visibility takes practice.

    Does technical SEO matter for AI search and AI Overviews?

    It does. Google’s AI features documentation says that to be eligible as a supporting link in AI Overviews or AI Mode, a page has to be indexed and eligible to show in Search with a snippet. There aren’t separate technical requirements. Other answer engines also need to reach and parse a page before they can cite it.

    Good technical health won’t guarantee a citation. Technical problems can absolutely rule one out, though.

    The next step

    If you’ve been publishing for months and don’t have much to show for it, don’t publish faster. Look underneath.

    Crawl the site. Compare it to what Google has indexed. Fix the blockers on your most important pages. Then build on top.

    And if you’d like an experienced set of eyes on it, take a look at our SEO and AI search services or get in touch. I’m happy to walk you through what an audit of your site would involve, and what it can and can’t tell you.

  • How to write meta descriptions that earn clicks

    How to write meta descriptions that earn clicks

    Let me guess. At some point, someone handed you a checklist that said meta descriptions must be exactly 155 characters, stuffed with your keyword, or Google would punish you.

    Forget most of that.

    A meta description is still one of the most useful bits of writing you’ll do for a page. Just not for the reasons those 2019 checklists claimed. It isn’t a secret ranking lever. It isn’t a hard 160-character field. And Google won’t always show the sentence you wrote.

    So what is it? A pitch. A short, accurate summary you offer up so that, when Google does use it, the right person has a reason to click.

    Updated November 2026.

    In this post I’ll show you how meta descriptions actually work now, how I write them so they earn the click, and how they fit alongside the on-page writing that AI Overviews, AI Mode, and other answer engines really read. For title tags (and how titles and descriptions should work as a pair), check out the companion guide, How to write title tags that Google will actually use.

    What a meta description is, and what it isn’t

    A meta description is an HTML element in your page’s head. Visitors don’t see it on the page. Search engines can use it as the snippet, that short blurb under the title link in the results.

    Google’s documentation on how snippets are created is what I point clients to. It’s a lot more precise than most blog posts on the subject:

    • Snippets are automatically created from page content, and they’re designed to preview the part of the page that best matches a specific search.
    • That means Google can (and often does) show different snippets for different queries.
    • Google may use your meta description when it thinks the tag describes the page better than an extract from the body.

    So you aren’t programming the snippet. You’re handing Google a well-written default, and making sure the page itself has clear sentences that can step in if Google rewrites it.

    Both jobs matter. A description that promises something the page doesn’t deliver will get ignored. Or worse, it’ll earn a click that bounces right back.

    That same document is clear on length, too. There’s no official character limit. Google truncates the snippet as needed, usually to fit the device width. The old “stay under 160 characters or Google will punish you” rule was never Google’s rule. And it’s even less useful now that most clicks happen on a phone.

    I still draft to a rough range of 140 to 160 characters, with the point of the sentence inside the first 120, because a phone only has so much width. But if a longer description is more accurate, I’ll take accurate and accept that the last clause might get cut. What I won’t do is pad a weak sentence with synonyms, or chop a good one at some arbitrary 155-character wall.

    Write unique descriptions that match search intent

    Google’s first best practice is one I still see broken on otherwise solid sites: create unique descriptions for each page.

    One site-wide blurb, or a template that just swaps the city name, doesn’t help anybody tell two pages apart. It also gives Google less reason to trust the tag.

    Unique means the description is true of this page:

    • A service page should name the service, who it’s for, and a reason to click.
    • A location page should be specific to that place, not a find-and-replace of your homepage.
    • A blog post should say what the reader will learn.
    • A product page can pull together facts scattered around the page (what it is, who it’s for, a distinguishing detail). Google actually encourages that use.

    Next, match the intent behind the searches you actually want. If people search “how to write meta descriptions,” they want a method, not a sales pitch for an agency. If they search “Miami SEO agency,” they want to know who you help and how. Write that answer in plain language. Google’s snippet guidance says outright that long strings of keywords don’t give users a clear idea of the page and are less likely to be shown.

    What about your brand name? It helps recognition. It shouldn’t eat the first line. I usually put the brand at the end, or leave it out if the title link already carries it. Lead with the benefit or the answer.

    Writing for SERP click-through rate

    A description is a tiny ad. An honest one.

    The goal is to earn a click from someone who’ll be glad they showed up. Not to trick a bigger crowd onto a page that won’t help them.

    When I write one, I run through four quick questions:

    1. Does the first clause answer the search?
    2. Is there a specific detail a competitor’s snippet probably doesn’t have?
    3. Would I still stand behind this sentence if Google showed it for a different search on the same page?
    4. Does it still make sense if a phone cuts off the last eight words?

    Specifics beat adjectives. “Expert, industry-dominating, results-driven SEO” tells a skeptic absolutely nothing. “A Miami SEO studio for brands that need to rank in Google and be cited in AI answers” tells them who it’s for.

    The title link does most of the heavy lifting. Your description should close the gap, not repeat the title word for word.

    Examples: before and after

    Quick note: these are illustrative rewrites, not results from a test.

    Service page, before (stuffed and vague): “Miami SEO, SEO Miami, best SEO company Miami, SEO services, search engine optimization, AI SEO, affordable SEO agency.”

    Ouch.

    After: “Technical SEO, on-page work, content, and local search for Miami brands, plus the foundations that help pages get cited in Google AI Overviews, ChatGPT, Gemini, Perplexity, and Grok.”

    Blog post, before (too short and generic): “Learn about meta descriptions and SEO.”

    After: “Google may rewrite your snippet, and AI Overviews need a page that can show one. Here is how to write unique meta descriptions that match intent, hold up on mobile, and still earn the click.”

    See the difference? The better versions name the real offer, drop the keyword list, and leave a clear takeaway. They’d still be useful if Google ignored the tag and grabbed a sentence from the opening paragraph instead. That’s exactly why I hold the on-page intro to the same standard.

    Mobile display, and why the opening sentence matters twice

    Most of your audience sees a shorter snippet than you do on your desktop monitor. Since truncation follows device width, the words that matter need to come early. Put the qualifier, the location, or the call to action at the end, not the start.

    Here’s the part people skip.

    The opening of the page matters just as much as the meta tag. When Google builds the snippet from the body, it often reaches for a short sentence near the top. An answer-first intro (one or two sentences that state the point before the background) gives Google, and anyone who lands on the page, the same clear start.

    Writing for AI Overviews, AI Mode, and other answer engines

    Classic SEO and AI search aren’t two separate crafts here. Google’s documentation on AI features and your website says there are no extra technical requirements and no special schema you need to show up in AI Overviews or AI Mode. To be eligible as a supporting link, a page has to be indexed and eligible to be shown in Google Search with a snippet.

    That last bit is easy to miss.

    A nosnippet robots directive, or a max-snippet of 0, doesn’t only hide your description in the blue links. Google’s robots meta tag documentation says those controls also stop your content from being used as a direct input for AI Overviews and AI Mode. So if you want to be eligible for those features, don’t switch snippets off. Google’s note on succeeding in AI search makes the same point: tighter preview permissions limit how your content can be featured.

    What about ChatGPT search, Perplexity, Gemini, and Grok? None of them publish a Google-style guide that says “we display your meta description.” I’m not going to pretend they do.

    In practice, answer engines cite pages they can fetch and understand. And the on-page habits that help Google pick a good snippet are the same ones that make a page easier to quote accurately:

    • A title, H1, and opening paragraph that agree with each other and name the entity clearly.
    • An answer-first introduction.
    • Headings that map to real questions.
    • Structured data that matches the visible text, as Google asks in its introduction to structured data.

    Open Graph title and description tags keep your public summary consistent when someone shares a link. Google also lists og:title as one of the sources for title links.

    Think of your meta description as the short public abstract for the page. If it’s unique, accurate, and lines up with the first screen of the page, you’ve given Google, an LLM, and a person pasting your URL into a chat the same honest summary.

    How to test descriptions in Search Console

    You won’t get a tidy lab result saying one sentence raised CTR by some exact percentage. Not unless you run a careful test over time.

    What you can do (and what I do after a round of rewrites) is watch the page in Google Search Console’s Performance report.

    Google’s guide to debugging Search traffic drops is the framework I lean on. If impressions hold steady but clicks drop, your title and snippet may not be helping people understand the page, or a competitor may have a more appealing result. After you update descriptions and titles, compare a period after the recrawl with a similar period before. Do it by URL and by query. The site-wide average hides too much.

    A few habits that keep me honest:

    • Write down the date you changed the copy.
    • Give Google time to recrawl before judging anything.
    • Watch position alongside CTR. If CTR shifts at the same time as a big jump in average position, it’s probably about rank, not your writing.
    • If you have enough data, check mobile and desktop separately.

    Don’t cloak, and don’t run an open-ended experiment that shows Google one title and users another. Google’s guidance on website testing is the standard I follow.

    One last thing. CTR is just clicks divided by impressions. It’s a useful diagnostic, not a trophy. A higher CTR on the wrong search is still a bad outcome if those visitors leave right away.

    A practical checklist

    When I review a page’s description, this is my list:

    • The page has its own meta description. Not shared with another URL, and not a lightly edited site tagline.
    • It accurately sums up this page, in the same language as the content, and matches the intent of the searches the page should win.
    • No keyword stuffing. The first clause still makes sense if a phone cuts off the rest.
    • The title, the H1, and the opening paragraph tell the same story. The description adds a reason to click. It doesn’t contradict them.
    • The facts a searcher wants (who it’s for, where, what they get) are in the sentence, not implied.
    • There’s no nosnippet or max-snippet:0 on a page I want eligible for AI Overviews or AI Mode.
    • Open Graph and social descriptions match the search description.
    • After publishing, I watch the URL in Search Console (queries, impressions, and CTR) once it’s been recrawled.

    FAQ

    Do meta descriptions still matter if Google rewrites them so often?

    Yes. Google says it may use your description when the tag is the more accurate summary, and a strong description is still the best default you can give it. Even when Google pulls the snippet from the page, a clear, unique summary and an answer-first intro are what it has to work with.

    Is there a character or pixel limit I must hit?

    Not officially. Google truncates the snippet to fit the device. Get the meaning in early, and treat roughly 140 to 160 characters as a planning range, not a rule.

    Are meta descriptions a ranking factor?

    Google doesn’t describe them that way. They affect whether someone understands your result and clicks. Useful, unique descriptions can improve the quality of the traffic you get. They don’t stand in for a page that deserves to rank.

    Should I write a different description for AI search than for Google?

    No. Write one accurate description, then make sure the intro, headings, and title agree with it. Eligibility for Google’s AI features depends on being indexed and snippet-eligible, not on some second hidden summary.

    The next step

    If you only do one thing after reading this, open your highest-impression pages and read their snippets like a stranger scrolling on a phone. Rewrite the ones that are vague, duplicated, or stuffed. Then line up the title, the H1, and the first paragraph, so the page would still earn the click even if Google ignored your tag completely.

    Titles deserve the same care, and the companion guide linked above covers them. If you’d like this done across your whole site, take a look at Seopya’s SEO and AI search services or get in touch.

  • How to write title tags that Google will actually use

    How to write title tags that Google will actually use

    You spend twenty minutes crafting the perfect title tag. You publish. A week later you search for your page and Google is showing something else entirely.

    Annoying? Sure. But it’s not a bug.

    The title tag is the most visible line of SEO writing you’ll ever do, and it’s also the one Google is most willing to swap out. Google’s systems build a title link (the clickable headline in results) from several sources. If your title is weak, stuffed, or out of date, Google will happily reach past it for something that describes the page better.

    Updated November 2026.

    In this post I’ll show you how title tags work now, how to write ones Google is more likely to actually use, and how titles and meta descriptions should be written as a pair. For the description half, see the companion guide, How to write meta descriptions that earn clicks.

    What a title tag is, and what a title link is

    The title tag is the HTML <title> element in your page’s head. It shows up in browser tabs, bookmarks, and a lot of shares.

    The title link is what Google actually shows in Search.

    They’re often the same text. They’re not guaranteed to be.

    Google’s documentation on influencing title links is the doc I keep open whenever I audit a site. It says title-link generation is automated, that it considers the page and references to it around the web, and that Google draws from a list of sources, including:

    • The <title> element
    • The main visual title on the page
    • Heading elements, such as the H1
    • The og:title meta tag
    • Other large, prominent text
    • Other text on the page
    • Anchor text, including links that point to the page
    • WebSite structured data

    Your official <title> is still the source Google uses most. But it’s one vote, not a command. If the tag is empty, boilerplate, inaccurate, or stale, Google may swap in your H1 or some other line.

    And after you change a title, Google has to recrawl and reprocess the page before the title link can update. That can take days or weeks. Patience helps.

    This is exactly why I write the title tag, the H1, and the opening paragraph together. When those three agree, you give Google fewer reasons to invent a headline. You also give a reader (or an answer engine) one consistent name for the page.

    Best practices that still hold

    Google’s own best practices are short. And honestly, most title problems I find come from skipping these basics.

    Give every page a descriptive, concise title. Skip vague labels like “Home” or “Blog.” There’s no official limit on how long a title element can be, but the title link gets truncated to fit the screen. So the useful part has to show up early enough to survive a phone.

    Don’t stuff keywords. A few descriptive terms help. Repeating the same phrase in slightly different forms doesn’t. Google notes that this kind of stuffing can make a result look spammy. (It usually does.)

    Don’t repeat boilerplate. Every page needs its own text. A catalog where every page is titled “Cheap products for sale,” or a long template that changes one word, makes results impossible to tell apart.

    Brand the title concisely. On interior pages, put your site name at the start or the end, separated by a hyphen, a colon, or a pipe. Keep it short. Your homepage can say a bit more about the site. Google may also drop the site name from the title link when it’s already shown next to the result.

    Make the main visual title obvious. One clear H1, visually bigger than everything else. That’s the simple fix. If two headings have the same size and weight, Google may just pick the first one.

    Write the title in the same language and script as the page. A mismatch gives Google a reason to replace your tag.

    The same document also lists the usual suspects behind rewrites: half-empty template titles, old years, inaccurate tags, micro-boilerplate, and no clear main title. If you see a title link that doesn’t match your tag, check that list first. Nine times out of ten, the answer’s there.

    How I write a title that earns the click

    I start with search intent, not a keyword list. What’s the person typing? And what would a useful headline say back to them?

    A good title does three things. It names the topic in the words people actually use. It adds a qualifier that sets this page apart, like a method, an audience, a place, or a year. And it leaves room for the brand at the end when the brand helps.

    I draft to a rough range of 50 to 60 characters. That’s because of truncation, not because Google scores length. If a precise title runs a little long, I’d rather be precise and put the unique words first. I don’t pad a short, clear title just to “fill the pixels.”

    Your title and H1 can be close without being identical. The title tag can carry the brand and slightly tighter wording for the results page. The H1 can be a bit more conversational.

    What they can’t do is disagree.

    A title that says “pricing” and an H1 that says “features.” A title that says 2024 on a page about 2026. Those are exactly the mismatches Google looks for.

    Then I read the first two sentences of the page. If the title promises a guide, the intro should start delivering on that promise right away. That’s good traditional SEO. It’s also how you write for systems that summarize a page from the top down.

    How titles and descriptions work together

    A search result is a tiny unit of copy: title link, URL, snippet.

    The title says what the page is. The description says why it’s worth opening. If they repeat each other, you’ve wasted the second line. If they contradict each other, you’ve told Google two stories and given the searcher a reason to hesitate.

    So I write them as a pair. The title carries the noun and the qualifier: “How to write title tags that Google will actually use.” The description finishes the thought: “Google builds title links from your title tag, your H1, and other sources, and it will rewrite a weak one. Here is how to write unique titles that match intent and work with your descriptions.”

    Keep both unique to the page.

    And remember, Google may replace either line. Your on-page H1 and your answer-first intro are your backup title and backup description. There’s a full method for the second line in the companion guide on meta descriptions.

    Examples: before and after

    These are illustrative rewrites, not measured test results.

    Homepage, before (vague, stuffed): “SEO, SEO Services, Best SEO Company, Search Engine Optimization, AI SEO | Seopya SEO SEO”

    After: “Expert SEO and AI Search Visibility | Seopya”

    Service page, before (boilerplate plus keyword pile): “Miami SEO | SEO Miami | SEO Agency Miami | Affordable SEO Services Miami Florida”

    After: “SEO and AI Search Services in Miami | Seopya”

    Article, before (outdated): “How To Write Industry-Dominating Meta Titles & Descriptions”

    After: “How to write title tags that Google will actually use”

    Yes, that last “before” was this very post. We all have a past.

    The better titles are specific, you can read them out loud without wincing, and they still make sense if the last word gets cut. They leave the details for the description and the first paragraph.

    Entity clarity, structured data, and social titles

    Search engines and answer engines have an easier time with a page that’s clear about who and what. Use the real organization name, the real service, and the real place in your title and opening text. Once, in natural language. Not a vague slogan repeated five times.

    Structured data won’t force a particular title link. And Google’s AI features documentation says you don’t need special schema to show up in AI Overviews or AI Mode.

    What structured data can do is confirm what’s already visible. Google’s introduction to structured data asks that markup match the text on the page. So the names and headlines in your Article or Organization markup should be the same ones a person can read.

    Social titles matter too. The Open Graph og:title and og:description tags, described in the Open Graph protocol, control how a link looks when it’s shared. Google also lists og:title as a source it may use for a title link.

    My advice? Keep your social title close to your search title. A dramatic social headline that doesn’t match the H1 creates the same kind of mismatch Google already rewrites.

    Writing titles for AI Overviews and other answer engines

    Google’s guidance on AI features lines up with the rest of Search: there’s no separate optimization track. To be eligible as a supporting link in AI Overviews or AI Mode, a page has to be indexed and eligible to show in Search with a snippet. A nosnippet rule, or max-snippet:0, can block that eligibility, as Google explains in its robots meta tag documentation.

    When it comes to the writing, I treat an AI answer the way I’d treat a busy reader.

    A model deciding whether to cite you will look at your title, your headings, and a short opening that states the point. If those lines are clear and agree with each other, you’re easier to quote accurately. If the title is stuffed, the H1 is cute, and the first paragraph takes six sentences to get going? You’re making the model (and the person) work way too hard.

    I won’t claim that ChatGPT search, Perplexity, Gemini, or Grok display your title tag in some documented, Google-like way. They don’t publish that. They cite pages they can crawl and parse.

    The practical work is the same anyway:

    • A precise title.
    • A matching H1.
    • An answer-first intro.
    • Visible text instead of text locked inside an image.
    • A description that could double as a summary.

    That’s also what Google’s note on succeeding in AI search keeps coming back to: useful, crawlable, people-first content, with preview controls that don’t needlessly restrict how your pages can be shown.

    How to test titles in Search Console

    After a round of title changes, I watch the affected URLs in the Search Console Performance report.

    Google’s guide to debugging Search traffic drops points out a handy pattern. If impressions stay roughly flat but clicks drop, your title and snippet may not be helping people understand the page.

    Compare a period after the recrawl with a similar period before, by page and by query. Read CTR as clicks divided by impressions, and check average position at the same time. If a title change lands alongside a big swing in rank, it’s not a clean test of your writing.

    Make one thoughtful change. Date it. Wait for the recrawl.

    And don’t show Google a different title than the one people see. Google’s website testing guidance is clear about cloaking, even during experiments.

    A practical checklist

    • Every indexable page has a unique, descriptive <title>. No “Home,” no site-wide slogan, no keyword list.
    • The important words come early enough to survive mobile truncation. The brand goes at the end, briefly, unless there’s a good reason to lead with it.
    • The title, the H1, and the first paragraph agree. The H1 is the clearest visual title on the page.
    • The title and the meta description do different jobs. They don’t contradict each other, and they’re unique to this URL.
    • The language and script match the page. Dates and claims in the title match what the page says now.
    • og:title lines up with the search title. Structured data, if you use it, matches the visible headline.
    • The page can be indexed and can show a snippet, if you want it eligible for AI Overviews or AI Mode.
    • After publishing, you watch the URL in Search Console (queries, impressions, clicks, and position) once Google has recrawled it.

    FAQ

    Why does Google show a different title than the one I wrote?

    Usually because the tag is vague, duplicated, outdated, stuffed, or out of step with your H1. Google lists these cases in its title-link guidance and may substitute other text. Fix the mismatch, request a recrawl, and wait.

    Should the title tag and the H1 be identical?

    They should mean the same thing. They don’t have to match word for word. Use the title tag for a tight, brand-aware headline in results, and the H1 for the on-page heading. Just don’t let them conflict on topic, date, or promise.

    How long should a title be?

    Google doesn’t set a character limit. The title link gets cut to fit the screen. I draft around 50 to 60 characters and put the distinctive words first. A slightly longer, accurate title beats a short, generic one.

    Do I need a special title format for AI search?

    No. Google says there are no extra requirements for AI Overviews or AI Mode beyond ordinary Search eligibility. Write a clear, unique title, line it up with the H1 and the intro, and keep the page snippet-eligible.

    The next step

    Open your highest-impression pages and compare three things: the title tag, the H1, and the title link Google is actually showing. Where they disagree, fix the writing, not the symptom. Then write the meta description so it finishes the thought the title started.

    If you’d like a second set of eyes on it, take a look at Seopya’s SEO and AI search services or get in touch.