The on-page SEO articles in this series have set out, in some detail, how to do on-page SEO well. This article approaches the same subject from the other side, by setting out how it most often goes wrong.
The two approaches are not redundant. A business can read an account of good practice, agree with all of it, and still carry several of these mistakes on its own site without recognising them — because the mistakes are quiet, and a page that suffers from one rarely announces the fact. This article names eight of them, explains why each happens and what it costs, and is meant to be read as a diagnostic the business can run on its own pages.
A note on sources is in order. Peer-reviewed research is cited by author and year and listed at the end; Google’s own published guidance is cited as a primary source and identified as such; and any claim resting on the common practice of the SEO field rather than on a specific source is identified as practitioner consensus.
What this article covers
This article covers eight on-page SEO mistakes — eight ways that the optimisation of a page’s own content commonly fails — and treats each in enough depth to be recognised and corrected. It is a companion to the on-page SEO guide earlier in this series, which a reader who wants the positive account should treat as the foundation.
The article first explains how the eight mistakes relate to one another, since they are not eight unrelated faults but fall into three groups. It then treats each in turn, and closes by identifying the single attitude that lies beneath several of them — because a business that corrects that one attitude has gone most of the way to correcting the rest.
How to read this list
The eight mistakes are easier to hold, and to act on, when their relationship is clear. They are not a random list; they fall into three groups, according to which part of a page’s job they damage.
The first group concerns whether the page is a clear answer to any question at all. The on-page SEO guide established the governing principle — that a page is, for the purposes of search, an answer to a question — and the mistakes in this group are the ways a page fails to be a clear answer to a specific one.
The second group concerns the quality of the answer. These are pages that do address a clear question, but address it badly — pages that are the answer to something, but a poor answer. The third group concerns delivery: pages that contain a genuinely good answer but present it in a way that the customer, or the search engine, cannot easily use. The figure below sets out the three groups and the eight mistakes within them.
Why these mistakes are so common
Before the eight themselves, it is worth asking why on-page mistakes are so widespread — because the answer explains why a diagnostic like this one is needed at all.
The first reason is that the mistakes are invisible from the inside. A page that suffers from one of them looks, to the business that made it, finished: it has a title, it has content, it loads, it can be visited. Nothing about it signals a fault, because the fault is not a breakage but an absence — an absence of clarity, of substance, of purpose — and an absence does not announce itself.
The second reason is that nothing corrects the business. A page with a genuine technical error may produce a visible message; a page that simply fails to rank produces only silence. The business sees no traffic and has no way, from the page itself, to know which of several possible faults is the cause. The mistake persists because nothing forces it to the surface.
The third reason is that most SEO advice, including the earlier articles in this series, describes how to do on-page SEO well — and a business can absorb all of it and still not recognise the same faults in pages it built before, or built quickly, or built without thinking. Knowing the right way to write a page is not the same as auditing the pages already written. That gap is what this article exists to close: it describes the mistakes from the symptom inward, so that a business can find them in work it has already done.
There is one more reason, and it is less about the pages than about the business’s attention. Correcting these mistakes is unglamorous work: it produces no new page to point to, no visible addition to the site, only quieter improvements to pages that already exist. Unglamorous work is easily deferred, and a business naturally prefers building the next thing to auditing the last one — so the mistakes survive not only because they are hard to see but because, even once seen, they compete for attention with more satisfying work and tend to lose. A diagnostic helps here too, by making the invisible faults concrete enough that fixing them feels like a real task rather than a vague intention, and a task, unlike an intention, can be scheduled and finished.
Mistake one: the page that answers no clear question
The first mistake is the most fundamental, and the easiest to miss because the page in question looks finished. It is the page built without anyone deciding what question it answers — a page that exists, that has content on it, but that is not a clear answer to any specific search.
The mistake happens because pages get created for the business’s internal reasons rather than the customer’s. Someone decides the site “should have a page about” a topic, and a page is duly made, without the prior question being asked: what does a customer search when they want this, and is the page the answer to that search. A page made without that question is aimed at nothing.
The cost is that the page ranks for nothing in particular. A search engine matches pages to queries, and a page that is not a clear, focused answer to a specific query does not match any query well enough to win it — so the work of making the page returns no visitors. The page is not harmful so much as inert.
The mistake is spotted with one question, asked of every page: what question is this page the answer to? A business that cannot state the answer in a single plain sentence has found a page with this problem. The fix is to give the page a clear question and rebuild it as the answer to that question, or, if it has no question worth answering, to remove it.
One distinction helps when deciding which way to resolve such a page. Some purposeless pages cover a topic that customers do genuinely search for, and these are worth rebuilding as proper answers; others cover a topic no customer searches at all, and these are worth removing rather than rescuing. The test is whether the page’s potential question is one anyone actually asks — a page can be given a purpose only if a real purpose is available to give it.
Mistake two: ignoring search intent
The second mistake is subtler, because the page that suffers from it has a clear subject and may even rank. It is the page built around the right keyword but for the wrong kind of question — a page that meets an informational searcher with a sales pitch, or a ready-to-act searcher with a long explanatory essay.
The mistake happens because keyword research, done on its own, is misleading. Research into keywords tells a business what words customers use; it does not tell the business what those customers want. A business that sees a keyword and builds a page on it, without asking what the searcher behind the keyword is trying to do, builds a page that is about the right thing and for the wrong purpose.
The cost is a mismatch between the page and its visitors. The taxonomy of search distinguishes the searcher who wants to learn from the one who wants to act (Broder, 2002), and a page that serves the wrong one of these loses the visitor it attracted — and a page that ranks for a query it does not serve well is, in a sense, worse than one that does not rank, because it draws visitors only to disappoint them.
The mistake is spotted by asking, for each page, what the searcher it is meant to receive actually wants — to understand something, or to do something. The fix is to build the page to that intent: a page serving informational intent should genuinely inform, and a page serving the intent to act should make acting easy and obvious.
There is a quick way to read the intent behind a search, and it is to look at what a search engine already shows for it. A search engine has, in effect, already inferred the intent of a query and assembled results to match — so a business uncertain what kind of page a search wants can perform the search and see. If the results are explanatory articles, the intent is informational; if they are pages built to transact, it is transactional. The existing results are a reading of the intent the business can simply consult.
Mistake three: duplicate and near-duplicate content
The third mistake is the site that contains several pages saying nearly the same thing — multiple pages competing to be the answer to one question, and sometimes sharing the same title as well.
The mistake usually accumulates rather than being committed at once. A business adds pages over the years — variations on a service, pages for different locations, posts on overlapping topics — and does not notice that several of them have drifted into answering the same question. Occasionally it is deliberate, the result of a belief that more pages on a topic mean more chances to rank.
The cost is that the duplicate pages compete with one another. A search engine deciding which page to show for a query cannot tell which of several near-identical pages is the answer, and so none of them is the strong, distinct answer a single good page would have been; the site’s effort on the topic is divided among pages that undercut each other. Google’s guidance addresses duplicate content directly as a problem to resolve (Google Search Essentials, 2022).
The mistake is spotted by looking across the site for pages whose subjects overlap, and asking whether each genuinely answers a distinct question. The fix is to consolidate overlapping pages into one strong page, or, where the pages should be distinct, to make them genuinely distinct — so that each is the clear answer to a question the others do not answer.
One version of this mistake is common enough to name on its own: the set of near-identical pages a business creates for different locations, each one the same page with a place name changed. The intention is reasonable — to be found in each town the business serves — but the execution produces exactly the duplication this mistake describes. A location page works only if it is genuinely about that location, with content true of that place and not merely a name substituted into a template.
Mistake four: keyword stuffing
The fourth mistake is the oldest, and although it is widely known to be a mistake it still appears: keyword stuffing, the practice of repeating a target term unnaturally through a page on the theory that more repetitions produce a higher rank.
The mistake rests on a misunderstanding of how search works — the belief that a search engine ranks a page by counting how often the target term appears on it. There was a distant period when something like this was partly true, and the belief has outlived the era that produced it by a long way.
The cost is threefold, and worth being precise about. Keyword stuffing does not raise rank, so the effort is wasted; it makes the page read badly to the human customer, so it costs conversion; and Google’s guidance names it explicitly as a spam practice, so a page that does it can be actively penalised rather than merely unrewarded (Google Search Essentials, 2022). It is a mistake that fails at its own goal and damages the page besides.
The mistake is spotted by reading the page aloud: a stuffed page sounds wrong, repeating its term in places natural language would not. The fix is the principle from the on-page guide — write the page for the customer, in natural language, and let the target term appear as often as ordinary writing requires and no more.
Modern keyword stuffing is often subtler than the crude repetition the term suggests, and so easier to commit without noticing. It can take the form of cramming slight variations of a term together, or of writing every heading on a page around the same phrase, or of forcing the keyword into sentences where a natural writer would use a pronoun. The test catches all of these forms: if a sentence was shaped by the wish to place a term rather than the wish to say something, it is stuffing, however subtle.
Mistake five: thin content
The fifth mistake is the page that promises an answer and does not deliver one — the page with a title and a heading and almost no substance beneath them. This is thin content: a page that exists, on its subject, but that does not actually answer the question it raises.
The mistake happens when a page is made in order to have a page, without the harder work of genuinely answering. A business decides it needs a page on a topic, creates one, and fills it with a few sentences that gesture at the topic without addressing it — or spreads the substance it does have so thinly across many pages that no single page carries enough.
The cost is that a thin page satisfies no one. The customer arrives expecting the answer the title promised, finds little, and leaves; the search engine sees a page that does not answer its own question, and Google’s guidance treats thin, low-value pages as a problem rather than a contribution. A thin page is not a small positive contribution to the site — it is a non-contribution that also dilutes the site around it.
The mistake is spotted by asking whether a page genuinely answers, in full, the question it raises. The fix is binary: either do the work to make the page a complete and genuine answer, or remove the page. There is little value in a page that exists without answering.
Thin content is sometimes produced not by laziness but by over-division — a business takes a topic that would make one genuinely substantial page and splits it across five, leaving each of the five too slight to satisfy. The fix in that case is not to add filler to each thin page but to merge them: one complete page on the topic does what five thin ones cannot, and it is the same total effort gathered where it works.
Mistake six: generic, interchangeable content
The sixth mistake is harder to see than thin content, because the page that suffers from it can look substantial. It is the page whose content is generic — written in language so general that it could appear, word for word, on any competitor’s page.
The mistake happens because generic writing is easier and feels safer. Abstract, flattering, non-specific phrasing — about quality, commitment, a tailored approach — can be produced without the effort of saying anything particular, and a business unsure what to write defaults to it, often echoing the same empty phrasing its competitors use.
The cost is that generic content does not do a page’s job, even though it fills the page. It tells the customer nothing that helps them choose this business over another, and it gives the search engine no reason to regard this page as a distinct answer rather than one of many identical ones — in the framing this series draws from Akerlof, a string of general claims does not convey the quality the customer cannot otherwise see (Akerlof, 1970). Generic content is, in effect, thin content wearing the costume of substance.
The mistake is spotted with a test the service-page article introduced: take any sentence on the page and ask whether a competitor could publish it unchanged. If they could, the sentence is generic. The fix is to replace the generic sentences with concrete, particular ones — the specific things that are true of this business and no other.
Generic content feels like the safe choice, which is the reason it is so common and the reason it is worth resisting deliberately. Vague, flattering, abstract phrasing commits the business to nothing and so cannot be wrong — but a page that cannot be wrong also cannot be persuasive, because persuasion requires saying something specific enough that it could, in principle, have been otherwise. The safety of generic content is the safety of having said nothing.
Mistake seven: poor on-page structure
The seventh mistake concerns a page that may contain a genuinely good answer and yet fails, because the answer is delivered badly. This is poor on-page structure — the wall of undivided text, the answer buried far down the page, the absence of headings, the page that cannot be scanned.
The mistake happens when a business writes a page’s content and gives no thought to how that content is organised. The substance is poured onto the page as one undifferentiated block, in the order it occurred to the writer, with nothing to mark its parts or to bring its main answer near the top.
The cost falls on both of a page’s readers. A customer scanning the page for the part of the answer they need cannot find it, and leaves; a search engine, which uses a page’s structure to understand how its content is organised, has a harder time making sense of an unstructured page. The answer is present, and effectively inaccessible — which, from the reader’s side, is much the same as the answer being absent.
The mistake is spotted by looking at a page as a scanning customer would: can its main answer be found quickly, and are its parts marked. The fix is the structural discipline the on-page guide described — headings that name the parts of the answer, the main answer placed near the top, internal links to related pages, a page built to be navigated rather than waded through.
The tells of poor structure are easy to recognise once looked for. A page with no headings, or with paragraphs that run for many lines without a break, or whose main answer sits below the point where a phone screen ends, is a page with this mistake. The remedy is not to cut the content but to organise it — the same answer, divided into marked parts and arranged so the most important of them is reached first.
Mistake eight: neglected titles and meta descriptions
The eighth mistake is the one most often made by simple omission: leaving a page’s title tag and meta description as whatever the website’s software generated by default — blank, generic, or duplicated across many pages — and never writing them deliberately.
The mistake happens because these elements are invisible on the page itself. A business looking at its own pages sees the headings and the content, not the title tag and the description, and so forgets they exist — or does not know that they matter, or how to set them.
The cost was the subject of an entire article in this series. The title tag is both a ranking signal and the headline a searcher reads in the results, so a default or duplicated title costs the page ranking and clicks together; the meta description, while not a ranking factor, is the snippet that helps win the click. A page carrying neglected versions of these elements can rank and still go unvisited, because nothing about its result invites the click.
The mistake is spotted by checking, page by page, whether each important page has a deliberate, distinct title and description. The fix is to write them — a task that, as the article on titles and descriptions argued, is quick, fully within the business’s control, and neglected by enough competitors that doing it is a genuine advantage.
Finding this mistake is unusually easy, because the elements, though invisible on the page, are visible elsewhere. The title tag shows in the browser’s tab when the page is open, and both the title and the description show in the page’s search result — and a search engine’s free tools for site owners will list the titles across a site, making duplicates and omissions plain. The mistake is one of the simplest of the eight to both detect and correct.
How the mistakes interact
The eight have been treated one at a time, for clarity, but they rarely appear one at a time. A single weak page commonly carries several of them at once, and the mistakes tend to keep company in predictable ways.
A page built with no clear question, for instance, is very often also thin and generic — having no specific question to answer, it has nothing specific to say, and so it says little, and what it says is vague. A page that ignores intent often has a neglected title as well, because the same inattention that skipped the question of intent skipped the writing of the title. The mistakes share causes, so they share pages.
This has a practical consequence for the audit. Fixing one mistake on a page that has three leaves the page still failing, and a business that corrects a page’s structure while leaving its content thin has improved the delivery of an answer that is still not worth delivering. The pages worth fixing should be fixed for all of their mistakes at once, or the effort is partial.
It also has a consequence for judgement. A page that carries most of the eight mistakes is often not worth repairing at all — it is quicker and better to rebuild it, or to remove it, than to correct fault after fault on a page that was never a genuine answer to begin with. The audit is partly a triage: it sorts the pages worth fixing from the pages worth replacing.
The common root: writing for the algorithm instead of the customer
The eight mistakes are distinct, and a business could correct them one at a time. But several of them grow from a single root, and naming that root is more useful than any individual fix.
Keyword stuffing, generic content, the page built with no real question, even some thin content — these are what a business produces when it writes for an imagined algorithm rather than for the actual customer. The business pictures a machine that counts keywords, that rewards pages for existing, that can be satisfied by the appearance of content; and it writes pages aimed at that picture.
The picture is wrong, and writing for it produces exactly the pages that fail. A real search engine, as the foundational accounts make plain and as Google’s own guidance repeats, is built to find the page that genuinely serves the person searching (Brin & Page, 1998; Google Search Essentials, 2022). A page written for the imagined keyword-counting machine is therefore aimed away from what the real machine rewards.
This is why the deepest correction for the eight mistakes is a single principle rather than a checklist. The on-page SEO guide stated it: write the page for the customer who will read it, because the algorithm is built to reward that page. A business that genuinely holds that principle does not stuff keywords, does not write generic filler, does not publish purposeless pages — not because it remembers a rule against each, but because none of those things is what writing for a real customer produces.
It is worth being clear that this does not make the eight mistakes interchangeable, or the individual fixes unnecessary. A business still has to find each mistake on each page, and the table that follows is the tool for that. The point of naming the common root is different: a business that corrects its underlying stance — that genuinely begins writing for the customer rather than the imagined machine — stops producing new instances of these mistakes, even as it works through the old ones. The audit clears the existing faults; the principle prevents the next ones.
An on-page self-audit
The table below turns the eight mistakes into a self-audit. For each, it gives the question a business can ask of its own pages to find the mistake, and the fix once it is found.
A word on how to use it. A business does not need to audit every page at once, and should not try to. The sensible order is to start with the pages that matter most — the service pages, the pages that are meant to bring in customers — and to work through the eight questions for each of those before moving outward to the rest of the site. A mistake found on a service page is worth more to fix than the same mistake found on a page few customers were ever going to reach.
| Mistake | How to spot it on your own site | The fix |
|---|---|---|
| No clear question | You cannot say in one sentence what question the page answers | Give the page a clear question, or remove it |
| Ignoring search intent | The page’s kind does not match what its searcher wants to do | Rebuild the page to serve its intent |
| Duplicate content | Several pages answer the same question or share a title | Consolidate, or make the pages genuinely distinct |
| Keyword stuffing | The page repeats a term where natural writing would not | Rewrite in natural language for the customer |
| Thin content | The page does not actually answer the question it raises | Complete the answer, or remove the page |
| Generic content | A competitor could publish your sentences unchanged | Replace generic claims with concrete particulars |
| Poor page structure | The answer cannot be found quickly; parts are unmarked | Add headings, lift the answer up, structure the page |
| Neglected titles and descriptions | Titles and descriptions are default, blank, or duplicated | Write a deliberate, distinct one for each page |
Concluding remarks
The eight on-page SEO mistakes treated here are quiet faults: a page that carries one rarely shows it, which is why a business can do on-page SEO with good intentions and still lose traffic to errors it has not noticed. The eight fall into three groups — the page that is not a clear answer to any question, the page that is a poor answer, and the page that delivers a good answer badly.
Within those groups: a page may have no clear question, may ignore the intent of its searcher, or may duplicate another page; it may stuff keywords, run thin, or fill itself with generic filler; or it may bury a good answer in poor structure, or carry a neglected title and description. Each has a way it can be spotted on a business’s own site, and each has a fix, and the table above gathers both.
Beneath several of the eight is one root: writing for an imagined algorithm rather than the real customer. The single most effective correction is therefore the principle the on-page SEO guide built on — write the page for the customer, because the search engine is built to reward the page that genuinely serves them.
This article closes the on-page SEO part of the series. The next articles turn to technical SEO — the part concerned not with what a page says but with whether a search engine can reach, read, and is willing to serve the site at all.
Future developments
On-page SEO mistakes do not disappear as search changes; they change form. AI-driven search, which composes answers from the pages it reads, alters which of the eight mistakes hurt most and how.
Some of the eight become more costly. A thin or generic page gives an AI system nothing it can usefully draw on, and a page with no clear question is not the clear, citable answer such a system is built to find — so the mistakes of substance and clarity are, if anything, punished more sharply when the page is being read by a system that wants to extract a genuine answer from it.
The reassuring point is that the corrections do not change. A page that is a clear, complete, concrete, well-structured answer to a real customer’s real question is what serves a traditional search engine, and it is also what an AI system can read, trust, and cite. The eight mistakes are worth correcting now, and the work of correcting them is not made obsolete by the shift in how customers search — it is, if anything, made more valuable.
Related reading
- On-page SEO for small business websites: a complete guide
- How to write a service page that ranks and converts
- Title tags and meta descriptions: what actually moves rankings
- A technical SEO audit a non-developer can actually do
References
Akerlof, G. A. (1970). The market for “lemons”: Quality uncertainty and the market mechanism. The Quarterly Journal of Economics, 84(3), 488–500.
Brin, S., & Page, L. (1998). The anatomy of a large-scale hypertextual web search engine. Computer Networks and ISDN Systems, 30(1–7), 107–117.
Broder, A. (2002). A taxonomy of web search. ACM SIGIR Forum, 36(2), 3–10.
Google Search Essentials. (2022). Google Search Central documentation. Google. [Primary source — official platform documentation, not peer-reviewed.]
