Common SEO Mistakes: 11 Errors and How to Find Yours

An SEO mistake is any decision that stops a page being found, ranked or clicked, and the eleven below are not worth the same. Some cost you the whole page, some a handful of positions, and two cost almost nothing measurable. A list that treats them as equal leaves you no clearer about your own site. This is for founders and marketers at SaaS and B2B companies who are publishing steadily and still not ranking, and who cannot yet tell whether the problem is technical, editorial or strategic. Every entry names what it costs, a free way to confirm you have it, and roughly how long the fix takes.

The short version

• The expensive mistakes are the ones that stop a page entering the index at all. Until that is fixed, a sharper title changes nothing, because nothing is competing.
• Publishing more is not a remedy for not ranking. We ran that at scale on this site, and the figures are in the final section.
• Two of the eleven cost less than their reputation suggests: keyword density and raw backlink counts absorb most of the worry and cause least of the damage.
• All eleven checks here are free. Search Console, one crawl and the live results page will diagnose every one.
• Work in cost order, not in the order you happen to notice things.

Which SEO mistakes actually cost you anything?

Four tiers, and only the first is urgent. A mistake's cost depends entirely on which stage of the pipeline it breaks: being crawled and indexed, then ranked, then clicked. Break the first stage and everything downstream is irrelevant, because the page is not in the race at all.

Tier	What it costs you	When to fix it
1. Lost indexing	The entire page. No position, no impressions, no clicks, and the work that went into it returns nothing.	First, always
2. Lost rankings	Positions. The page is indexed and collecting impressions, but sits too deep to be clicked.	Second
3. Lost clicks	Traffic only. The position is already good, so the gap is between the result and the reader.	Third, and it is the cheapest win
4. No measurable cost	Your attention. Fixing it changes a number that was never holding you back.	Last, or never

The practical consequence: an unindexed page with a brilliant title is worth exactly as much as an unindexed page with a terrible one. That is why the order below is cost, not difficulty.

Which mistakes stop your pages being indexed?

These four are tier one. Each removes a page from search results completely, and each is cheap to confirm once you know which report to open.

1. Publishing volume without checking whether the domain can rank for what you publish

This is the mistake I hold the strongest opinion about, because we made it here, deliberately and at scale. Volume multiplies whatever one page is worth, so when that value is near zero the total stays near zero. Our guide to setting a weekly publishing number from your own data works through the full record; the short version is in the last section of this article.

• What it costs: Every hour spent on pages that were never winnable. On a young domain those pages frequently go unindexed as well, so the loss compounds.
• How to confirm it: In Search Console, open the Pages report for the last 90 days and count how many of your posts took zero clicks. A large unclicked majority means topic selection, not output, is the constraint.
• The fix, and how long it takes: An afternoon to re-qualify your topic list, then ongoing discipline about what clears the bar. The rule we adopted is in the final section.

2. Pages blocked by noindex or robots.txt that were meant to be live

Staging rules that shipped to production are the usual culprit, and the damage is silent because the page looks perfect to a human visitor. The two mechanisms also interact in a way that catches experienced teams.

• What it costs: The page, entirely. A noindex rule drops it from results regardless of how many sites link to it.
• How to confirm it: Run the URL through URL Inspection and read the indexing verdict, then open yourdomain.com/robots.txt and look for a Disallow covering that path. The two mechanisms can also cancel each other out. Google's noindex documentation is explicit that the rule only takes effect on a page Googlebot can still reach, so disallowing the same URL in robots.txt means the tag is never read, leaving the page eligible to appear with no clean way to pull it back.
• The fix, and how long it takes: Minutes to change the tag or rule, then days to weeks for Google to recrawl. Our write-up on pages that are crawled but not indexed covers the cases where the tag is not the cause.

3. Canonical tags pointing somewhere unintended

A canonical tag pointing at the homepage, at a parameter-stripped duplicate, or at a page that no longer exists will quietly consolidate a good page out of the index. Templated canonicals are the usual source.

• What it costs: Indexing for every page the rule touches, which on a template means a whole section at once rather than a single URL.
• How to confirm it: URL Inspection reports the user-declared canonical and the Google-selected canonical side by side. When they disagree, Google has made a different choice from yours. Per Google's guidance on consolidating duplicate URLs, a rel="canonical" annotation is "a strong signal", not an instruction that must be obeyed.
• The fix, and how long it takes: An hour at template level if the rule is centralized, longer if canonicals were hand-written per page.

4. Orphan pages with no internal link pointing at them

A page reachable only from the sitemap is one Google has little reason to prioritise and readers cannot find at all. It is the one tier-one mistake no single report hands you, because it is defined by an absence.

• What it costs: Discovery, and the crawl priority that comes with being linked. Orphans tend to sit unindexed longest.
• How to confirm it: Crawl the site from the homepage, then compare the URLs the crawler reached against your sitemap. Anything in the sitemap and not the crawl is an orphan.
• The fix, and how long it takes: A few hours. Point links at it from topically adjacent pages Google has already indexed, because an edge from a page it has never crawled passes nothing.

Which mistakes cost you rankings rather than indexing?

Tier two. The page is in the index and collecting impressions, so the diagnosis shifts from "does Google have it" to "why is it ranked where it is".

5. Two pages targeting the same query

When two of your URLs compete for one query, Google picks between them and often picks inconsistently, so neither accumulates the signals it needs. The symptom is a query where your ranking URL changes from week to week.

• What it costs: Positions on the query, split across pages that would rank better combined.
• How to confirm it: In the Search Console performance report, filter to a single query and open the Pages tab. More than one URL appearing there is the signal. Our guide to diagnosing and resolving cannibalization covers which page to keep.
• The fix, and how long it takes: A day. Either consolidate into one page and redirect the loser, or genuinely re-target one of them at a different query.

6. Content that answers a different intent from the one the query carries

A query like "best project management software" wants a comparison, and a product page will not rank for it no matter how well optimised that page is. Intent mismatches are invisible in every technical report, because nothing is broken.

• What it costs: Positions, usually severely. You are competing in a format the results page has already rejected.
• How to confirm it: Search your target query and read what currently ranks. If nine of the ten results are a different content type from yours, the mismatch is yours.
• The fix, and how long it takes: Hours to days per page, because it usually means rewriting rather than editing.

7. Thin pages kept because deleting feels like a loss

Pages that exist to hold a keyword, without the substance to answer it, dilute the impression a site gives of its own expertise. The instinct to keep everything is understandable and mostly wrong, though the correction is narrower than it sounds.

• What it costs: Rankings across the section they sit in, not just their own.
• How to confirm it: List every page with zero clicks over 90 days, then read them. The test is whether the page answers its query better than what currently ranks, not word count.
• The fix, and how long it takes: Variable, and selective. Improve the ones close to page one, merge the ones that overlap, and leave the rest until there is a reason to act.

Which mistakes cost you clicks you have already earned?

Tier three, and the cheapest work on this list. The page already ranks, so these fixes convert existing impressions rather than creating new ones.

8. Titles and meta descriptions written for the keyword, not the click

A title assembled to contain a phrase usually reads like one. The reader is scanning ten results and choosing between them, which is a different job from matching a string.

• What it costs: Clicks only, but at zero cost to recover, which makes the ratio the best on this page.
• How to confirm it: Sort the Search Console pages report by impressions and look for pages in positions roughly five to fifteen whose click-through rate sits well below your average at that position.
• The fix, and how long it takes: Ten minutes a page. Lead with the specific answer or number the page delivers, and keep titles under about 60 characters so they survive truncation.

9. Core Web Vitals failures that live in the templates, not the pages

Performance problems almost always belong to a layout, not an article, so fixing them once in the template beats fixing them page by page. Treat this as a tier-three cost: it refines results you already have rather than rescuing a page.

• What it costs: Clicks and completed visits. Useful to fix, rarely the reason a page is absent.
• How to confirm it: In Search Console, the Core Web Vitals report clusters URLs by similar behaviour, and those clusters usually map to templates. Google's Web Vitals documentation publishes the pass marks, and each one is assessed at the 75th percentile across mobile and desktop rather than on a single lab score: LCP at or under 2.5 seconds, INP at or under 200 milliseconds, CLS at or under 0.1.
• The fix, and how long it takes: A developer day per template, usually spent on images, fonts and reserving space for anything that loads late.

Which common mistakes cost less than their reputation?

Tier four. Both of these appear on nearly every list of SEO errors, and both absorb attention out of proportion to what they return.

10. Counting keyword density and exact-match anchor ratios

Keyword stuffing is a genuine policy violation. A density percentage is not how Google describes it. The spam policies define the practice as filling a page with keywords "in an attempt to manipulate rankings", and the test offered is repetition "so often that it sounds unnatural". No threshold appears in that definition at all, so tuning a page to a percentage optimises for a number Google never published.

• What it costs: Usually nothing, unless the writing has been distorted to hit a target, at which point the cost is readability.
• How to confirm it: Read the page aloud. If the phrasing sounds forced, that is the problem, and it is an editing question rather than a measurement one.
• The fix, and how long it takes: Stop counting. The time you get back is the actual return.

11. Chasing backlink volume before the pages are worth linking to

Links amplify a page that already deserves to rank; they do not substitute for one that does not. Acquiring links to thin or mistargeted pages is the most expensive way to learn that.

• What it costs: Budget and months, with little to show, though genuine authority does matter in competitive categories.
• How to confirm it: Ask whether anything you have published in the last quarter would be worth citing by someone with no relationship to you. If not, link acquisition is premature.
• The fix, and how long it takes: Reorder the work. Make a small number of pages genuinely worth referencing first.

How do you work out which one is yours?

In three passes, in this order, and the order is what saves the time. When I pick up a site that has flattened out, Search Console comes before the crawler every time, because the first question is not what is wrong with a page but whether Google is holding it at all. Answering that reorders everything else.

1. Pass one, coverage: Open URL Inspection for the handful of URLs that matter most, then read the Pages report to see what Google has excluded and why. This finds mistakes 2, 3 and 4, and it is the only pass that can tell you the rest do not matter yet.
2. Pass two, the crawl: Crawl the site and look for orphans, canonical conflicts and template-level performance failures. These are structural, so one finding usually implicates many pages at once.
3. Pass three, the content review: With coverage and structure clean, compare queries to pages. This is where cannibalization, intent mismatch and thin pages surface, and it is slow because it needs judgement rather than a report.

Both of the first two passes are largely automatable. GrowthHasten puts a site through 130+ deterministic checks, sorts what it finds into 12 scored categories, and orders the pages so the costliest problem sits at the top; you can run it against your own domain on a free plan that reads one website of 900 pages or fewer and allows three re-analyses, without a card or a trial clock. The design decision that matters for a diagnostic job: every check is a deterministic function of markup already parsed, so AI is handed verified findings to explain and order rather than asked to produce them, and the report names the dimensions it could not measure instead of scoring them zero. That is the argument for why a score should admit what it did not measure: a confident number over a check that never ran is worse than no number. The manual equivalent is laid out step by step in our technical SEO checklist.

What did the first mistake cost us on this site?

Fifteen clicks. This site published 149 posts in nine weeks and now carries 156 published articles. Over the 28 days to 1 October 2026, the whole domain recorded 15 clicks from 14,142 impressions at an average position of 67.21, measured in our own Search Console. An average position in the high sixties means the typical impression sat several pages deep, where clicks effectively do not happen.

Every one of those posts met the structural standards: headings, tables, internal links, FAQs, sources. The checklist was not the problem. The topics were. They were questions established sites already own, and a domain whose Ahrefs Domain Rating measured 22 was never going to take them by writing a cleaner definition.

One thing that record is not: evidence of a penalty. Over the same stretch our average position improved rather than slipped, and clicks held flat while impressions fell. What disappeared were impressions from positions deep on page nine that had never produced a click. A penalty moves position the other way. The cost here was wasted effort, which is the far more common version of this mistake and the one almost nobody writes about.

What changed as a result: a ceiling of three published posts a week, down from an average of roughly sixteen, and a rule that every topic must carry a measured difficulty figure and at least two nameable, beatable results on the live first page before a word is written. The 149 posts are all still live. The decision governs what gets written next, not what already exists.

When is one of these not worth fixing?

Several of them, more often than the lists admit. A diagnosis is only useful if it also tells you when to stop.

Leave it alone when: the page is correctly excluded, as thin, duplicate and internal pages often should be. When Core Web Vitals already pass at the 75th percentile, further tuning is engineering for its own sake. When a thin page serves a real purpose, such as a pricing or policy page, its search performance is not the measure that matters.

Be especially careful with deletion: removing large amounts of older content in the hope of a ranking improvement is its own mistake, and it is irreversible in a way that editing is not. Fix or merge before you delete.

The honest limit on everything above: these are the mistakes that are findable with free tools. A site can do all eleven correctly and still not rank, because it is competing for queries its authority cannot reach yet. That is a strategy problem, and no audit will diagnose it for you.

Start with five URLs, not eleven mistakes. Put the five pages that matter most through URL Inspection and read the verdict on each. Anything coming back as something other than indexed is a tier-one problem, and it outranks every other finding on this page.

What changes after that is the order you work in. Plenty of sites carry three or four of these together, and the ones that are pleasant to fix are almost always tier three. Confirm the cost tier first, then spend the time where the tier says to.