An SEO issue is a technical or editorial choice that stops Google from crawling, indexing or showing a page for the search it's meant to capture. That definition leaves out a good part of what tools call an error: a 72-character title or a missing meta description shows up in red in the reports and often doesn't move a single position.
SE Ranking analysed 418,125 audit reports generated by its own tool: the most widespread problem, images without alt text, affects 74.43% of sites. The most damaging one for traffic, an important page left out of the index, can't appear in a frequency ranking at all, because the crawler doesn't know which pages bring in customers.
Below, SEO issues are split into two groups: the frequent ones and the ones that cost traffic. They overlap less than you'd think, and neither group includes the SEO challenges that have nothing to do with the site itself, such as competing with the big portals.
The most common SEO issues, measured across 418,125 audits
SE Ranking's study, published in March 2025 and updated in May, is the largest dataset we found with percentages issue by issue. The sample is global, not UK-only: it measures how common a problem is across websites in general. Next to it we've put the figure from HTTP Archive's Web Almanac 2025, which measures similar items on millions of real pages with a different method, and our own assessment of how much each item weighs on rankings.
| Issue flagged | Sites with the problem (SE Ranking) | Web Almanac 2025, mobile pages | Does it affect rankings? |
|---|---|---|---|
| Images without alt text | 74.43% | alt present on 60% of images (median) | Only in image search |
| Pages with no incoming internal links | 69.32% | n/a | Yes |
| Title too long | 67.52% | title present on 98.54% of pages | No |
| Missing meta description | 65.38% | meta description on 67.2% of pages | No, it affects the click |
| Missing H1 | 54.67% | H1 on 70% of pages | A little |
| Pages set to noindex | 50.58% | noindex on 2.4% of pages | Yes, if the page matters |
| Pages returning 4xx | 35.73% | n/a | Yes, if they have links or traffic |
| Slow pages (lab test) | 34.54% | good Core Web Vitals on 48% of pages | A little, beyond a threshold |
| Redirect chains | 21.58% | n/a | A little, except in migrations |
| No sitemap | 14.78% | n/a | Rarely |
| Identical duplicate content | 5.04% | n/a | Rarely |
The column that matters is the last one. Of the five most frequent items, only one has a direct effect on visibility: pages that no internal link reaches. The other four are the ones that fill automated reports and, as the next section shows, the first page of Google too.
Do the pages that explain SEO issues have them?
On 10 October 2026 we took the six organic results on Google UK for "seo issues" that are articles or product pages (leaving out a Reddit thread) and downloaded them from a server with a declared user agent. Then we checked the items those same pages list as errors.
| Position | Page | Title length | Meta description | H1 | Images with empty or missing alt |
|---|---|---|---|---|---|
| 1 | screamingfrog.co.uk | 54 characters | yes | 1 | 706 of 714 |
| 3 | semrush.com | 49 characters | yes | 1 | 0 of 28 |
| 4 | seoclarity.net | 50 characters | yes | 1 | 4 of 46 |
| 5 | linkedin.com | 44 characters | yes | 1 | 34 of 63 |
| 6 | edge45.co.uk | 50 characters | yes | 1 | 6 of 13 |
| 7 | wellows.com | 61 characters | yes | 1 | 0 of 100 |
On titles, the UK page is tidier than the Italian one we checked a week earlier on Google Italy, where three of the top six went over 70 characters, the threshold above which almost every crawler flags an error. Here none does: the longest is 61. The rule hasn't changed, though. Google says in its title link documentation that there's no limit on title length; in the results it's simply truncated to fit the width of the screen. A long title is a presentation problem, like the one Google decides to rewrite (here's why that happens).
Alt text is the clearer case. The page in position 1 is Screaming Frog's list of the 300-plus issues its crawler flags, missing alt text included, and 700 of its 714 images have no alt attribute at all: they're the little severity icons next to each issue. For decorative images like icons and dividers, the W3C recommends exactly alt="", so that screen readers skip them. An empty alt is correct practice, and even a missing one hasn't kept the page off the top spot. Our guide to alt text separates the cases where it really matters.
The last finding is the most telling. Leaving aside Screaming Frog's product page and LinkedIn's page of collected posts, four results are articles. Semrush's, published in April 2026, doesn't give a single figure with a source, seoClarity's has one (Google's "more than half" of sites on mobile-first indexing, from 2018), Edge45's gives percentages from its own case studies without saying how many, and only Wellows quotes the Web Almanac. On this search Google isn't rewarding the most precise page: it's rewarding the one that matches the intent, a list of issues, on a strong domain.
The SEO issues that really cost traffic
There are six, in order of damage. A crawler can't spot the first three, because seeing them means knowing which pages bring in customers and for which searches.
1. Keeping a page that matters out of the index
50.58% of the sites SE Ranking analysed have at least one page set to noindex. It's almost always deliberate: basket, account area, thank-you pages. The damage starts when noindex lands on a service or category page, usually after a release or a theme change. On WordPress all it takes is the "Discourage search engines from indexing this site" box left ticked after moving from staging.
There's also the opposite mistake: blocking a page in robots.txt to remove it from the results. Google explains that if the crawler can't access the page it will never see the noindex, and the page can stay in the results. You check this in the Page indexing report in Search Console; the difference between the two directives is in our guide to noindex.
2. Writing pages nobody searches for, or with the wrong intent
Ahrefs measured that 96.55% of the pages in its index, around 14 billion, get no traffic from Google at all. The first cause it gives is lack of demand: nobody searches for that topic. Another is the wrong intent, for example a product page where Google shows buying guides.
No audit tool flags this, because technically the page is perfect. You find it by comparing the queries in Search Console with the format of the results on page one, the method described in our guide to search intent.
3. Two pages competing for the same search
It happens through accumulation: an article from 2021 and one from 2025 on the same subject, or a service page and an article with the same title. Google tends to show one at a time, and not always the one you'd want. In the Search Console Performance report, filter by query and open the Pages tab: if two URLs take turns on the same impressions, it's worth merging them with a 301 redirect and keeping the stronger one.
SE Ranking finds duplicate content on only 5.04% of sites because it looks for identical text. Cannibalisation between different pages on the same topic is something no crawler sees. The difference between the two cases is explained in our guide to duplicate content.
4. Migrating without a redirect map
35.73% of sites have pages returning a 4xx error and 21.58% have redirect chains. On their own these are minor problems. They become serious at one specific moment, a change of domain, CMS or URL structure: that's when every page with traffic or backlinks that ends up as a 404 loses what it had built up. The rule is a page-by-page map of 301 redirects, prepared before the release. If the company name changes too, follow the rebranding checklist.
5. Leaving pages without internal links
It's the only issue at the top of both rankings, frequency and damage. 69.32% of sites have pages with no incoming internal links at all, and 67.11% have pages with just one. Google struggles to find an orphan page and, when it does, the page gets little authority from the rest of the site. For an order of magnitude, in the Web Almanac 2024 the median page has 41 internal links, rising to 129 on the thousand most visited sites.
The fix is cheap: start from the pages that generate enquiries and check that they're linked from your most visited content, with anchor text that describes the destination. The rules are in our guide to internal links.
6. Slowness measured in the field, not in the lab
These numbers need reading carefully. SE Ranking finds an LCP above 2.5 seconds on 26.89% of sites in the lab test, but on only 9.46% with real user data. According to the Web Almanac 2025, 48% of mobile pages pass all three Core Web Vitals thresholds, against 56% on desktop. A low PageSpeed Insights score isn't necessarily a problem: it becomes one when the data Chrome collects from real users, the data in Search Console's Core Web Vitals report, falls outside the thresholds. How to read the two measures is explained in our guide to Core Web Vitals.
Issues that tools flag and Google ignores
This list saves more hours than any other. Every item is backed by a Google document, not an opinion.
- Missing meta description (65.38% of sites). Google uses it for the snippet when it thinks it fits; rankings don't change. Write one on the pages that matter because it affects the click: how to do it is in our guide to the meta description.
- More than one H1, or headings out of order. In its SEO starter guide Google says it doesn't matter if headings are out of order and that there's no ideal number of headings per page.
- Keyword density. There's no percentage to hit. Keyword stuffing does exist, and the same guide lists it as against the spam policies.
- Word count. In its guidance on helpful content Google asks whether you're writing to a particular word count because you've heard Google has a preferred one, and answers: no.
- Missing meta keywords. Google Search doesn't use the meta keywords tag.
- E-E-A-T as something to optimise. Google says that E-E-A-T itself isn't a specific ranking factor.
More beliefs of this kind, each with the document that disproves it, are collected in our article on SEO myths.
A case: the problem was a page that didn't exist
When we started working with Macropix, a Milan manufacturer of custom LED screens, in 2020, the main problem didn't show up in any technical audit. For "ledwall" (LED wall), 4,400 searches a month on Google Italy, the site didn't rank because it had no page built for that search. For "monitor pubblicitario" (advertising monitor) it was in position 88, on page nine.
By 2025 "ledwall" was in position 1 and "monitor pubblicitario" in position 2. On the ledwall cluster, in July 2026, the domain had a share of voice of 25.55% against amazon.it's 18.08% (Semrush data). The table with all nine keywords is in the manufacturing case study.
The most expensive mistake, in that case, wasn't a wrong tag. It was a search with real demand that the site didn't answer with any page.
How to find your site's SEO issues in an hour
In the UK, 81.4% of businesses with 10 to 49 employees had a website in 2019, according to the ONS E-commerce and ICT activity release (the latest edition, published in February 2021), and in many of them nobody checks it regularly. For a first pass, Search Console and an hour are enough, in this order:
- Page indexing: look for important pages among the excluded ones and read the reason.
- Performance filtered by page: do the pages that generate enquiries get impressions? If they don't, the problem is further upstream, in the index or the intent.
- Performance filtered by query: see whether two URLs take turns on the same search.
- Links: in the internal links report, check how many the pages that sell receive.
- Core Web Vitals: only look at URLs with real user data rated as poor.
Whatever is left, long titles, missing alt text on decorative images, meta descriptions, goes to the back of the queue. If the site has already lost traffic the route is different, and it's described in our guide to a website traffic drop. For a full check with priorities in order there's our SEO audit guide; if you'd rather someone else did it, that's the work we do in our SEO service.
The wrong number to put in a report
Something the industry rarely says: the number of warnings in a crawler report tends to go up as a site grows, not down. More pages, more images and more templates mean more items in red. That's why an audit that ends with "errors reduced from 1,200 to 300" tells you nothing about traffic. The number to put in an SEO report is a different one: how many of the pages that bring in enquiries are indexed, reached by internal links and ranking for the right search.
FAQs on SEO issues
By frequency, according to 418,125 audits analysed by SE Ranking: images without alt text (74.43% of sites), pages with no incoming internal links (69.32%), titles that are too long (67.52%) and missing meta descriptions (65.38%). Of these, only missing internal links have a direct effect on rankings.
No. Google states that there's no length limit for titles: the part that doesn't fit in the space available is truncated in the results. It's still worth putting the words that matter at the start.
Usually not. When Google finds identical pages it picks one as canonical and shows that one. The practical damage is different: signals and links split across several URLs and, sometimes, the wrong version in the results.
It depends on the size of the site. For a company website of a few hundred pages, a monthly Search Console check and a full technical check every quarter are enough, plus an extra check after every release or migration.
Partly, yes. A page that is out of the index or unreachable stays invisible to systems that rely on search engine results as well. On top of that there's a new mistake: blocking AI crawlers in robots.txt without having decided to.
