The most common website copy errors (and how to catch them)
Most of the copy mistakes that reach a live website are not exotic — they are the same handful of error types, recurring on the pages nobody re-reads. Knowing the categories makes them far easier to catch. Here is the practical taxonomy, with examples of each and where it tends to hide.
Last reviewed: 2026-06-21
The most common website copy errors fall into six recurring kinds: spelling errors and typos (including homophones a spell checker passes, like "your" for "you’re"); grammar slips (subject-verb disagreement, wrong tense, doubled or dropped words); punctuation and spacing problems (missing or stray commas, double spaces, mixed quote and dash styles); unclear sentences a first-time visitor has to reread; inconsistent style (the same term, capitalization, or format written two ways across pages); and leftover placeholder text (lorem ipsum, "your text here", untouched theme defaults).
They cluster on the pages that get built once and never revisited — deep service pages, older posts, templated and auto-generated pages — because the home page is re-read a hundred times and the long tail is not. Catching them reliably means checking every published page, not just the obvious ones. Verant crawls the whole live site and flags exactly these six kinds, quoting your exact text on each, and verifies each correction with a second AI agent before showing it.
Spelling errors and typos
The most familiar category: transposed letters ("the" as "teh"), doubled or dropped words ("the the", "to to"), and outright misspellings. The dangerous subset is the homophone — a correctly spelled word that is the wrong word: "your" for "you’re", "its" for "it’s", "their" for "there", "complimentary" for "complementary". A basic spell checker passes every one of these, because each is a real, correctly spelled word; only reading for meaning catches them.
Grammar slips
Subject-verb disagreement ("the list of features are growing"), tense shifts mid-paragraph, dangling modifiers, and pronoun mismatches. On a website these often survive because copy is edited in fragments — a sentence is half-rewritten, a list item is added without re-reading the lead-in — and the seam is grammatically broken but visually unremarkable. They read as fine in a quick skim and wrong on a careful read.
Punctuation and spacing
Missing or extra commas, a stray double space, an apostrophe in the wrong place ("its" vs. "it’s" again, now as punctuation), and inconsistent typography: straight quotes mixed with curly ones, a hyphen used where an en dash or em dash belongs, "10-20" on one page and "10–20" on another. Web copy is especially prone to this because it arrives from many sources — a word processor, a CMS field, pasted text — each carrying its own marks.
Unclear sentences
Copy that is technically correct but hard to read: a sentence a first-time visitor — who lacks your context — has to read twice, an ambiguous pronoun, a noun pile-up, or a clause that could be parsed two ways. Clarity problems are not "wrong" in the grammar sense, which is why they slip past error-only checks, but they cost the reader effort exactly where you most want them to keep going.
Inconsistent style
The same choice made two different ways across a site: "sign up" here and "signup" there, "email" on one page and "e-mail" on another, the Oxford comma on the home page and not on the pricing page, a feature called "Auto-Save" in one place and "autosave" in another. None is wrong in isolation; the inconsistency is the error. It accumulates as a site grows and copy is written by different people at different times, and it quietly makes polished writing read as careless.
Leftover placeholder text
Filler a template shipped with that never got replaced: lorem ipsum dummy paragraphs, literal prompts like "your text here" or "Add your content here", framework starters like "Hello world!", and stock demo copy. It is grammatically fine and correctly spelled, so the usual checks pass it, and it survives most often in low-attention sections and on templated pages. It is its own category precisely because no spelling or grammar rule will ever flag it — you have to recognize the copy as a default.
What is NOT a copy error
Several common website problems are real but are not copy errors, and a proofreading pass does not catch them: broken links and redirects (a link checker’s job), wrong facts and stale figures (a human fact-check — proofreading reads how copy reads, not whether it is true), accessibility issues like missing alt text or poor contrast (an accessibility audit), and SEO metadata problems (an SEO tool). Keeping these on separate lists matters: a clean proofreading pass does not mean the page is otherwise correct, and an SEO or link tool will not catch a typo. Each check has its own tool.
How to catch these errors across a whole site
- 1
List every public page, including the templated and auto-generated ones — the long tail is where these errors hide, not the home page.
- 2
Check the rendered, published page (what a visitor sees), not a draft or content export — the live copy is whatever the CMS, theme, and JavaScript actually produced.
- 3
Run all six kinds in one pass rather than only spell-checking: paste your URL into Verant and run a full-site scan so every page is checked for spelling, grammar, punctuation, clarity, style, and placeholder text.
- 4
Read the homophones and clarity flags for meaning — these are the ones a spell checker structurally cannot catch, so they reward a careful review.
- 5
Run the non-copy checks — links, facts, accessibility, SEO metadata — with their own tools, then fix the copy in your CMS and re-scan to confirm.
How verification works
Most proofreading agents show you every suggestion and make you sort the good from the bad. Verant runs an adversarial second pass — Claude Sonnet proofreads, then GPT-5 tries to break each correction. What survives is what we show you. Verbatim is sacred: every flag quotes your exact text; we never auto-apply fixes.
Keep going
- Website proofreading software What website proofreading is, the full "what Verant catches" taxonomy, and the explicit list of what it does not do.
- Homophone The correctly spelled wrong word — "your" for "you’re" — that a spell checker passes and a careful read catches.
- Pre-launch website proofreading checklist Run the six-kind copy pass before a site goes live, with the non-copy checks flagged separately.
Related reading: website proofreading software, homophone, and pre-launch checklist.
Frequently asked questions
What are the most common copy errors on websites?
They fall into six recurring kinds: spelling errors and typos (including homophones like "your" for "you’re"), grammar slips, punctuation and spacing problems, unclear sentences, inconsistent style, and leftover placeholder text. They cluster on the pages built once and never revisited.
Why does a spell checker miss so many website errors?
Because several of the most common errors are correctly spelled. Homophones ("its" for "it’s"), inconsistent style, unclear sentences, and leftover placeholder text all pass a dictionary check — the words are real, just wrong or inconsistent in context. Catching them means reading for meaning and consistency, not just spelling.
Where do website copy errors usually hide?
On the pages nobody re-reads: deep service pages, older posts, and templated or auto-generated pages cloned from a starter. The home page is re-checked constantly; the long tail is built once and shipped. A full-site crawl checks all of them the same way.
Is a broken link or a wrong fact a copy error?
No. Broken links, wrong facts, accessibility issues, and SEO metadata problems are real but separate from copy errors — each needs its own tool (a link checker, a human fact-check, an accessibility audit, an SEO tool). Proofreading checks the six kinds of copy issue; it does not validate links, facts, accessibility, or metadata.