Text rewriting means restating an existing text in new words while keeping its meaning, as opposed to producing a fresh article from scratch. It saves a website when you need to revive outdated posts, bring dozens of pages into one voice or turn a dense legal document into plain language, and it hurts when it becomes a race for "100% uniqueness" and breeds pages that are twins in substance. Below we look at where the line runs, how Google actually treats rewritten content and how to check the result so the budget is not wasted.
An article rewrite starts from a finished source: someone else's or your own publication, technical documentation, an old service page. The writer reads it, grasps the logic and lays the same substance out differently: a new structure, new phrasing, often a new order of arguments. The facts stay, the wrapping changes. Copywriting starts from a task and an audience instead of a text: the writer gathers material, interviews the client, studies competitors and produces something that did not exist in that form before.
The third option, which people now confuse with both, is rephrasing through AI. A model does it in seconds and the output looks decent, but it knows nothing about your product, cannot tell that a number in the source is stale and happily "fills in" details that were never there. At Query we use these tools for drafts, yet the final text is always read by a person who understands the subject.
| Criterion | Rewriting | Copywriting | AI rewriting |
|---|---|---|---|
| Source of content | An existing text | Brief, interviews, research | Same text + the model |
| New information | Almost none | Yes, the main value | May be invented |
| Risk of errors | Low | Depends on the author | High without proofreading |
| Speed | Hours | Days | Minutes |
| Price | Below copywriting | Highest | Near zero + editing |
| When it fits | Updates, unification | New pages, expertise | Drafts, headline options |
The most common case: a blog that has been running for five years, where a third of the posts describe tools, prices and interfaces that no longer exist. Deleting them is painful because they hold links and rankings. Writing them again is expensive. This is where a rewrite shines: the structure has already been validated by traffic, you need to swap the outdated fragments, add fresh data and tighten the language.
The second case: a site whose texts were written by different people in different years. One page addresses the reader formally, another chats casually, a third reads like a press release. Unifying the voice through rewriting costs a fraction of full copywriting for every section, and the effect is noticeable: the visitor stops feeling like they are wandering across several different websites.
The third, less obvious one: legal and policy texts. Terms of service, privacy policy, return rules. The substance cannot change, but it can and should be phrased so that a human actually finishes reading and understands it. That is also a rewrite, with extra responsibility attached: every rewritten sentence is checked against the original by a lawyer or the owner.
The most widespread mistake: rewrite a competitor's top-ranking article, get "98% unique" from a checker and wait for rankings. Google sees a page that adds nothing to a dozen already indexed ones and simply has no reason to rank it higher. Formally the text is unique; in substance it is a duplicate. There are millions of such pages, and the latest algorithm updates target exactly them.
The second mistake: text uniqueness as a goal in itself. To squeeze out the coveted 100%, the writer breaks established terms, replaces "keyword research" with "search phrase discovery" and writes "mobile communication device" instead of "phone". The reader stumbles, the search engine cannot find the query the page was meant to rank for. The paradox: the text became more unique and worse at the same time.
The third, internal one: rewriting your own pages that already compete with each other. If you have "Air conditioner repair" and "Air conditioner repair in Kyiv" and you order a rewrite of both so they "look different", you end up with two pages for one query. The search engine picks one of them itself, and not necessarily the one you wanted. This is called cannibalization, and the cure is merging or clearly separating intent, not rephrasing.
There is no "uniqueness percentage" anywhere in Google's algorithms. There is an assessment of helpfulness: does the page answer the query better, or at least differently, than existing ones, does it show the author's experience, do users trust it. Google's official guidance on creating helpful content includes a self-check question asking whether the article merely summarizes what others have already said without adding value.
The practical conclusion: a rewritten text ranks not because it "does not match" the original but because it contains something of its own. That can be fresher information, a clearer structure, an example from your practice, a table the source lacked, or simply an answer to a question the original skipped. When a rewrite adds that layer, it works. When it only shuffles words, it is neutral at best.
A separate matter: technical duplicates. If the same text is reachable at two URLs on your site, you need rel="canonical" or a redirect, not a rewrite. Rephrasing a page to "get around" a duplicate caused by URL parameters is like putting out a fire with a hair dryer.
A plagiarism check is necessary, but not for the reason it is usually run. Its purpose is to confirm the writer did not lift chunks from other sites and hand you somebody else's text, not to prove to Google that the page is "new". Tools such as Copyscape, Quetext or the built-in checks on freelance marketplaces compare the text against indexed pages and show which fragments match and where they came from.
What the numbers mean. A score of 95% or higher is normal for a rewrite: the matches fall on terms, names and set phrases. Anything from 80 to 90% deserves a fragment-by-fragment look: it might be a quote or a spec list that cannot be phrased otherwise, or the writer might have cut corners in one section. Below 70% for a rewritten text means it is not a rewrite but copying with a cosmetic touch. And the reverse holds: 100% guarantees nothing except that no word chains matched.
Different services give different numbers on the same text because they slice it into "shingles", chains of words used for comparison, in different ways. So specify a particular tool and threshold in the brief rather than an abstract "at least 90% unique".
A rewrite is usually cheaper than copywriting: the writer skips the research stage, and that is the main saving. But the gap is smaller than clients expect, because a quality rewrite with fact-checking and restructuring is close to editorial work. A simple rewrite of a news item and a rewrite of a technical article about equipment are priced differently, and the second can easily approach the cost of writing from scratch.
Timing: one article of average length is rewritten in a day or two including proofreading. A package of several dozen pages for style unification takes weeks, and what matters there is not speed but getting the first batch approved so the rest follows the same pattern. If you would rather not manage that yourself, we handle website text rewriting end to end, including checks and publishing; and when a page needs new substance rather than a restatement, it is more honest to order website copywriting from scratch straight away.
.jpg "Website Promotion by a Freelance SEO Specialist: Pros, Risks and How to Choose")
.webp "Website Usability: Components, Metrics and the Problems That Drive Users Away")
.webp "Link Shortening: A Detailed Guide to Services, Safety and UTM Tracking")