A page should be as long as it takes to finish the job the searcher came for, and no longer. Google says the length of content alone does not matter for ranking and that there is no magical word-count target, minimum or maximum [3]. Pick length by working out what the query needs, then check your draft against that, not against a number.
What Google's guidance says about length
Google's documentation does not give a length rule. It does give a way to think about the question.
- Length alone is not a ranking input, and there is no minimum or maximum [3].
- "Main content" is any part of a page that directly helps the page achieve its purpose [1]. A paragraph that does not serve the purpose is not main content in that sense, however many words it adds.
- Quality guidance asks for originality, effort, and talent or skill. It describes automatically generating large amounts of text without manual oversight or curation as little to no effort [1].
Third-party SEO sources agree. Semrush says the ideal length depends on search intent, topic complexity, and what it takes to make a better piece than the competing pages [4]. It also says a 1,000-word page can outrank a 3,000-word article if it matches searchers' needs better [4]. Ahrefs goes further: if you are writing to reach a word count, you are creating content wrong [5].
Why word-count targets produce filler
A target like "write 2,000 words" is usually derived from averaging the top results. The average tells you what competing pages happened to contain. It does not tell you what the searcher needs.
Once a writer has a number to hit, the easiest way to hit it is to add text that costs nothing to write. The common patterns:
- A long preamble that restates the title.
- A "what is X" section on a topic the reader already knows.
- Generic lists ("10 benefits of…") with no specifics.
- Repeating the same point in the intro, the body and the conclusion.
- Sections added because a competitor has a section with that heading.
Before and after, on a page that answers "how do I check a page's canonical URL":
Before (64 words before the first useful step):
In today's world, canonical URLs are more important than ever.
Whether you are a beginner or an expert, understanding canonicals
can help your site. In this guide we will explore what canonicals
are, why they matter, and how you can check them step by step.
After (useful on line one):
To check a page's canonical URL, view the page source and search for
rel="canonical", or run the URL through URL Inspection in Search Console.
The second version is shorter and stronger. The rest of the page can then go deeper on the cases that need it.
Set depth by intent
Start from the query, not the count. Look at what the results page actually shows. Ahrefs' example is "how to tie a tie": the top result is a video, so the right answer to that query is not a long essay [5]. The dominant format tells you the depth the searcher expects.
| Intent | What a complete answer needs | Typical shape |
|---|---|---|
| Quick fact or definition | The answer, one qualifier, one example | A few short paragraphs |
| How-to with a single path | Ordered steps, the command or setting, how to verify | Numbered steps, a snippet |
| How-to with branches | Steps plus the conditions that change them | Steps plus a decision table |
| Comparison | Criteria, a side-by-side, a recommendation per case | Table plus short sections |
| Broad guide or "complete" topic | Every subtopic the reader must understand to act | Multiple sections, internal links to deeper pages |
| Transactional or navigational | Price, specs, availability, next action | Often short; clarity beats volume |
These shapes are guidance from general practice, not Google rules. The point is that "complete" has a different size for each row.
Thin versus short
A short page can be complete. A long page can be thin. Semrush defines thin content as pages that offer little or no added value and do not answer search intent or help users complete a task [6]. Length is not in that definition.
Google's spam policies give one concrete case: affiliate pages can be considered thin when they are distributed across a network without additional value, especially when they repeat the same or similar content across multiple domains or languages [2]. Google's guidance also treats mass-generated text without oversight as low effort [1]. Semrush dates Google's crackdown on thin content to the Panda update in 2011 [6].
Use these checks to tell the two apart:
| Signal | Short but complete | Thin |
|---|---|---|
| Answers the query | Yes, directly | Circles it or answers vaguely |
| Original element | A specific fact, test, example or data point | Rephrases what other pages say |
| Reader's next step | None needed, or clearly linked | Reader must search again |
| Duplication | Unique on your site | Near-copies of other pages, templated with a swapped keyword |
| Effort visible | Yes: specifics, screenshots, commands | Generic text that could sit on any site |
A 300-word page giving a store's return policy, with the exact window, conditions and how to start a return, is short and complete. A 1,500-word category page that repeats the same paragraph with a different product name on 200 URLs is thin.
A process for choosing length
- Write the query's job in one sentence. "The reader wants to find which URL Google chose as canonical and fix it if it is wrong."
- Open the results. Note the formats (video, list, tool, guide) and what the best pages cover. Note what they leave out. This is where keyword research and SERP reading overlap.
- List the questions a reader must have answered to finish the task. Each one gets a section or a sentence. Anything outside the list is a candidate for cutting or for its own page.
- Draft without a target count. Stop when the list is covered.
- Add what only you can add. Original data, a tested example, a screenshot, a decision rule. This is the part that reads as effort [1].
- Cut. Delete any paragraph that, if removed, would not change what the reader can do. Check the intro first.
- Split if it sprawls. If a section answers a different query, move it to its own page and link to it. Internal links let a hub stay focused while the depth lives elsewhere.
Structure matters more as pages get long. Clear headings let a reader jump to the part they need, so a long page does not force them to read all of it.
Fixing pages that may be thin
Audit by purpose, not by word count. Sorting your pages by length and editing the shortest first is the wrong order. Instead:
- Merge pages that answer the same query into one stronger page, and redirect the others.
- Rewrite pages that are on a real topic but add nothing. Add the specific, original material the topic needs.
- Remove or noindex pages with no purpose for a searcher, such as empty tag archives or near-duplicate templates.
- Leave alone short pages that finish their job.
Experience and demonstrated knowledge are part of what makes a page worth its length; see E-E-A-T explained for how those signals show up on the page.
What is not known
No public source gives a length at which a page becomes thin or complete, and none should be trusted if it claims to. Studies that correlate length with rankings describe what ranking pages happen to look like, not what caused them. The notes behind this article do not include such studies, so none are quoted here.
What to do next
- Pick three pages you suspect are thin. For each, write the query's job in one sentence and check whether the page does it.
- Delete the first paragraph of your longest post and see whether it still reads correctly. If it does, leave it deleted.
- Drop word-count targets from your content briefs. Replace them with a list of questions the page must answer.
- Mark short pages that already finish their job as done, and do not pad them.