You submitted your site and Google AdSense came back with: “Low value content. Your site does not yet meet the criteria we look for in publisher partners.” No specifics, no per-page feedback. Most people panic about word counts, but length is rarely the real problem.
Fastest fix: find your weakest 20-30% of pages and either add a genuinely original element to each (your own data, screenshot, opinion, or worked example) or delete them, then declare a one-sentence site purpose on the homepage and About page. That clears the majority of “low value” rejections. The rest of this guide is the full diagnosis-to-reapply path, ordered by what fixes the rejection most often.
AdSense’s program policies don’t print a tidy checklist, but the reviewer rubric in 2026 comes down to three questions: does each page add something you can’t already find on the web (the policies explicitly flag “copied content from others without additional commentary, curation, or otherwise adding value”), does the site have an obvious purpose, and can a reviewer navigate it. “Low value” almost always means one of those is missing.
What changed in 2026
If your last AdSense knowledge is from 2024, three things tightened:
- Information Gain is the bar now. Google’s reviewers (and crawlers) ask whether a page adds information that isn’t already on the top results for that query. A correct-but-redundant “How to Start a Blog” with the same 7 steps everyone else has is low value by default, even at 2,000 words.
- AI content is allowed, unedited AI content is not. Policy does not ban AI-assisted writing; it bans thin, templated, unhelpful output regardless of how it was made. The thing reviewers reward is visible human editorial investment: specific examples, first-hand observation, real data, a named author.
- E-E-A-T, with Trust weighted heaviest. Experience, Expertise, Authoritativeness, Trustworthiness. Trust pages (privacy, contact, identifiable author) and accurate claims now matter as much as the prose.
Which bucket are you in?
Run this before touching anything. Most rejections are one or two of these, not all six.
| Symptom you can observe | Likely cause | Where to fix it |
|---|---|---|
| Articles match the top Google results step-for-step | No information gain (cause 1) | Step 1 + Step 2 |
| Pages read like the same author wrote all of them, no images, no “I tried this” | Unedited AI cadence (cause 2) | Step 1 |
| A stranger can’t say who the site is for in 10 seconds | No clear purpose (cause 3) | Step 3 |
| You can’t reach every article in 2 clicks from the homepage | Not navigable / orphans (cause 4) | Step 4 |
site:yourdomain.com shows 1-screen pages, empty tag/author pages | Thin pages diluting the average (cause 5) | Step 1 + Step 5 |
| Text overlaps or is unreadable on a phone | ”Site Behavior: Navigation” (a real 2026 sub-reason) | Step 5 |
| Topic is medical/financial/crypto/borderline-download | Policy conflict, not really “value” | Step 6 |
Common causes, by hit rate
1. No information gain vs. what’s already on the web
If your page is a paraphrase of the top 3 Google results for that query, reviewers see a near-duplicate. The number of articles doesn’t help if every article reads like a slightly reworded encyclopedia entry. The 2026 framing is “information gain”: what does this page add that the existing results don’t?
How to spot it: take any article, copy the H1 into Google, open the top 3 results. If your structure, examples, and conclusions match theirs without adding anything new (your own data, your own opinion, a new framing), the page is in trouble.
2. Pure AI-paste with no editing
Reviewers spot unedited LLM cadence fast: lists of 3-5 generic items, “in conclusion” closers, no images, no first-person voice, no specific numbers. This is the single most common 2025-2026 rejection trigger. AI as a drafting tool is fine; AI as a publishing button is not.
How to spot it: read 5 of your pages out loud. If they sound like the same author wrote all five and nothing surprises you, that’s the cadence reviewers are trained to flag.
3. The site has no clear purpose
About, Contact, and a one-line “what this site is about” need to make the topic and audience obvious in under 10 seconds. A grab-bag site that posts about productivity, crypto, and dog training without a clear thread fails on “purpose.”
How to spot it: show your homepage to someone who doesn’t know you. Ask “who is this for and what will they get?” If they hesitate, your site lacks a declared purpose.
4. Pages exist but aren’t navigable
20 articles with no category pages, no internal links, no related posts, no menu — reviewers can’t browse around. The crawl looks like a flat list of orphans.
How to spot it: open your homepage. Can you reach every published article in 2 clicks? If not, you have orphan / island content.
5. Thin pages disguised as “content”
Pages with one paragraph plus an image, tag archives with 1-2 results, empty author pages, category pages with no description, and search-result pages exposed to crawlers all drag down the average. On mobile, overlapping or tiny text triggers a separate “Site Behavior: Navigation” flag, and because reviewing is largely mobile-first now, the phone layout is often what a reviewer sees first.
How to spot it: run site:yourdomain.com, click 10 random results, and open the same pages on a phone. If 3+ feel auto-generated, are 1-screen-long, or are unreadable on mobile, that’s your dilution.
6. Topic conflicts with AdSense policy
Even if the page is excellent, certain topics (medical claims, copyright-borderline downloads, regulated finance/crypto advice) won’t pass review. This is rarer and usually comes with different rejection wording — but “low value” is sometimes the catch-all.
Shortest path to fix
Ordered by ROI. Steps 1-3 fix the large majority of “low value” rejections.
Step 1: Build the “value triage” spreadsheet
List every published URL with these columns:
| URL | Word count | Original element (0-3) | Inbound internal links | Decision |
|---|
“Original element” means: your own data, your own screenshot of you doing the thing, your own opinion / framing, a quote from someone you interviewed, or a worked example you built. Generic numbered lists do not count.
Decision rule:
- 0 original elements -> rewrite or delete.
- 1 original element and under 500 words -> expand or merge.
- 2-3 original elements and 800+ words -> keep, only polish.
Aim for at least 70% of pages in “keep” before reapplying. If a page was AI-drafted, the cheapest way to add an original element is to actually do the thing once and screenshot the result.
Step 2: Rewrite the homepage and About to declare purpose
Replace any vague tagline with a one-sentence promise that names the audience and the outcome:
Bad: "Tips and tutorials about AI."
Good: "Hands-on debugging guides for indie devs shipping AI apps on Vercel and Firebase."The About page should answer: who runs it, your background (1-2 lines; a link to LinkedIn / GitHub is enough), what topics you cover, what you don’t cover, and how to contact you. If you use AI to draft, a one-line footer such as Researched and written with AI assistance, reviewed by [Your Name] is an E-E-A-T signal reviewers read as editorial oversight, not a red flag.
Step 3: Fix orphans and navigation
For every article in “keep”:
- Add it to at least one category or tag page.
- Add 3-5 internal links from related articles in.
- Add 3-5 internal links to related articles out.
A crude orphan scan for an Astro content site (finds articles never linked from another article):
node -e "
const fs = require('fs');
const path = require('path');
const root = 'src/content/articles/en';
const files = [];
(function walk(d){for(const f of fs.readdirSync(d)){const p=path.join(d,f); fs.statSync(p).isDirectory()?walk(p):files.push(p)}})(root);
const linked = new Set();
for (const f of files) {
const t = fs.readFileSync(f,'utf8');
for (const m of t.matchAll(/\/en\/articles\/([a-z0-9-]+)\//g)) linked.add(m[1]);
}
for (const f of files) {
const slug = path.basename(f,'.mdx');
if (!linked.has(slug)) console.log('ORPHAN:', f);
}
"Step 4: Add the missing trust pages
AdSense reviewers visibly check for these, and a missing one alone can read as “low value / not ready”:
/privacy/— must disclose third-party cookies and ad networks (Google and partners) and user consent for data collection. This is mandatory once you run ads./terms/— short is fine./contact/— a real email or form, not just a social link.- An identifiable author byline on every article (your name + a 1-line bio at the bottom).
You don’t need /ads.txt to get approved, but having a valid one ready at your root domain removes friction the moment ads start serving.
Step 5: Fix thin and mobile-broken pages
Delete or merge the 1-screen pages from Step 1. Then open the site on an actual phone (or Chrome DevTools device mode) and confirm nothing overlaps, text is legible, and the menu works. Mobile breakage is its own rejection bucket and the first thing a 2026 reviewer sees.
Step 6: Wait for a re-crawl, then reapply
Do not reapply the next day. Google logs the rejection, and a same-snapshot resubmit reads as “they didn’t actually fix anything.” The 2026 community and reviewer consensus is to wait at least 2-3 weeks — long enough for Googlebot to re-crawl and re-index your edits so the reviewer sees a genuinely different version of the site. Don’t gate on a calendar date; gate on whether your changes are live and indexed.
To check the crawl actually happened, open Google Search Console, run a few edited URLs through the URL Inspection tool, and confirm “Last crawl” is after your edits. Then reapply. In the optional “what changed” field, list specifics (“rewrote 38 articles, deleted 12 thin pages, added category descriptions, fixed mobile layout”). Reviewers do skim it.
How to confirm it’s fixed
Before you reapply, you should be able to answer yes to all of these:
- At least 70% of your live pages have a verifiable original element (you can point at it on the page).
- A stranger reads your homepage and correctly states who the site is for in under 10 seconds.
- Every article is reachable in 2 clicks; your orphan scan prints nothing.
/privacy/,/terms/,/contact/, and per-article author bylines all exist and load.- The site is legible and navigable on a phone.
- Search Console shows your edited pages were re-crawled after the changes.
If any answer is no, fix that first — that’s the bucket you’ll get rejected on again.
FAQ
Does word count cause “low value content”? No, not directly. A 400-word page with a real original insight can pass; a padded 2,000-word page that restates the top search results will not. Reviewers weigh information gain and originality, not length.
Can I get approved with AI-written content? Yes, if it’s edited. Policy bans thin/templated/unhelpful content regardless of origin, not AI assistance itself. Add specific examples, real data, first-person observations, and a named author, and AI-assisted pages pass routinely.
How long should I wait before reapplying? At least 2-3 weeks as of June 2026, and only after Search Console confirms your edited pages were re-crawled. Reapplying within a day or two almost always gets the same rejection because Google reviews a near-identical snapshot.
How many articles do I need before applying? There’s no official minimum, but 2026 practitioner guidance clusters around 15-25 solid posts on one clear topic, each with at least one original element. Quality and focus matter more than count.
The rejection didn’t name a page. How do I know which one to fix? You don’t get per-page feedback. Use the triage spreadsheet in Step 1 and the “which bucket are you in” table above — fix the lowest-quality 20-30% of pages plus the site-level issues (purpose, navigation, trust pages), which is what the rubric actually scores.
Is “low value content” a permanent ban? No. It’s a not-yet-ready signal, not a penalty or blacklist. Sites that fix the underlying issues and reapply after a re-crawl get approved regularly.
Prevention
- Don’t apply before you have 15-25 articles where each has at least one original element (your data, your screenshot, your opinion).
- Write an explicit one-sentence site purpose on the homepage and stay on that topic for ~90% of posts.
- Heavily edit any AI-drafted content: change the structure, add your own examples, strip “in conclusion” cadence, add a real screenshot.
- Build category and tag pages with real intro paragraphs, not just lists of links.
- Make every article reachable in 2 clicks from the homepage, and confirm the layout holds up on mobile.