SERP Competitor Analysis
Find out what the pages on page one actually contain, then see exactly where your page sits against the middle of them, measure by measure.
This tool runs on your own search data account key, saved in the Your keys card on the Your SEO Tools page of your dashboard; each run makes one search on that account.
What gets counted on every ranking page
Each page is fetched and parsed the way a crawler sees it. Nothing is scored or weighted: you get the counts, and the conclusion stays yours.
- Position, domain and final URL
- Where the page sits in the organic results and the address it lands on after redirects.
- Title and meta description
- The text of each and its character count, so you can see how the top results frame the same search.
- H1 and the real H2 text
- The main heading, how many H1s the page has, and up to 40 of its H2 headings word for word.
- H2 and H3 counts
- How deeply each page breaks its subject into sections.
- Word count and words per H2
- Visible words on the page, and that total divided by the number of H2 sections.
- Images and alt text
- How many images each page carries and how many of them have alt text filled in.
- Internal and external links
- Links that stay on the same site against links that point to other domains.
- Schema types
- Every schema.org type found in the page's JSON-LD, including types nested inside a graph.
How to run a SERP comparison
Type the search
Use the exact phrase a customer would enter, not a topic label. One phrase per run.
Choose the market and device
Pick the country and desktop or mobile. Results differ between both, so match where your visitors really search.
Set how many pages to read
Anything from 3 to 10. Ten gives the steadiest medians; three is quicker when you only care about the very top.
Add your own page
Optional. Paste your URL and every measure gets a Yours, Their Median and Difference column.
Ranking pages are read three at a time, with about 15 seconds and the first 2 MB of HTML allowed for each.
Making sense of the comparison
- Pages Read
- How many ranking pages could be fetched out of those returned. Medians come only from the pages that were read.
- Your Page Against The Middle Of Theirs
- Your count, their median and the gap. A plus means you have more than the middle page, a minus means fewer, and level means equal.
- Across The Pages That Were Read
- Median, average, lowest and highest for each measure. A wide spread between lowest and highest tells you length is not what this result rewards.
- Structured Data On The Pages That Rank
- Each schema type, how many ranking pages use it and the share. Types used by at least half of them that your page lacks are flagged.
- Not read
- The page returned an error, timed out, blocked automated readers or was not a public address. It is never counted as an empty page.
- Your Outline
- Your own H2 headings in order, ready to hold against the headings the ranking pages use.
Gaps this comparison often reveals
Your page covers far fewer subtopics than the top results
Fix: Compare your H2 count and outline with theirs. Add sections for the questions most ranking pages answer and you skip, rather than padding existing ones.
Most ranking pages use a schema type yours does not
Fix: Check whether the type fits your content, such as FAQPage, Product or HowTo. Add it only where it truthfully describes the page.
Very few internal links compared with competitors
Fix: Link out to your related guides, categories and product pages with descriptive anchors. It helps readers and spreads authority across the site.
Images without alt text
Fix: Describe each meaningful image in a few plain words. Leave decorative images with an empty alt instead of keyword lists.
Title much longer or shorter than the rest
Fix: Look at how the ranking titles are phrased and sized, then rewrite yours to lead with the search phrase and a clear reason to click.
Who gets the most from it
- Content writers sizing up a search before drafting a single word.
- SEO consultants explaining to a client why a page sits on page two.
- Site owners refreshing an older article that has slipped down the results.
- Agencies building briefs from evidence instead of guesswork.
A median is context, not a target
Matching the median word count of the top ten does not make a page rank, and a shorter page can win when it answers faster. Use the differences to spot missing subjects and structure, then judge each one on whether it helps the reader.
SERP Competitor Analysis FAQ
Short answers about the SERP Competitor Analysis.
The organic results for your search in the country and on the device you chose. Ads and other search features are not counted as competitors.
Live search results come from a paid search data account. The tool uses the key you saved in your dashboard, so the search is made on your account and the results belong to you.
Large sites often block automated readers, and some pages are too slow or return an error. Those rows say why, and the medians are worked out from the pages that could be read.
No. A weighted score built from counts reads like a measurement but is an opinion. You get what was counted and how your page differs.
Only public pages can be fetched. Publish the draft on a public URL, or run the comparison without your page and use the medians as a guide while writing.
Yes. It counts words, headings, links and markup, so it works for pages in any language your chosen market returns.
Related SEO tools
Content Outline Generator
A content outline from the pages that already rank.
Your own keyRank Tracker
Where your site ranks for up to 25 keywords.
Your own keyOn-Page Audit Pro
70 checks on one page, scored against the basics.
1 free run a daySchema Intelligence Analyzer
Score the structured data already on your pages.
1 free run a day