On-Page SEO Checklist

I do not tick boxes in order. I start with the live SERP for the target query and write down the job the page must do in one sentence. Then I run through three layers: intent, content, technical.

Intent first. If the query is "best noise cancelling headphones", the page needs a comparison with pros, cons and prices. A specs list alone fails. I check the title and H1 answer that job within a glance. For a review page I recently optimised, the original H1 was "Noise Cancelling Headphones" – too vague. I changed it to "Best Noise Cancelling Headphones in 2025: Reviewed & Compared". Clicks jumped 12%.

Content next. The primary keyword must appear in the title, H1 and first 100 words naturally. I count words in the first paragraph – if the keyword is missing after 80 words, I rewrite. I then skim the H2 hierarchy. Each H2 should support the main topic, not introduce new themes. I once saw a page about "on-page SEO" whose H2s included "What is a landing page" – that belongs on a different page, not here.

Technical last. I check the URL slug is short and hyphenated (e.g., /on-page-seo-checklist/, not /blog/2025/03/on-page-seo-checklist-for-beginners/). I open the page source and confirm there is no accidental noindex. I test mobile layout: if the answer is below three scrolls, the page loses. I run images through a compressor and write alt text that describes the image for accessibility, not keyword stuffing. For decorative images, I use alt="". That is not lazy; it is correct.

I link internally to the next useful page. For this note, I would point to [on page SEO](/on-page-seo/) for the broader strategy and [image SEO](/image-seo/) for the technical side of images. External links go to trustworthy sources that back a claim, like Google's own guidance on [title tags](/title-tag/). That is the whole list. It takes me about 30 minutes per page.

The hardest lesson for me was learning when to stop optimising. I used to spend hours polishing meta descriptions, tweaking H2 keywords and adding internal links to every third paragraph. For most pages, the return on that effort is near zero.

Here is my rule: if the page answers the query better than the top three results on SERP, and the title and H1 clearly match intent, I stop. I do not touch the meta description if it already summarises the page without baity language. I do not run a grammar checker unless the prose is genuinely broken. I do not add schema unless I see that the SERP already shows rich results for that query type.

I remember a blog post about SEO copywriting that ranked #12 despite a perfect on-page score. The problem was intent: the query was "how to write SEO copy" but the page was a list of tools, not a tutorial. I rewrote it to follow the AIDA model and it climbed to #4. That was a content fix, not a checklist fix.

A counter-argument: "What about page speed?" Yes, it matters, but I treat it as a separate audit, not part of a per-page checklist. I run Lighthouse once a month for the whole site. If I check Core Web Vitals on every page I touch, I never get through the backlog.

So when to stop? When the page passes the intent test, the title and H1 are clear, the primary keyword is in the first 100 words, the slug is clean, and images have alt text. Everything else is polish that can wait until you see it underperforming. I have checked many pages with no meta description that still got a good CTR because the title did the work. That is not ideal, but it is not fatal.

I have made enough mistakes with checklists to fill a separate note. The biggest: I treated the checklist as a score. I gave myself a 9/10 on a page that ticked every box – title, H1, keywords, links, images – and the page flopped. Why? Because the content did not address what users actually asked. The checklist measured form, not substance.

Second mistake: I optimised for a keyword the page did not answer. I once wrote a page optimised for "SEO content writing tips" but the actual content was about content calendars and blog scheduling. The keyword was in the title and H1, but the body never taught writing tips. The bounce rate was 80%. I had to rewrite the entire page. Now I check intent alignment before I even draft the checklist.

Third mistake: I shipped a page without clicking through on mobile. The checklist had a line for "mobile friendly", but I only ran the Google Mobile-Friendly Test. The page looked fine in the emulator. On a real iPhone 12, the main answer was hidden behind a sticky header and a massive hero image. That cost me ranking for a month.

Fourth mistake: I used to think every H2 must contain the primary keyword. That leads to awkward, forced headings. Logical structure matters more. For an article on [blog SEO](/blog-seo/), I used H2s like "Blog SEO strategies" and "Blog SEO tools" when the natural flow should have been "Why blog structure matters" then "Tools I use". Google understands synonyms and context.

Finally, I overused internal links. I added links to every sentence that contained a relevant phrase, which diluted the value and made the page look spammy. Now I link only when the next page genuinely helps the reader continue, like a link to a [content audit](/content-audit/) tool when discussing identifying weak pages. Fewer links, better context.