In this article
Fix three things and most pages improve inside a month: the title tag, the first sixty words, and the internal links pointing at the page. That is the short version of the on-page SEO tips below. Everything else — heading hierarchy, alt text, schema, readability — still matters, but it compounds more slowly. I've audited a few hundred pages over the years, and the order you fix things in decides how quickly you see movement in Search Console. Start with the levers that touch the snippet.
What Are the Most Important On-Page SEO Tips?
The highest-impact on-page SEO tips are: put the primary keyword near the front of a click-worthy title tag, answer the search query in the opening two sentences, add three to five relevant internal links from strong existing pages, use one H1 with a logical H2/H3 structure, and cover the subtopics competitors rank for.
Notice what's missing from that list. Keyword density targets. Exact-match anchor text everywhere. Stuffing the meta keywords tag, which Google stopped using in 2009. Those obsessions survive in old checklists and waste real hours.
Here's a concrete example of prioritisation. A B2B client had a comparison page stuck at position 12 for a 4,000-searches-per-month term. We rewrote the title from "Product Comparison | Brand Name" to a version leading with the query, moved the summary table above the fold, and added links from four blog posts that already ranked. Position 4 within nineteen days. Nothing else changed — no new backlinks, no word count inflation.
The lesson? On-page work is triage, not a to-do list you march through top to bottom. Identify which signal is weakest on that specific URL, fix it, then measure. If you want the full sequence written out, our step-by-step on-page SEO playbook walks through the same process in more depth.
Write Title Tags That Earn Clicks, Not Just Keywords
Your title tag is the only on-page element that doubles as ad copy. Treat it that way.
Google has rewritten titles since its August 2021 update, and a widely cited Zyppy study by Cyrus Shepard found rewrites happened on roughly six out of ten queries tested. You can't stop that entirely. You can make rewrites less likely by keeping titles under about 60 characters, avoiding pipe-stuffed keyword lists, and making sure the title genuinely reflects the page's H1 and content.
A formula I keep returning to: primary keyword + differentiator + year or format. "On-Page SEO Checklist: 27 Checks for 2026" beats "SEO Checklist — Best SEO Tips and Tricks | Agency Name" every time. The first promises a countable deliverable. The second promises nothing.
Two gotchas nobody warns you about. First, if your CMS appends the brand name automatically, your carefully-written 58-character title becomes 74 characters and gets truncated mid-word. Check the rendered source, not the CMS field. Second, seasonal titles with a year in them need a calendar reminder — a page titled "Best SEO Tools for 2024" bleeds clicks silently for months while its rankings stay flat.
Test titles the way a paid-search manager would. Pull the query's impressions and CTR in Search Console, change one variable, wait fourteen days, compare like-for-like. If impressions hold and CTR climbs, keep it. If both fall, you drifted from intent and Google noticed.
Build a Heading Hierarchy That Mirrors Search Intent
Headings are not decoration. They're a table of contents that both readers and crawlers use to decide whether the page covers the ground.
One H1 per page, matching the title tag closely but not identically. Then H2s for each major subtopic, H3s only for genuine children of those subtopics. Skipping from H2 to H4 because the H4 style looks nicer is a classic developer-driven mistake — fix the CSS, not the semantics.
To decide which H2s you actually need, do this: search your target query, open the top five results, and list every subheading they use. Cluster them. Anything that appears in three or more results is table stakes. Anything that appears in one is your opportunity to differentiate. This takes twenty minutes and outperforms guessing.
Phrase at least a couple of headings as the questions your readers type. "How long should a title tag be?" pulls more extractable answers into AI Overviews and featured snippets than "Title Tag Length Considerations" ever will. Then answer immediately underneath, in plain sentences, before you elaborate.
One more thing. Long pages need jump links. Add an anchored contents list at the top for anything past 1,500 words — it improves the reading experience, and Google sometimes surfaces those anchors as sitelinks under your result. Free real estate on the SERP for about ten minutes of work. Your keyword research process should feed this structure directly, not sit in a separate spreadsheet nobody opens.
Answer the Query in the First Sixty Words
Most writers bury the answer. They warm up with context, define the term, mention how much the industry has changed. By then the reader has bounced and the snippet-extraction engine has grabbed something useless.
Lead with the payoff. If the query is "how long does SEO take," your first sentence should say three to six months for competitive terms, faster for long-tail. Context follows.
This isn't only about bounce rate. Google's systems and the answer engines built on top of them lift self-contained passages. A crisp 45-word answer sitting directly under a question heading is far more liftable than the same information scattered across four paragraphs. I've watched pages win featured snippets purely from reformatting existing content into that shape.
Practical constraints that help: keep the answer paragraph between 40 and 60 words, use the query's phrasing at least once, avoid pronouns that depend on earlier sentences, and don't split the answer across a bullet list and a paragraph.
A small habit that pays off — after drafting, read only your opening paragraph aloud and ask whether it would satisfy someone who read nothing else. If the answer is no, rewrite it. On a client's 90-post blog we did this to the twelve pages with the highest impressions and lowest CTR. Average CTR on those URLs went from 1.9% to 3.4% over six weeks. Same rankings. Just clearer openings.
Internal Links: The Cheapest Ranking Lever You Own
Nothing in on-page SEO gives a better return per hour than internal linking, and almost nobody does it systematically.
Here's the method. Export your pages by organic traffic. Take the top twenty. For each new or underperforming page, find three of those strong pages where a link would make contextual sense, and add it in the body copy — not the footer, not a related-posts widget. Body links carry more weight and get clicked more.
Anchor text should be descriptive and varied. "Click here" tells crawlers nothing. Exact-match anchors on every link look manipulative and read badly. Mix natural phrases: "technical crawl issues," "our on-page audit process," "keyword clustering."
Watch depth. Any page more than three clicks from your homepage is getting crawled less often — run a Screaming Frog crawl and sort by crawl depth to find the orphans. I've found entire product categories sitting at depth six on ecommerce sites because they were only reachable through a faceted filter.
Orphan pages are the quiet killer. Zero internal links means Google discovers the URL through your sitemap alone and treats it as low priority. Fix those first, before you write anything new.
One caution from experience: don't add fifteen internal links to a 900-word post because a tool told you to. Links should serve the reader's next question. If you'd struggle to explain why a link is there, delete it. Pair this work with a proper technical SEO review so crawl budget isn't wasted on redirect chains.
Readability Changes Behaviour, and Behaviour Changes Rankings
Readability is not a direct ranking factor. It is, however, the thing that determines whether someone stays long enough to convert or scroll to your call to action — and those behavioural patterns feed back into how Google evaluates satisfaction.
Concrete formatting rules I apply to every draft:
- Paragraphs of one to four sentences. Never a wall of eight.
- Sentence average under 20 words, with deliberate short ones for rhythm.
- Bold the phrase a skimmer needs, not whole sentences.
- One idea per paragraph. If it needs "also," it's probably two paragraphs.
- Body text at 18px minimum with line height around 1.6.
That last point gets ignored constantly. A brilliantly written article set in 14px grey-on-grey is unreadable on a phone, and roughly two-thirds of search traffic is mobile according to Google's own reporting. Design decisions are SEO decisions.
Skip the obsession with Flesch scores. Aiming for grade 8 reading level makes sense for consumer content; forcing it on a technical audience produces condescending prose. Write for the actual reader. An SEO director reading about log file analysis doesn't need "crawling means visiting your website" explained.
Tables deserve a mention. For any comparison — tools, pricing, before-and-after metrics — a table beats prose and gets extracted into answer boxes far more readily. Build them in HTML, not as an image.
Images, Alt Text and the Speed Signals That Matter
Images are the most common source of both accessibility failures and Core Web Vitals problems on content pages.
Alt text first. Describe what the image shows, in a sentence a screen reader user would find useful. "Screaming Frog crawl report showing 47 broken internal links" is good. "seo image on page seo tips seo tools" is spam, and it's the kind of thing that quietly signals low quality across a whole site. Decorative images get an empty alt attribute — alt="" — not a keyword.
Then the technical side. Serve WebP or AVIF, define explicit width and height attributes to prevent layout shift, lazy-load everything below the fold, and never lazy-load your hero image — that's the single most common cause of a slow Largest Contentful Paint on blog templates.
Google's thresholds are public and specific: LCP under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, Cumulative Layout Shift under 0.1. INP replaced First Input Delay as a Core Web Vital in March 2024, so if your monitoring still reports FID, it's out of date.
File naming still helps image search. on-page-seo-heading-structure.webp beats IMG_4471.jpg. Small thing, five seconds of effort, and image search sends real traffic in visual niches — recipes, interiors, fashion, product photography.
Compress before upload. A 3MB screenshot in a blog post is inexcusable in 2026.
How Do You Know Your On-Page SEO Changes Worked?
Compare the same URL's average position, impressions and CTR in Google Search Console for the 28 days after your change against the 28 days before, filtered to the queries you targeted. Allow two to four weeks for recrawling and reindexing. Isolate one variable per change so the attribution stays clean.
Annotate everything. A dated changelog — spreadsheet, Notion, whatever — with URL, what changed, and the date, saves you months of guessing later. Without it, you'll be staring at a traffic dip in November wondering whether it was your title rewrite or a core update.
Check indexing before you conclude anything failed. Use the URL Inspection tool to confirm Google has actually fetched the new version. I've seen teams declare a rewrite unsuccessful when the cached version was still six weeks old.
Metrics worth watching beyond position: number of unique queries the page ranks for (a good proxy for topical coverage), scroll depth, and assisted conversions. A page that drops from position 3 to 5 but doubles its query count is usually winning, not losing.
Set realistic timelines. Title and snippet changes show within days to two weeks. Content depth and internal linking changes typically take four to eight weeks. Anything you judge after 72 hours is noise. Run everything through an on-page SEO audit checklist quarterly rather than reacting to daily rank fluctuations.
Your Next Two Weeks
Pick ten pages with high impressions and CTR below 2%. Rewrite their titles and opening paragraphs. Add three body internal links to each from your strongest existing content. Log the date. That's it — one focused fortnight, no new content required.
Most sites have more upside sitting in pages that already rank on page two than in anything they haven't written yet. These on-page SEO tips work best applied narrowly and measured properly, not sprayed across 400 URLs in a weekend.
Frequently Asked Questions
How many times should the primary keyword appear on a page?
There's no magic number. Aim for natural inclusion in the title tag, H1, first paragraph, one or two H2s, and the URL — then write normally. On a 2,000-word page that usually lands around 10 to 20 mentions, roughly 1%. Google's systems understand synonyms and related entities, so forcing exact matches hurts readability without improving relevance.
Do meta descriptions still affect rankings in 2026?
Not directly — Google confirmed years ago that meta descriptions aren't a ranking factor. They do influence click-through rate when Google displays them, which happens roughly a third to two-thirds of the time depending on the query. Write 140 to 155 characters that include the query and a reason to click. Treat it as ad copy, not a keyword slot.
Should every page target a single keyword?
One primary keyword per page, plus its close variants and the questions surrounding it. Modern pages routinely rank for hundreds of related queries from a single topic cluster. What you must avoid is two separate URLs targeting the same intent — that splits signals and creates cannibalisation. Merge them, redirect the weaker one, consolidate the internal links.
How long should an SEO-optimised article be?
As long as the query needs. A definition page might satisfy intent in 400 words; a comparison guide may need 3,000. Check the top five results, then match their coverage of subtopics rather than their word count. Padding to hit an arbitrary number is the fastest way to raise your bounce rate and lose the snippet you were chasing.