Short answer

An on-page SEO service fixes what sits on the page itself — search intent match, title and headings, structure, internal links and content depth — so the page earns its ranking before you spend on backlinks. SerpInsight audits the page, rewrites what’s weak, and says plainly if links aren’t the problem.

Most stalled rankings get blamed on authority. Some deserve it. But plenty of pages are stuck for a cheaper reason: they answer a slightly different question than the one the results page rewards. That’s on-page work — changeable this afternoon, without waiting on a publisher, and usually worth fixing before you buy a single link insertion (also called a niche edit).

What on-page work actually changes

On-page optimization moves the levers you own outright. No outreach, no waiting on an editor.

On-page SEO, defined

Every ranking input you can change without anyone’s permission: the copy, title tag and meta description, heading hierarchy, URL, image alt text, structured data, and the internal links pointing at the page from the rest of your own site. Off-page needs a third party to say yes.

A content optimization service engagement works through five things, in this order:

  • Intent match: does the page’s format match what the current results reward?
  • Title tag, meta description, and the first sixty words of copy
  • Heading hierarchy and the order sections appear in
  • Internal links in and out, with anchors that describe the destination
  • Coverage gaps: subquestions every ranking competitor answers and you don’t

Keyword density isn’t on that list. It was a reasonable proxy twenty years ago; chasing it now produces copy written for a parser.

Why can’t a good page rank? Usually it answers a different question

The most common reason a page can’t rank isn’t a link shortage — it’s an intent mismatch. Google has already worked out, from how people behave on that query, what kind of page satisfies it. If the top ten are comparison round-ups and you’ve published a product page, more backlinks drag you from position 34 to 22 and then stop.

A hypothetical example. You sell scheduling software and target “best scheduling software for clinics”. If eight of the top ten are round-ups naming six to ten tools, the reward goes to a comparison, not a pitch. Publish a real comparison that names rivals and admits where they win, or move the page to a query where a product page fits, like “clinic scheduling software pricing”.

Reading intent properly takes about twenty minutes per keyword. We do it first, and sometimes the honest finding is that the page is fine and the target keyword was wrong.

Structure, headings and internal linking

Answer in the first sixty words. If someone asks how long a rebate takes, the sentence after the heading says “usually six to eight weeks”, not “rebates are an important part of the customer journey”. Then one idea per section, under an H2 a person would plausibly type. “Processing times” is a filing label. “How long does a rebate take?” is a real question.

Internal links are the most underused lever on most sites we see. A page with nothing pointing at it is a page you’ve quietly told Google not to prioritise. Links should come from topically related pages, sit inside sentences rather than a related-posts strip, and use anchors that describe where they go. For the mechanics, see our guide to wiring your own pages together.

You can be freer with descriptive anchors on your own site than off it. Just don’t point forty internal links at one URL using an identical phrase. The ratios we hold to on external placements are stricter — roughly 40-50% branded — because external anchors are what a manual reviewer would look at.

Depth beats length, and “make it 2,000 words” is a bad target

One opinion, and it’s the only one here: word count targets are the worst proxy in SEO. They’re popular because they’re easy to check, which is why they get gamed. Tell a writer to hit 2,000 words on a topic that takes 700 and you’ll get 1,300 words of throat-clearing.

Depth means covering the questions a reader has next, concretely. Our test: can a knowledgeable reader delete any paragraph without losing information? If yes, it’s padding, and padding dilutes the page. Length from coverage is fine. Length from a target is a tax on your reader.

Which symptoms are on-page problems and which are link problems

Before spending on outreach, match the symptom to its likely cause.

Symptom Most likely cause On-page or links?
Ranks 30-60, content looks reasonable Format doesn’t match what the results reward On-page. Links here are money burnt.
Ranks 6-30, content good, fewer referring domains than rivals Authority gap Links. This is where placements pay.
Plenty of impressions, very low click-through Title and meta don’t match the query On-page. Often an hour’s work.
The wrong URL on your site ranks for the keyword Cannibalisation, confused internal signals On-page. Consolidate, redirect, re-link.
Page isn’t indexed at all Crawl, duplication or quality threshold On-page and technical, never links.
Held page two, then slipped after a rival updated Coverage gap opened up On-page refresh first, links second.
New domain, thin site, under roughly DR 20 Nothing to rank yet Neither alone. Build pages first.

How an on-page engagement runs

  1. Scope and query confirmation

    You send the URL and the keyword it’s meant to win. We check the keyword is worth winning and that this URL is the right one for it. If it isn’t, we say so in writing before work starts.

  2. Results page read

    What’s ranking now: format, angle, depth, who those pages are written for. This is where intent mismatch surfaces, and most audits skim it.

  3. Gap and structure audit

    Subquestions the ranking pages answer and yours doesn’t, plus heading order, title, meta and internal links. Delivered as a marked-up document, not 60 pages of automated warnings.

  4. Rewrite and implementation

    Revised copy, headings, title and an internal link plan. You implement it, or we do.

  5. Measurement window

    Rankings typically move within four to eight weeks on indexed pages, faster for title changes. If nothing moves, links become the honest next step.

When on-page is the cheaper win than link building

Often, and we’d rather say it than sell around it. Link building suits one situation: the page deserves to rank, matches intent, and is stuck in positions six to thirty while competitors hold more referring domains. Outside that, links are an expensive way to move a page that would move anyway once the copy was fixed.

A rewrite is a fixed cost paid once. A Growth-tier placement is $147 per link and a page short on authority needs several — our pricing lists tiers and bulk discounts. On-page first also makes every later link work harder: authority pointed at a page that satisfies the query converts, authority pointed at the wrong answer evaporates. Our take on link pacing covers sequencing.

Where we draw the line. If a page can’t rank because of intent mismatch, we’ll tell you links won’t fix it, even when you came to us specifically to buy links.

On-page SEO service: common questions

How long does on-page SEO take to show in rankings?

On an already-indexed page, expect movement in four to eight weeks. Title and meta changes can register within days. If nothing has shifted after roughly ten weeks, the constraint is usually authority rather than the page.

Can on-page optimization alone get a page to rank?

In low and medium competition queries, frequently yes — especially long-tail ones where the ranking pages are mediocre. In commercial queries where every result holds hundreds of referring domains, no. The test is whether the pages above you are better, or merely better linked.

What word count should my page be?

There isn’t one, and any agency quoting a number is selling by the yard. Cover the question and its follow-ups, then stop. A 900-word page with specifics routinely beats a 2,400-word page that pads. For a sanity check, look at the shortest page in the top five.

Do you make the changes or just send recommendations?

Either. With CMS access we implement the rewrite, headings, title and internal links directly. Without it, you get revised copy plus a change list specific enough for a developer to action in an afternoon.

How much does an on-page SEO service cost?

It’s scoped per page, because a title fix and a full rewrite aren’t the same job. We quote a fixed price in writing before anything starts — no retainer, no auto-renewal. Our link placements, for comparison, are fixed at $87, $147 and $247 by host site tier.

Send us the URL and the keyword. We reply within 24 hours with an honest read on whether it’s an on-page problem or an authority problem — and we’ll say so if it’s the second. Get in touch, or see the rest of what SerpInsight offers.

How do we decide which pages to optimize first?

We rank candidates by the gap between what a page already earns and what it plausibly could, then divide that by the hours it takes to close. Pages already sitting in positions six to thirty usually win that maths, because the impressions exist and the fix tends to be structural rather than a rewrite from nothing.

Every site we look at has more optimization candidates than budget. So the first deliverable of an on-page SEO service isn’t a rewrite — it’s a queue, with the reasoning shown. Below is roughly how we sort the situations we see most, and what each is worth chasing.

Page situation Potential upside Effort First move
Positions 6-30, content broadly sound High — impressions already exist, clicks scale steeply between position 10 and position 4 Medium: title, intro, coverage gaps, internal links Fix coverage, then decide on links
High impressions, click-through well under the pages around it Medium-high — traffic arrives without any ranking change at all Low: an hour or two per page Rewrite title and meta against the actual query
Two or more of your URLs swapping for one keyword High — you’re competing with yourself and Google keeps picking differently Medium: merge, redirect, re-point internal links Pick the survivor, consolidate into it
Real traffic, no enquiries or sales High in revenue terms, invisible in ranking reports Low-medium: intent, offer placement, page structure Check the query intent matches your offer
Ranked well, slipped after a competitor refreshed Medium — you had the position once, so the page is close Low-medium: add what the new winner covers Compare section by section, close the gap
No internal links pointing at it Medium — cheap to fix, and it compounds across the site Low: an afternoon for a whole site Link from related pages with descriptive anchors
Positions 30-60, thin or off-intent Low per hour spent — the page needs rebuilding, not tuning High: it’s a new page in practice Park it until the quick wins are done

The last row is the one clients push back on, because a page at position 45 feels closer to page one than it is. Positions between 30 and 60 usually mean Google understands the page and has decided it isn’t a strong answer. Getting from 45 to 15 is a rebuild. Getting from 12 to 6 is an edit.

Title tags and meta descriptions move clicks, not rankings

Treat the title tag as a ranking factor and you’ll write it for a crawler. Treat both as advertising copy that has to win a choice against nine competitors, and the numbers move. Google rewrites titles it doesn’t like anyway, which tells you what it thinks the field is for.

The rewrite process is short. Pull the queries a page already gets impressions for, find the phrasing real people use, and make sure that phrasing appears in the first few words of the title where it isn’t truncated on a phone. Then write a meta description that gives a reason to click rather than a summary of the page. “Complete guide to commercial lease disputes” summarises. “How long a lease dispute takes, what it costs, and when to settle” promises three answers.

  • Front-load the phrase people actually type; brand name goes last or not at all
  • Keep titles near 55-60 characters so mobile doesn’t cut the useful half
  • Add the qualifier that separates you: a price, a year, a location, a limit
  • Never reuse one title pattern across forty pages; near-duplicates get rewritten by Google
  • Change one page’s title at a time, then wait a fortnight before judging it

Two honest caveats. Google may ignore your meta description entirely and pull a snippet from the body instead, which is one more reason the first sixty words of copy matter. And click-through varies enormously by position, so a rise after a title change only means something if the ranking held steady. We check both before claiming a title rewrite worked.

Using People Also Ask as a structure guide, not a keyword list

The related questions box and the “related searches” strip at the bottom of a results page are the closest thing to a public map of what a searcher wants next. Read them as an outline for your headings rather than as phrases to sprinkle into paragraphs.

Here’s the method. Search your target keyword, expand every question in the box, note the ones that keep appearing as you expand others, and check the related searches underneath. You’ll usually end up with eight to fifteen sub-questions. Cluster them: some are the same question worded differently, some belong on a different page entirely, and three to six are genuinely the sections your page is missing. Those become H2s, phrased the way a person would ask, each answered in the two sentences directly beneath.

What this is not is a keyword shopping list. Stuffing all fifteen phrases into one page produces a document that repeats itself and answers nothing well. The value is the ordering — you learn which question comes first in a reader’s head and which is a follow-up, and a page arranged in that order reads as though someone thought about it.

The section test

Every H2 should be a question a reader would recognise as theirs, and the sentence under it should be a usable answer even if they read nothing else. If a section needs three paragraphs of setup before it says anything, the heading is wrong or the section belongs somewhere else.

Schema markup: what’s worth adding and what isn’t

Schema tells search engines what the things on your page are — this is a product, that is its price, this block is an author. It does not directly raise rankings, and anyone who tells you otherwise is selling markup by the page. What it can do is change how your result looks, and a result with stars, a price or a breadcrumb trail attracts more clicks at the same position.

The types worth the effort for most sites are short:

  1. Breadcrumb

    Replaces the raw URL in the result with a readable path. Low effort, applies site-wide, and it also nudges you into fixing a messy site structure. Start here.

  2. Article and Author

    Marks up publication date, updates and who wrote the thing. Useful on anything time-sensitive, and it gives your author pages a job beyond decoration.

  3. Product and Offer

    Price, stock status and review ratings for ecommerce. This is where markup most reliably changes what the listing looks like, so it’s the one worth doing carefully.

  4. LocalBusiness

    Address, hours and phone, matched exactly to your Google Business Profile. Relevant if you have premises — our local SEO service covers how that fits the rest of the local picture.

  5. FAQ

    Still worth adding for machine readability, but be realistic: Google narrowed FAQ rich results sharply in 2023, so on most sites it no longer shows in the listing. Add it because it’s accurate, not because you expect stars.

Two rules we hold to. Markup must describe what a visitor can actually see on the page — marking up reviews or prices that aren’t visible is a structured data violation and gets rich results removed. And validate before you ship, because broken markup is silently ignored, which means you’ll never know it isn’t working.

How does on-page work interact with link building?

Links amplify a page that already deserves to rank. They can’t rescue one that doesn’t. That’s the whole relationship, and getting the order wrong is the most expensive mistake in SEO budgets we see.

Think of it as multiplication rather than addition. A page that satisfies the query has a real coefficient, so authority pointed at it produces movement. A page that misreads intent has a coefficient near zero, and you can multiply zero by as many referring domains as you like. This is why two sites can buy identical placements and get different outcomes — the difference wasn’t the links.

The practical sequence we recommend: fix intent and structure, wait four to eight weeks, look at where the page settles. If it lands in positions six to thirty and the pages above hold more referring domains than you, that’s the moment a link insertion (also called a niche edit) earns its cost, because every point of authority now lands on a page that converts the visit. If it lands at 40 and stays, spend the money on the page instead.

There’s a reverse case too, and it’s less discussed. On-page changes to a page that already holds strong links carry more weight than the same changes to an unlinked page, because Google recrawls and re-evaluates it faster. If you have one well-linked page that ranks for a near-miss keyword, optimizing that page is often quicker than building a new one. Our notes on internal linking cover how to push some of that existing authority onto neighbouring pages, and the SEO strategy guide sets out the sequencing across a whole site.

Sequencing rule. We won’t quote link placements for a page we’ve just told you has an intent problem. Fix the page, re-measure, then buy links if the numbers still say so — tiers are $87, $147 and $247, and none of them are worth spending twice.

What we change, and what we never touch without asking

Handing your pages to an outside team is a reasonable thing to be nervous about. Plenty of optimization work reads as though it were written by someone who has never met the business, because it was. So we’re explicit about the boundary before anything is edited.

Changed as part of normal work, with the diff shown to you either way: heading text and order, the opening sixty words, section content where coverage gaps exist, title tags and meta descriptions, image alt text, internal link anchors and destinations, and schema markup.

Never changed without written approval:

  • Your brand voice. If you write plainly and we’d write more formally, you win.
  • Deleting or merging a page. Consolidation is often right, and it’s always your call.
  • URLs and redirects. Cheap to propose, expensive to get wrong.
  • Claims about your product, pricing, guarantees or regulated wording.
  • Calls to action, forms and anything downstream of a click your sales team owns.
  • Removing content that no longer ranks but still serves customers or support.

Where we work directly in your CMS, changes go out one batch at a time with a dated note of what moved, so a ranking shift can be traced to a cause. Where you implement, you get the revised copy plus a change list a developer can work through without asking us a single follow-up question.

More on-page SEO service questions

Should I optimize an old page or write a new one?

If the old URL ranks anywhere in the top hundred for a related term, optimize it — it already has history and usually a few internal links. Write new when the existing page targets a genuinely different intent, since forcing two jobs onto one URL is how cannibalisation starts. A rough line: if the old page needs more than about 60% of its body replaced, it’s a new page wearing an old URL.

Does adding schema markup improve rankings?

Not directly, and we’d rather say that than sell markup as a ranking service. Schema affects how your listing can appear — breadcrumbs, prices, ratings — and a better-looking listing earns more clicks at the same position. One caveat worth knowing: markup describing things a visitor can’t see on the page breaks Google’s structured data policy and can get rich results pulled entirely.

How many pages can you optimize at once?

We scope in batches, and we’d usually rather start with five to ten pages than thirty, because you learn what your site responds to before committing the budget. Each page takes its own results-page read, so a batch of ten is real work rather than a template run across a spreadsheet. Send the URLs and the keywords and we’ll come back within 24 hours with a queue and a fixed price.

Not sure whether your problem is the page or the authority behind it? Send one URL and the keyword it should win. We’ll tell you which it is, in writing, before quoting anything — start here, or read how we think about pacing if you already know links are the gap.