On this page
Choose one URL, write down the query and task it should serve, open it on a phone, and search the query from a Malaysian location for a dated results screenshot. Crawling, indexing and canonicals belong to the technical SEO audit checklist, while speed belongs to the Core Web Vitals guide; the search visibility overview explains why they come first. Keyword density is not scored, for the reason below, and conversion fit is checked without any claim about conversion rates.
What task does the page complete?
The task is set by what searchers ask for, not by a keyword list. Search the query and note which result type Google shows most — guides, service pages, listings, a map — because the dominant type is evidence of what readers expect, rather than a rule that every competing page must follow. Google builds the title link it shows from the title element, the visible title, headings and other page text, and rewrites titles that are half-empty, outdated or inaccurate.
0 of 4 checked · saved on this device
One stated task, in one sentence, that no other page on the site claims.
Evidence: The task sentence in your page plan; a site search that returns one obvious page.
The page type matches the result type Google shows most for the query.
Evidence: A dated results screenshot from a Malaysian location, dominant type noted.
The title element is unique, describes the page plainly and names the task, brand kept short.
Evidence: The title element in the source against the title link in your screenshot; a Google rewrite is a partial.
The H1 agrees with the title element, and title, headings and body share the audience's language — Bahasa Malaysia, English, or each on its own page.
Evidence: The rendered H1 and lang attribute; the queries the page receives in Search Console.
Does the opening answer the task?
Google's helpful-content questions ask whether a reader leaves knowing enough to reach their goal or has to search again, and whether the title or main heading summarises the page without exaggeration; the first screen should make the page's purpose clear and help the reader judge whether to continue.
0 of 4 checked · saved on this device
The first paragraph gives the answer, or the decision to be made, before any company history.
Evidence: A screenshot of the first screen on a 390 px phone; the answer shows without scrolling.
The headings read as an outline of the task on their own.
Evidence: The heading list from the browser's accessibility tree; a heading over one thin sentence is a partial.
The meta description is unique, summarises the page in a sentence or two, and is not a keyword list.
Evidence: The meta description in the source and the snippet in your screenshot.
Title, H1 and opening promise nothing the page does not deliver.
Evidence: The three read together, then the page read to the end.
Which entities and attributes are missing?
An entity is the thing the page is about — a service, a product, a place, the business — and attributes are the facts a reader compares before deciding. A page missing them sends the reader elsewhere to finish, which Google's question about substantial, complete coverage is meant to catch. Build the attribute list first:
- The main entity in one noun phrase, named the same way in title, H1, opening and body.
- The attributes a buyer compares: for a service, inclusions, who does it, duration, price range in RM, areas covered, licences; for a product, specifications, variants, availability, warranty, delivery. Add the related entities that complete the page: locations served, alternatives, standards, the people doing the work.
- Follow-up questions after the main answer, answered here or linked to the page that answers them.
0 of 4 checked · saved on this device
The entity is named consistently, and business facts agree with the contact page and Google Business Profile.
Evidence: The three side by side; any mismatch in name, address, phone or area is a fail.
Every compared attribute is present and specific; prices in RM, lead times and coverage areas are dated.
Evidence: The attribute list ticked against the page; gaps logged as facts to add, not phrases to insert.
Images show the actual entity — premises, team, product, work — beside the text they illustrate.
Evidence: Each image's position and origin; generic stock beside a claim about your own work is a partial.
Meaningful images have alt text describing them and their relation to the content, plus descriptive file names; decorative images have empty alt.
Evidence: The img elements in the source or an alt text report; keyword-stuffed alt is a fail.
Are claims supported and links useful?
Google's helpful-content questions ask whether a page shows clear sourcing, evidence of the expertise involved and background about the author or site. Its link guidance asks for descriptive, reasonably concise anchor text, links that cross-reference your own content, and external links to sources, which it says can help establish trustworthiness; the internal linking strategy guide covers where inlinks should come from. Label every claim first:
- Measured — a number you produced and can reproduce, with how it was measured.
- Cited — a fact from a named source, linked to it.
- First-hand — an observation from doing the work, attributed to who did it.
- Opinion — a recommendation, written as one.
- Unsupported — none of the above: supply evidence or cut the claim.
0 of 4 checked · saved on this device
Every factual claim carries a label, guarantees and superlatives ("number one in Malaysia") are gone or sourced, and where readers expect a byline the page names who wrote or reviewed it.
Evidence: A marked-up copy with a label beside each claim; a byline linking to an author or About page where expected.
Every link is a real anchor with an href and descriptive text; paid or user-submitted links carry rel=sponsored or rel=ugc.
Evidence: Links exported with anchor text and rel attributes; "click here" anchors are a partial.
The page links to the steps before and after it in the reader's journey, and at least one page links to it with a descriptive anchor.
Evidence: Inlinks from a crawl or the Search Console links report.
The next step fits the task and stage — a related guide on a guide, a quote or call on a service page — and works on a phone.
Evidence: The call to action tapped on a phone; a broken form or WhatsApp button is a fail.
The task-completion rubric
| Criterion | Pass | Partial | Fail | Evidence |
|---|---|---|---|---|
| Purpose | One task, answered up front | Task shared with a sibling | No task, or several | Page plan; site search |
| Intent match | Dominant result type | Minority result type | Unlike anything shown | Dated results screenshot |
| Title and headings | Unique title; headings outline the task | Accurate title; thin headings | Boilerplate or stuffed title | Source; title link |
| Entity coverage | Every compared attribute, specific | Vague or undated | Attributes missing | Attribute list ticked |
| Evidence | Every claim labelled | Minor claims unsupported | Guarantees without sources | Marked-up copy |
| Media | Real images beside their text, good alt | Alt missing or generic | Stock filler or stuffed alt | Alt report; image origins |
| Internal links | Descriptive anchors, journey links, inlinks | Generic anchors | No contextual links or inlinks | Crawl or links report |
| Conversion fit | Fits, works on a phone | Generic next step | None, or broken | Tap test on a phone |
How to use the result
Mark every item pass, partial, unknown or not applicable, then read the rubric row by row rather than adding marks up: a fail on purpose outweighs eight passes elsewhere, so work fails in rubric order and record the date, reviewer and evidence location so the check can be repeated.
- Pass — the named evidence exists and you have looked at it. Record where it is.
- Partial — met in part or without evidence. Note what would complete it.
- Unknown — not yet measured: not a fail, but the next measurement to make.
- Not applicable — the item does not fit this page type; record why.
What to do next
- Pick the page that matters most this month, write its task sentence and run the checklist, marking unknown honestly.
- Turn each unknown into a measurement: the dated results screenshot, the link export, the alt text report.
- Fix in rubric order — purpose, intent, opening — before titles, images or links, and date each change.
- Where entities or evidence fail, get the facts from the people who hold them; the missing content is usually in the business.
Where to go next
- Technical SEO audit checklist →
Crawling, indexing and canonical controls are checked there, not on this page.
- Keyword research guide →
How to find the task and the intent a page should be built around before you audit it.
- Internal linking strategy guide →
Where the page sits in the site and which pages should link to it.
- Complete SEO audit guide →
How a single-page check fits into a whole-site audit.
Rather have this handled?
This is part of how we run seo programmes for Malaysian businesses. Ask us about your site — no pitch, a straight answer.
Sources and further reading
- Google Search Central — Search Engine Optimization (SEO) Starter Guide — read 15 September 2026
- Google Search Central — Creating helpful, reliable, people-first content — read 15 September 2026
- Google Search Central — Influencing your title links in search results — read 15 September 2026
- Google Search Central — Control your snippets in search results — read 15 September 2026
- Google Search Central — Google Images SEO best practices — read 15 September 2026
- Google Search Central — Link best practices for Google — read 15 September 2026