Skip to content
SEOWofSEOWof
Book an inspection
Browse the standard

Standard SW-1:2026.2 · Category 4 of 9

On-Page & Content Fitness

On-page faults are cheap to fix and consistently under-actioned, which makes this the highest return-per-hour category in the standard. We inspect the elements that determine how a page is understood and represented in results, and the internal competition that quietly suppresses them.

Weight
14 / 100
Checks
7
Pass-blocking
1
WOF analogue
Structural condition
OP-01Major failweight 3measured

Title tags are unique, present and descriptive

Does every indexable page have a unique, well-formed title?

Pass, advisory and fail criteria for OP-01
Pass100% of indexable pages have a unique title between 15 and 60 characters with the distinguishing term in the first half.
AdvisoryAll unique and present but some exceed 60 characters or lead with brand boilerplate.
FailAny missing or duplicated titles across indexable pages, or a template emitting identical titles at scale.

Why it matters

The title is the strongest on-page relevance signal and the primary thing a searcher reads before deciding to click. Duplicate titles across a template make pages indistinguishable to both the search engine and the searcher, and are a frequent trigger for the search engine rewriting your title into something you did not choose. Fixing titles is close to the cheapest work in SEO with a measurable effect on click-through.

How we inspect it

Titles are extracted post-render, then checked for presence, uniqueness across the site, pixel width at common SERP rendering widths, keyword placement and boilerplate ratio — the proportion of the title consumed by a repeated brand suffix.

Source: Rendered head extraction across all indexable URLs

Remedy

  1. 1Export all titles with their URLs and sort by duplication to find the offending templates.
  2. 2Write titles that lead with the distinguishing term; keep brand as a short suffix or drop it on deep pages.
  3. 3Keep within roughly 60 characters so the title is not truncated in results.
  4. 4For templated pages, build patterns from genuinely distinguishing fields, not just an ID.
  5. 5Review titles against actual query data and click-through rate, not against a keyword-density target.

Verification

Recheck re-extracts titles and confirms 100% presence and uniqueness across indexable URLs.

Usually fixed by
Content
Effort
Medium · 2–8 hrs
Est. impact
high
Typical cost
NZ$240–960
OP-02Advisoryweight 1measured

Meta descriptions present and non-duplicated

Does every important page have a distinct meta description?

Pass, advisory and fail criteria for OP-02
PassAll priority pages carry a unique description between 70 and 160 characters.
AdvisoryPresent but duplicated on secondary templates, or outside the length range.
FailMissing or duplicated across primary commercial pages.

Why it matters

Meta descriptions are not a ranking factor and this standard scores them as advisory only — a deliberate departure from tools that mark them as errors. They do influence click-through when used, and they are rewritten by the search engine most of the time. Writing them for your top commercial pages is worthwhile; writing them for ten thousand product pages is not.

How we inspect it

Descriptions are extracted, checked for presence, duplication and pixel length at standard SERP widths.

Source: Rendered head extraction

Remedy

  1. 1Write descriptions manually for the top pages by impressions, and leave the long tail to be generated from the page.
  2. 2Lead with the specific value or differentiator, not a restatement of the title.
  3. 3Include the term the page targets so it bolds when it matches the query.
  4. 4Do not duplicate across templates — a duplicated description is worse than none.

Verification

Recheck re-extracts descriptions and confirms presence and uniqueness on priority URLs.

Usually fixed by
Content
Effort
Medium · 2–8 hrs
Est. impact
low
Typical cost
NZ$240–800
OP-03Minor failweight 2measured

Heading hierarchy is semantic and singular

Does each page have exactly one H1 and a logical heading order?

Pass, advisory and fail criteria for OP-03
PassExactly one H1 per page, no skipped levels, headings describe the section beneath them.
AdvisoryMultiple H1s from a valid sectioning pattern, or a single skipped level.
FailNo H1, or headings used purely as styling with no structural meaning.

Why it matters

Headings are how both a screen reader user and a machine reader understand the shape of a page. A correct outline lets a search engine extract sub-sections as answers and lets assistive technology navigate by structure. The most common fault is not a missing H1 but headings chosen for their font size, which produces an outline that describes the visual design rather than the content.

How we inspect it

We extract the heading outline in document order and check for a single H1, absence of level skips, and headings used for text rather than for visual styling of non-heading content.

Source: Rendered DOM heading extraction

Remedy

  1. 1Ensure one H1 per page that states what the page is about, distinct from the site name.
  2. 2Nest headings without skipping levels; style with CSS classes, never by choosing a different heading level.
  3. 3Make each heading meaningful when read out of context, as a screen reader user will.
  4. 4Do not wrap logos or decorative text in heading elements.

Verification

Recheck re-extracts the heading outline and confirms a valid hierarchy on all inspected templates.

Usually fixed by
Content
Effort
Small · 0.5–2 hrs
Est. impact
medium
Typical cost
NZ$120–400
OP-04Minor failweight 3derived

Priority pages carry substantive content

Do your commercial pages have enough unique substance to compete?

Pass, advisory and fail criteria for OP-04
PassPriority pages have a unique-content ratio above 40% and cover the sub-topics common to the ranking set.
AdvisoryContent is adequate but noticeably shallower than the ranking set on several sub-topics.
FailUnique-content ratio under 20%, or priority pages are effectively template with a product name substituted.

Why it matters

This standard deliberately does not score word count. Long thin pages are still thin. What matters is whether the page answers the question completely enough that the visitor does not return to the results — and whether it contains information that exists nowhere else, which is the only durable defence against both competitors and generated content.

How we inspect it

We isolate main content from template boilerplate, measure the unique-content ratio, and compare depth and topical coverage against the pages currently ranking for the page's target query. There is no absolute word-count threshold in this standard.

Source: Main-content extraction, boilerplate ratio and comparative depth against ranking set

Remedy

  1. 1Extract the sub-topics covered by the current ranking set and identify which your page omits entirely.
  2. 2Add information only you can provide: original data, real pricing, actual photographs, specific constraints and failure cases.
  3. 3Remove padding written to hit a word count — it dilutes the page and wastes the reader's time.
  4. 4Answer the follow-up question the page currently forces the reader to search for separately.
  5. 5Consolidate several thin pages into one strong page rather than improving each in isolation.

Verification

Recheck re-extracts main content and re-runs comparative coverage against the current ranking set.

Usually fixed by
Content
Effort
Extra large · 3+ days
Est. impact
high
Typical cost
NZ$960–4,800
OP-05Advisoryweight 1measured

URLs are clean, stable and descriptive

Are your URLs readable, lowercase and free of unnecessary parameters?

Pass, advisory and fail criteria for OP-05
PassLowercase, hyphen-separated, under 100 characters, no session or tracking parameters in canonical URLs.
AdvisoryMinor inconsistencies such as mixed separators or excessive depth on a subset of templates.
FailSession identifiers in canonical URLs, mixed-case duplicates resolving separately, or unreadable identifier-only paths.

Why it matters

URLs are a weak ranking signal and this standard scores them as advisory accordingly. They matter more for the things around ranking: they are displayed in results, they are what people copy and paste as links, and mixed-case handling can create genuine duplicate URLs on case-sensitive servers. Do not restructure working URLs for cosmetic reasons — the migration risk exceeds the benefit.

How we inspect it

URLs are checked for case consistency, separator convention, depth, length, parameter presence, session identifiers and dates in paths.

Source: URL pattern analysis across the crawl set

Remedy

  1. 1Enforce lowercase with a 301 from mixed-case variants.
  2. 2Use hyphens as word separators, never underscores or camel case.
  3. 3Strip tracking parameters from canonical URLs and keep them out of internal links.
  4. 4Avoid dates in paths for content you intend to update and keep evergreen.
  5. 5Change URLs only when there is a substantive reason, and always with 301s in place before launch.

Verification

Recheck re-analyses URL patterns and confirms case and parameter compliance.

Usually fixed by
Developer
Effort
Large · 1–3 days
Est. impact
low
Typical cost
NZ$600–2,400
OP-06Minor failweight 2derived

Page type matches the intent of its target query

Does the format of the page match what the search results reward?

Pass, advisory and fail criteria for OP-06
PassPage type matches the dominant result type for its primary query.
AdvisoryMixed-intent SERP where the page matches a secondary but material intent.
FailClear mismatch, e.g. a product page targeting a query returning only guides and comparisons.

Why it matters

Intent mismatch is the reason technically flawless pages fail to rank. If every result for a query is a comparison article, a product page will not break in no matter how well optimised it is — the search engine has already determined what type of answer the query wants. This is the highest-leverage strategic check in the standard and the one automated tools almost never perform, because it requires looking at the actual results rather than the page.

How we inspect it

For each priority query we classify the dominant result type in the live SERP — informational, commercial, transactional, local or navigational — and compare it against the classification of the page targeting it.

Source: SERP composition analysis against page type for priority queries

Remedy

  1. 1Review the live SERP for each priority query and classify the dominant format.
  2. 2Where mismatched, either create the correct page type or retarget the existing page to a query it can satisfy.
  3. 3Look at what the ranking pages actually deliver — the format is a stronger signal than any keyword metric.
  4. 4Where the SERP is mixed, cover both intents: a guide that includes the commercial comparison, or vice versa.
  5. 5Re-check quarterly; SERP intent shifts, particularly as AI Overviews absorb purely informational queries.

Verification

Recheck re-samples the SERP composition for each priority query and confirms alignment.

Usually fixed by
SEO
Effort
Large · 1–3 days
Est. impact
high
Typical cost
NZ$640–2,400
OP-07Advisoryweight 2derived

Time-sensitive content is current

Is any indexed content stale enough to mislead?

Pass, advisory and fail criteria for OP-07
PassNo priority page carries outdated pricing, superseded statistics or a stale year reference.
AdvisoryOlder content exists but is clearly dated and not commercially critical.
FailPriority pages contain incorrect pricing, discontinued products, or references to years two or more behind current.

Why it matters

Freshness is not a universal ranking factor and this standard does not treat it as one. It matters where the query has a freshness expectation and where wrong information damages trust: a page quoting last year's prices costs conversions directly, and a "2024 guide" ranking in 2026 will lose to any competitor who updated theirs. The remedy is updating substance, not changing the date in the title.

How we inspect it

We identify pages carrying explicit year references, pricing, statistics or product data, cross-reference them against impression decay curves, and flag pages whose traffic is declining while their content has not changed.

Source: Date extraction, decay analysis of impressions, and stale-reference detection

Remedy

  1. 1Inventory pages with dates, prices, statistics or version references and set a review cadence per type.
  2. 2Update the substance genuinely — changing only the year in the title is detectable and ineffective.
  3. 3Remove or redirect pages about discontinued products or superseded practices.
  4. 4Show a meaningful last-updated date where it is genuinely accurate.
  5. 5Prioritise by declining impressions, not by publication date.

Verification

Recheck re-scans for stale references and re-evaluates impression trend on updated pages.

Usually fixed by
Content
Effort
Large · 1–3 days
Est. impact
medium
Typical cost
NZ$640–2,400