Technical SEO checklist for marketing sites in Singapore

Search Console flags 2,400 “Crawled - currently not indexed” URLs on a mid-market B2B site hosted out of Singapore, while the paid team is still quoting SGD 4-plus CPCs on the same commercial themes. Organic is not “slow”. It is blocked. Marketing leaders here feel it first as thin pipeline from non-brand queries, not as a pretty Lighthouse screenshot.

Another tactics list will not fix that. A scored technical audit will. You mark each control pass, weak, or fail against what Google and your own analytics can prove this week, then you order repairs so crawl, index, and page experience stop leaking demand before the next content sprint lands.

This checklist is built for marketing sites selling into Singapore: multi-template WordPress or headless builds, regional pages for Central, East, West, North and North-East, WhatsApp and form CTAs, and the usual mix of blog posts, service hubs, and gated assets. You will score the site, prioritise the fails that cap qualified traffic, and leave with a thirty-day repair order - not a new strategy deck.

Score every crawl, index, and speed control your SG marketing site ships

Work through the scorecard in one focused pass. Open Search Console, your CMS, and a live URL inspector. For each item, write pass, weak, or fail in a shared sheet with the evidence URL or report name. Do not score “vibes”. Score what you can inspect before the next leadership meeting.

Audit scorecard

• 1 Index coverage vs. what you meant to publish In Search Console, compare “Submitted and indexed” against “Discovered - currently not indexed” and “Crawled - currently not indexed”. Good looks like: primary service and hub URLs indexed within days of publish, thin tag/archive noise excluded, and no surprise parameter sprawl. Fail costs you invisible commercial pages while thin filters still burn crawl budget. Note the exact count of not-indexed URLs on money templates (services, solutions, locations) separately from blog archives.

In Search Console, compare “Submitted and indexed” against “Discovered - currently not indexed” and “Crawled - currently not indexed”. Good looks like: primary service and hub URLs indexed within days of publish, thin tag/archive noise excluded, and no surprise parameter sprawl. Fail costs you invisible commercial pages while thin filters still burn crawl budget. Note the exact count of not-indexed URLs on money templates (services, solutions, locations) separately from blog archives.

• 2 robots.txt and sitemap honesty Fetch robots.txt live. Confirm you are not blocking /wp-admin only while accidentally disallowing critical CSS/JS paths or entire /sg/ or /en-sg/ trees. Sitemap must list only canonical, 200, indexable URLs - no redirects, no noindex, no staging hosts. Good: one XML sitemap index, lastmod that moves when content changes, submitted in Search Console. Fail: dual sitemaps from old migrations, or a sitemap full of 301s that trains Google on the wrong URL set. Run a robots.txt checker and a sitemap checker against production, not staging.

Fetch robots.txt live. Confirm you are not blocking /wp-admin only while accidentally disallowing critical CSS/JS paths or entire /sg/ or /en-sg/ trees. Sitemap must list only canonical, 200, indexable URLs - no redirects, no noindex, no staging hosts. Good: one XML sitemap index, lastmod that moves when content changes, submitted in Search Console. Fail: dual sitemaps from old migrations, or a sitemap full of 301s that trains Google on the wrong URL set. Run a robots.txt checker and a sitemap checker against production, not staging.

• 3 Canonical and hreflang discipline Spot-check ten commercial URLs: self-referencing canonicals, HTTPS, preferred host (www or bare), and no “print” or UTM variants winning the canonical. If you run APAC variants, hreflang must map SG English to the SG URL, not to a generic .com that 302s. Fail looks like: parameter URLs self-canonicalising, or SG pages pointing canonical to a US template. Use a canonical checker on every template type (home, service, blog, landing).

Spot-check ten commercial URLs: self-referencing canonicals, HTTPS, preferred host (www or bare), and no “print” or UTM variants winning the canonical. If you run APAC variants, hreflang must map SG English to the SG URL, not to a generic .com that 302s. Fail looks like: parameter URLs self-canonicalising, or SG pages pointing canonical to a US template. Use a canonical checker on every template type (home, service, blog, landing).

• 4 HTTPS, mixed content, and host consistency Force HTTPS sitewide. Kill mixed-content console errors on key conversion paths (demo request, contact, pricing). One host only - permanent redirects from the other. WordPress sites on Singapore-hosted shared stacks (e.g. Vodien, Exabytes) commonly serve Google Fonts or WooCommerce product images over HTTP when the CDN or object storage bucket was provisioned before the SSL migration; that tanks trust signals and breaks modern browser features. Good: clean padlock, HSTS if your infra team supports it, zero mixed scripts on the primary CTA page.

Force HTTPS sitewide. Kill mixed-content console errors on key conversion paths (demo request, contact, pricing). One host only - permanent redirects from the other. WordPress sites on Singapore-hosted shared stacks (e.g. Vodien, Exabytes) commonly serve Google Fonts or WooCommerce product images over HTTP when the CDN or object storage bucket was provisioned before the SSL migration; that tanks trust signals and breaks modern browser features. Good: clean padlock, HSTS if your infra team supports it, zero mixed scripts on the primary CTA page.

• 5 Status codes on the money paths Crawl the top 50 organic landing pages from Search Console. You want 200s. Soft 404s on “thank you” and filtered category pages are a common SG e-commerce and directory leftover on marketing sites that absorbed a shop module. Soft 404s and long redirect chains dilute equity and confuse reporting. Soft fail: single-hop 301s you planned. Hard fail: chains of three-plus hops, or 200s that render empty shells.

Crawl the top 50 organic landing pages from Search Console. You want 200s. Soft 404s on “thank you” and filtered category pages are a common SG e-commerce and directory leftover on marketing sites that absorbed a shop module. Soft 404s and long redirect chains dilute equity and confuse reporting. Soft fail: single-hop 301s you planned. Hard fail: chains of three-plus hops, or 200s that render empty shells.

• 6 Title, H1, and meta description quality on templates Templates must not ship duplicate titles like “Home” or “Services | Brand | Brand”. One clear H1 per URL, matching the commercial intent of the page, not a design slogan. Meta descriptions should be unique enough to win the SERP snippet for branded and non-branded queries. Fail: fifty service child pages with the same title pattern and empty metas. Spot-check with a title tag checker, H1 checker, and meta description checker.

Templates must not ship duplicate titles like “Home” or “Services | Brand | Brand”. One clear H1 per URL, matching the commercial intent of the page, not a design slogan. Meta descriptions should be unique enough to win the SERP snippet for branded and non-branded queries. Fail: fifty service child pages with the same title pattern and empty metas. Spot-check with a title tag checker, H1 checker, and meta description checker.

• 7 Core Web Vitals on mobile Singapore traffic Use CrUX field data in Search Console and PageSpeed Insights for your origin and key URLs. As of the March 2025 CrUX report, only about 55.9% of origins pass all three Core Web Vitals (web.dev/explore/metrics). Marketing sites fail most often on LCP (hero image or web font) and INP (heavy JS on sticky headers and chat widgets). Good: LCP under ~2.5s on mobile for hub and landing templates, stable CLS on form pages, chat loaded after interaction. Fail: third-party tags blocking main thread before the primary CTA paints. Validate with a Core Web Vitals checker.

Use CrUX field data in Search Console and PageSpeed Insights for your origin and key URLs. As of the March 2025 CrUX report, only about 55.9% of origins pass all three Core Web Vitals (web.dev/explore/metrics). Marketing sites fail most often on LCP (hero image or web font) and INP (heavy JS on sticky headers and chat widgets). Good: LCP under ~2.5s on mobile for hub and landing templates, stable CLS on form pages, chat loaded after interaction. Fail: third-party tags blocking main thread before the primary CTA paints. Validate with a Core Web Vitals checker.

• 8 Mobile rendering of forms, WhatsApp, and sticky CTAs In Singapore, enquiry often starts on mobile with a WhatsApp tap or short form. Technical SEO fails when the CTA is covered by a cookie bar, when the click-to-chat link is not crawlable, or when the form lives only inside a non-indexed modal. Good: crawlable tel and wa.me links, visible primary CTA without layout shift, form fields that do not force horizontal scroll. Fail: “Book a call” that only exists after three JS-rendered steps Google never sees as content.

In Singapore, enquiry often starts on mobile with a WhatsApp tap or short form. Technical SEO fails when the CTA is covered by a cookie bar, when the click-to-chat link is not crawlable, or when the form lives only inside a non-indexed modal. Good: crawlable tel and wa.me links, visible primary CTA without layout shift, form fields that do not force horizontal scroll. Fail: “Book a call” that only exists after three JS-rendered steps Google never sees as content.

• 9 Structured data that matches the visible offer Organisation, WebSite with SearchAction if you have site search, Article or BlogPosting on thought leadership, FAQ only where FAQs are visible, and Service or Product markup only when the page truly offers that entity. Invalid or spammy FAQ schema is a fast way to lose rich results. Validate with a schema markup validator and, if you use FAQs, a FAQ schema generator against live HTML, not the CMS preview.

Organisation, WebSite with SearchAction if you have site search, Article or BlogPosting on thought leadership, FAQ only where FAQs are visible, and Service or Product markup only when the page truly offers that entity. Invalid or spammy FAQ schema is a fast way to lose rich results. Validate with a schema markup validator and, if you use FAQs, a FAQ schema generator against live HTML, not the CMS preview.

• 10 Internal link graph to commercial hubs Orphan service pages and blog posts that never link into money hubs waste the crawl Google already granted. From each new article, one contextual link to the matching service or solution page; from hub pages, links to proof and pricing or contact. Good: shallow depth (money URLs within three clicks of home), descriptive anchors, no “click here”. Fail: 200-post blog with zero links into the three pages that should own non-brand commercial queries. An internal link analyzer makes the orphans obvious.

Orphan service pages and blog posts that never link into money hubs waste the crawl Google already granted. From each new article, one contextual link to the matching service or solution page; from hub pages, links to proof and pricing or contact. Good: shallow depth (money URLs within three clicks of home), descriptive anchors, no “click here”. Fail: 200-post blog with zero links into the three pages that should own non-brand commercial queries. An internal link analyzer makes the orphans obvious.

• 11 Pagination, facets, and parameter control If the site inherited filters (industry, location, resource type), decide which combinations are indexable. Everything else: noindex or blocked parameters in Search Console, consistent canonicals. Fail: infinite faceted URLs ranking for junk long-tails while your real “SEO agency Singapore” style hubs stay thin. Marketing sites rarely need faceted indexation; treat it as debt unless product marketing explicitly owns those landing pages.

If the site inherited filters (industry, location, resource type), decide which combinations are indexable. Everything else: noindex or blocked parameters in Search Console, consistent canonicals. Fail: infinite faceted URLs ranking for junk long-tails while your real “SEO agency Singapore” style hubs stay thin. Marketing sites rarely need faceted indexation; treat it as debt unless product marketing explicitly owns those landing pages.

• 12 JS rendering and critical content in the first HTML View source and “view rendered” for H1, primary copy, and main nav. If the commercial paragraph only appears after client-side render and Google’s fetched HTML is empty, you have a rendering debt. Soft fail: hydration delay. Hard fail: entire service body behind JS with no server-rendered fallback. This shows up as “Crawled - currently not indexed” more often than teams expect on modern marketing stacks.

View source and “view rendered” for H1, primary copy, and main nav. If the commercial paragraph only appears after client-side render and Google’s fetched HTML is empty, you have a rendering debt. Soft fail: hydration delay. Hard fail: entire service body behind JS with no server-rendered fallback. This shows up as “Crawled - currently not indexed” more often than teams expect on modern marketing stacks.

• 13 Log-level crawl waste and soft bot traps If you can access server or CDN logs, sample Googlebot hits. High volume on calendar URLs, session IDs, or old UTM copies means budget spent off-strategy. Even without logs, Search Console crawl stats plus a site crawl reveal loops. Good: bot time concentrated on hubs, services, and recent content. Fail: bot hammering /page/47 of a tag that should have been noindexed two redesigns ago.

If you can access server or CDN logs, sample Googlebot hits. High volume on calendar URLs, session IDs, or old UTM copies means budget spent off-strategy. Even without logs, Search Console crawl stats plus a site crawl reveal loops. Good: bot time concentrated on hubs, services, and recent content. Fail: bot hammering /page/47 of a tag that should have been noindexed two redesigns ago.

• 14 Tracking integrity on organic landing pages Technical SEO without conversion instrumentation is theatre. Confirm GA4 (or your analytics) fires on thank-you and WhatsApp click events, UTMs do not break canonicals, and landing pages load tags without tanking INP. When paid and organic disagree on lead counts, fix the dual-fire or missing event before you “scale content”. Pair this with a quick UTM link builder standard so campaigns stop creating duplicate URL variants.

Technical SEO without conversion instrumentation is theatre. Confirm GA4 (or your analytics) fires on thank-you and WhatsApp click events, UTMs do not break canonicals, and landing pages load tags without tanking INP. When paid and organic disagree on lead counts, fix the dual-fire or missing event before you “scale content”. Pair this with a quick UTM link builder standard so campaigns stop creating duplicate URL variants.

If several commercial templates fail items 1, 3, 6, and 7 together, pause net-new publishing volume until the template is healthy. Shipping more AI-assisted drafts onto a broken template multiplies the mess - and yes, plenty of top pages now contain AI-assisted copy, but Google still has to fetch, render, and trust the URL.

When the scorecard is ugly on crawl and index, a specialist technical SEO programme is usually cheaper than another quarter of content that never enters the index cleanly.

How to mark pass, weak, and fail so repairs beat vanity polish

Pass means you have evidence in Search Console, a crawl, or live HTML that the control works on the templates that drive pipeline. Weak means it works on homepage and brand, but breaks on service, location, or campaign landing templates. Fail means the control is wrong on a money URL or systematically wrong across a template.

Weight fails in this order: (1) anything that blocks indexation or sends the wrong canonical, (2) anything that breaks mobile conversion paths, (3) Core Web Vitals on high-impression URLs, (4) on-page uniqueness and internal links, (5) nice-to-have schema polish. Do not burn a sprint on FAQ rich results while half your service URLs are “Crawled - currently not indexed”.

Score by template, not by homepage pride. Homepage often passes because design and performance budget live there. The SGD pipeline usually sits on /services/, /solutions/, industry pages, and a handful of comparison or pricing URLs. Mark each template once, then multiply the fail across URL count so leadership sees scale - “fourteen service URLs fail canonical” lands harder than “canonicals feel off”.

Reconcile organic sessions to assisted pipeline the same week you score. If non-brand clicks rise while qualified enquiries stay flat, look at post-click friction before you declare technical SEO “done”. Session behaviour tools such as HeyLead Insights help you see whether slow LCP, buried forms, or confusing proof are eating the traffic you just unlocked.

DIY tools for this audit

• Google index checker

• Core Web Vitals checker

• Canonical checker

• Schema markup validator

• AI SEO audit

• Local SEO checklist

Thirty days of technical repairs after the score, not another content calendar

Treat the month as four repair blocks. You are clearing fails that stop organic from compounding. Content cadence still matters - especially sustained publishing with human strategy - but it sits behind a clean template.

30-day fix plan

• 1Days 1-7 - Index truth. Export Search Console coverage. List every money URL that is not indexed. Fix robots, noindex mistakes, canonical loops, and sitemap junk. Request indexing only after the live HTML is correct. Confirm host and HTTPS redirects in one hop.

• 2Days 8-14 - Template titles, H1s, and internal links. Rewrite titles and H1s on weak service templates so each URL owns one clear commercial job. Add internal links from high-traffic articles into those hubs. Remove or noindex pure thin archives. Run title/H1/meta checks again on a sample of ten URLs.

• 3Days 15-21 - Speed and interaction on conversion URLs. Compress and properly size hero assets, defer non-critical JS, load chat after intent, fix CLS on forms. Re-test mobile LCP/INP on the five URLs with the most organic impressions. Do not “optimise the whole site” - win the pages that already get demand.

• 4Days 22-30 - Schema, rendering, and measurement lock. Validate Organisation and page-level schema. Fix SSR/hydration gaps on service bodies. Confirm GA4 events and WhatsApp clicks. Document what still fails and why (vendor tag, design system, legal cookie banner). Only then reopen the content throttle.

Once templates pass, sustainable publishing becomes leverage instead of noise. That is where a managed cadence product like HeyLead Auto Blogger fits: region- and niche-aware articles under human strategy, shipped without turning your marketing lead into a full-time production desk - on top of a site Google can actually crawl and rank.

If paid search is still carrying the number while organic rebuilds, keep SEM budgets honest: wasted query spend and misaligned landings are a separate leak. Technical SEO will not fix a homepage-only handoff from ads, but it will stop organic from repeating the same mistake.

What marketing leaders are seeing

On one B2B SaaS client we audited in Q1, three service hubs were stuck in ‘Crawled - currently not indexed’ for eleven weeks while paid covered the gap at roughly SGD 5 CPC. Fixing the canonical on the template - not writing more blogs - is what finally moved non-brand enquiries.

On one professional services client we audited, the chat widget loaded before the hero on mobile. INP looked fine on desktop lab tests and awful in the field. Once we deferred the widget, the same landing page started converting WhatsApp taps again without touching the offer.

FAQs

How often should a Singapore marketing site run this technical SEO checklist?

Full scorecard quarterly, or after any redesign, CMS migration, tag-manager overhaul, or major template change. Spot-check index coverage and Core Web Vitals monthly on the top twenty organic URLs. Waiting for a traffic cliff is how you discover a robots mistake three months late.

Is technical SEO enough if competitors publish more content?

No. Technical health is the floor. You still need intent-led pages, proof, and a publishing system. But more content on a failing template rarely beats fewer clean URLs that index, load, and convert. Fix the floor first, then out-publish on substance.

What should we fix first if budget only covers one sprint?

Indexation and canonical correctness on commercial templates, then mobile LCP/INP on those same URLs, then internal links from existing traffic. Schema and micro-copy polish come after Google can reliably fetch and rank the pages that sell.

Do Core Web Vitals still matter when AI Overviews reduce clicks?

Yes. Zero-click behaviour raises the bar on the clicks you do win. Slow or unstable pages waste scarce non-brand visits and poison engagement signals. Treat CWV as conversion infrastructure, not a vanity badge - especially while only a bit over half of origins pass all three vitals in field data.

Can we score this without an enterprise crawler?

Yes for a first pass: Search Console, URL Inspection, free validators, and a limited crawl of money templates. Enterprise crawlers help at scale, but most SG mid-market leaks show up in coverage reports and ten carefully chosen live checks.

Putting it to work

This week

• Export Search Console coverage and mark every commercial template pass / weak / fail for index status.

• Run robots, sitemap, and canonical checks on production home, one service hub, one blog post, one campaign landing page.

• Record mobile field CWV for your five highest-impression organic URLs.

• List orphan or thin URLs that should be noindexed or redirected rather than rewritten.

• Confirm form submit and WhatsApp click events fire once on thank-you paths.

• Share the scored sheet with whoever owns CMS and tags so repairs have a named owner.

Next 30 days

• Clear index and canonical fails on all money templates.

• Ship title/H1 uniqueness and internal links into hubs.

• Win LCP/INP on the conversion set; defer non-essential third parties.

• Validate schema and lock analytics.

• Only then scale publishing - with a system, not heroics.

Pull your last 28 days of Search Console coverage for service and solutions URLs, score the fourteen controls above in a single sheet, and assign owners to every hard fail before you approve the next content batch. When the leak is crawl, index, and template performance on a Singapore marketing site - the unglamorous work that decides whether organic can carry pipeline while paid CPCs stay elevated - a partner like HeyLead can run the technical programme end to end so your team stays on positioning, offers, and the number. Start the conversation at [email protected].