What CRO is and why it belongs here
Conversion Rate Optimisation (CRO) is the practice of increasing the percentage of users who complete a desired action on the site. Form submissions, phone clicks, purchases, bookings. The work blends analytics, user research, copywriting, design and statistics, depending on how formal the programme is. At its best, CRO compounds with SEO. At its worst, it operates as a separate budget and the two disciplines fight each other.
It's like bringing a knife to a gunfight when you run SEO without CRO. Good SEO brings traffic to a page that does not convert; the spend produces sessions and impressions but not revenue. The fix is not more traffic. The fix is fixing the page. The CRO discipline is what does that fixing.
The reason CRO sits inside the SEO measurement pillar rather than being treated as a separate discipline: it shares almost all the same measurement infrastructure. GA4 conversion events. Event tracking. Landing-page-level analysis. Attribution. The KPI hierarchy that the rest of this pillar covers. Keeping CRO inside the SEO measurement discipline avoids the trap of doubling organic traffic while leaving the conversion rate unchanged.
Picking the right pages
Not every page is worth testing. The high-impact CRO work concentrates on three categories of page.
Category 1: High-traffic high-intent pages
The service-area landing pages, money pages, product pages and category pages that drive most of the commercial organic traffic. These pages have the highest baseline conversion rate, the highest absolute lead or revenue volume, and the most upside from a percentage-point improvement. A 10 percent lift on a page producing 50 leads a month is 5 extra leads; the same lift on a 5-lead page is half a lead. Start where the traffic is.
Pull the GA4 landing-page report filtered to Organic Search. Sort by sessions descending. Cross-reference with the GSC Performance Pages tab to confirm the pages with the highest organic clicks. Pick the top 5 to 10 by sessions that also have meaningful commercial intent.
Category 2: High-intent low-conversion-rate pages
Pages where the conversion rate looks low for the intent. A "plumber Fremantle" landing page converting at 0.8 percent when the comparable pages on the site convert at 3 percent is a CRO problem, not a traffic problem. The fix is on-page: form design, CTA placement, trust signals, page speed, copy clarity. See content depth and E-E-A-T explained.
Category 3: Strategically important pages with growth runway
Pages that do not yet have huge traffic but will. A new service-area launch, a category page for a new product line, the comparison pages a content publisher is using to drive monetisation. CRO on these pages early avoids baking in low conversion rates that need rebuilding later.
Pages not worth CRO testing
Blog posts with informational intent rarely benefit from formal CRO. The intent is wrong for conversion; the reader is researching, not buying. The right move on informational pages is to optimise for the next step (newsletter signup, related-service exposure, calls to high-converting commercial pages) rather than for the primary conversion event. About pages, contact pages and other utility pages also rarely benefit from full A/B testing.
Diagnosing the conversion problem
Before testing anything, diagnose what is actually wrong. Most CRO work fails because the team tested a button colour when the real problem was a confusing form three sections down the page.
The diagnostic stack we use on every CRO engagement.
- Heatmaps. Click maps and scroll maps. Show where users click and how far they read. Catches the "users never see the CTA because they bounce at the hero" pattern. Microsoft Clarity is free and covers this; Hotjar is the paid equivalent.
- Session recordings. Watch real users navigate the page. Catches form-field abandonment, rage clicks, scroll loops where users miss something, and dead-end navigation. Microsoft Clarity is free and covers this.
- Form analytics. Field-level form analytics catch where users abandon. Form fields that block 20 percent of users are CRO gold. HubSpot, Zuko and Hotjar all do form analytics.
- User testing. Five real users (UserTesting, PingPong) attempting the conversion task. Five users will surface 80 percent of the obvious issues. Cheap and underrated.
- Quantitative GA4 analysis. Exploration reports for funnel drop-off. The "user took these three steps and then left" diagnostic that suggests where the funnel breaks.
The diagnostic phase usually surfaces three to five obvious issues per high-traffic page. The CRO testing then prioritises the issues by expected impact and tests one at a time.
Designing the test
The test design that works for SMB organic traffic.
One hypothesis per test
Multi-variant tests on small samples produce noise. State the change being made, the metric being moved, and the expected mechanism. "Moving the lead form above the fold will increase form submissions by 20 percent because users no longer have to scroll to see the conversion ask." One hypothesis. One change. One outcome metric.
Sample size and duration
For most SMBs with 1,000 to 5,000 organic sessions per month per landing page, plan on a 4 to 8 week test duration to reach significance on conversion-rate changes of 15 to 30 percent. Smaller effects (under 10 percent) take much longer (12 weeks or more) and often are not worth testing.
Below 1,000 sessions per month per page, formal A/B testing is usually not worth it. The honest workflow is to make informed changes one at a time, give each change 30 to 60 days, and watch the conversion-rate trend on the page. Less rigorous, but matched to the data volume.
Ship the winner, kill the loser
Once a test reaches significance, roll out the winning variant as the new default. Document the result. Move on. Do not leave A/B tests running once the result is in; the ongoing test loses revenue on the losing variant.
Document the test log
Keep a simple test log: page tested, hypothesis, change made, start date, end date, result, learnings. After 12 months the log is more valuable than any single test because the patterns emerge across pages. See reporting to stakeholders for how to surface CRO wins in the monthly report.
CRO without hurting SEO
The two things that hurt SEO during CRO work.
Cloaking. Showing different content to Googlebot versus users. Google penalises this. Server-side A/B testing tools handle it correctly by serving the same content to Googlebot consistently while testing variants on users. Client-side tools (VWO, Optimizely) handle it correctly by default if configured properly. Configure incorrectly and you ship a page that fails SEO and CRO simultaneously.
Dropping primary keywords from H1 or title tag during a test. A CRO test that rewrites the H1 to be more conversion-friendly without the primary keyword often costs ranking. Test the H1 wording within the constraint that the primary keyword stays. Better: test the conversion elements (forms, CTAs, trust signals, layout) rather than the H1 or title tag itself. See title tags for the SEO-side constraints.
The honest framing: CRO and SEO compound when they respect each other's constraints. The CRO team avoids cloaking and protects the on-page SEO elements. The SEO team avoids treating CRO as a separate budget and shares the landing-page-level data. The wins come from the intersection.
CRO tools worth paying for
The tool stack we use across Perth and WA clients.
- Microsoft Clarity. Free. Heatmaps, session recordings, basic form analytics. The default starting point for any SMB doing CRO. Generous data retention; minimal setup.
- Hotjar. Paid alternative to Clarity. Better polling and survey features. Worth it if you want on-page surveys alongside the heatmaps.
- VWO or Optimizely. Paid A/B testing tools. Both client-side, both SEO-safe when configured properly. Pick based on pricing fit; the feature parity is high.
- Google
Optimize. Retired in 2023. Do not use; replaced by VWO, Optimizely, or built-in GA4 experiments. - Zuko or HubSpot form analytics. For sites where form drop-off is the suspected bottleneck.
- UserTesting or PingPong. For five-user qualitative testing. Cheap, underrated, often surfaces things no analytics tool catches.
For most Australian SMBs, the starting stack is Clarity (free) plus a basic A/B testing tool when traffic justifies it. The expensive enterprise CRO stack adds value at scale; below 10,000 monthly sessions on the test pages, it rarely pays back.
The conversion rate optimisation service we run for clients bundles the diagnostic stack with the testing workflow because most clients do not have the bandwidth to maintain both internally.
Common mistakes
- Starting CRO on the highest-traffic high-intent pages, not on the homepage.
- Diagnosing the actual conversion problem with heatmaps and session recordings before testing anything.
- Testing one hypothesis at a time on samples large enough to reach significance.
- Keeping primary keywords intact in H1 and title tag during CRO testing.
- Treating CRO as compounding with SEO rather than competing for budget.
- Running A/B tests on pages with under 1,000 monthly organic sessions and trusting the results.
- Testing button colours when the form three sections down is the real problem.
- Cloaking variants and accidentally tripping Google's anti-cloaking signals.
- Dropping primary keywords from H1 during a CRO test and losing ranking.
- Treating CRO as a separate budget line item without sharing the measurement infrastructure.
Perth and WA context
Three CRO patterns specific to Perth and WA businesses.
Phone CTAs convert better than form CTAs for trade businesses. Trades, plumbing, electrical, automotive: the conversion ask should be the phone number prominent at the top and the bottom of the page, with the form as a secondary path. Many Perth trade sites we audit have buried the phone number and led with the form, which fights the dominant lead behaviour. See trades SEO.
Trust signals matter more for high-consideration services. Healthcare, legal, financial: testimonials, professional registrations, named credentialled authors, named partners, About-page bios. The CRO wins on these sites usually come from trust elements rather than from CTA placement. See E-E-A-T explained, legal SEO and healthcare SEO.
Local trust signals matter for service-area pages. Photos of actual Perth crew, named local team, suburb-specific testimonials, references to local landmarks. Generic-feeling service-area pages convert worse than pages with visible local credibility. See Local SEO and Local SEO Perth.
For the wider context, the SEO KPIs chapter covers how CRO wins feed the primary KPI, the GA4 for SEO chapter covers the conversion-event setup that CRO depends on, the reporting chapter covers how to surface CRO wins in the monthly report, the content depth chapter covers the content-side of conversion, the E-E-A-T explained chapter covers the trust-signal layer that drives conversion on high-consideration pages, and the CRO service is what most clients engage when they want the discipline run for them. Clients ready to start can engage the SEO service with CRO bundled, or begin with a free SEO audit.