CRO basics: what actually moves revenue

Conversion rate optimisation has a reputation for being about testing button colours. Done properly it's the highest-return work in digital marketing, because it improves every channel at once.

Start by finding where people leave

Don't guess. Look.

In GA4: build a funnel exploration from landing page through to conversion. The step with the largest drop-off is where to work.

On mobile specifically. Mobile conversion rates are typically far lower than desktop, and mobile is usually the majority of traffic. Optimising desktop when 70% of visitors are on phones is solving the wrong problem.

Watch session recordings if you have a tool for it. Twenty recordings of people abandoning a form teaches you more than any amount of theorising.

The four places revenue leaks

1. The landing page doesn't match the promise. Someone clicks an ad about a specific thing and lands somewhere generic. Highest-impact, cheapest fix.

2. The form asks too much. Every field costs conversions. Ask for what you need to have the next conversation, nothing more.

3. Checkout friction. Forced account creation, unexpected shipping costs at the final step, limited payment options. For South African e-commerce, not offering the payment methods people actually use is a large and quiet leak.

4. Nothing answers the obvious objection. Price, delivery time, returns, whether you're legitimate. If a visitor has to hunt for it, many will leave instead.

What to test, in priority order

1. Message match between ad and page
2. Form length — remove fields and measure
3. The primary CTA — position, wording, prominence
4. Proof placement — move testimonials and trust signals above the fold
5. Price presentation — clearer, earlier
6. Page speed — see why your website is slow

Everything below that is refinement.

Testing honestly

You probably don't have the traffic for A/B testing. Detecting a 10% improvement reliably needs thousands of conversions. Most sites don't have that.

What to do instead: make substantial changes, one at a time, and compare periods. It's less rigorous and it's what's available. Just be honest that you're observing rather than proving, and be wary of seasonality.

Test big changes. A completely different page structure produces a difference you can actually detect. Two similar headlines don't.

The metric to watch

Not conversion rate alone — revenue per visitor.

A change can raise conversion rate while lowering average order value and leave you worse off. Revenue per visitor captures both.

Where to start this week

1. Open GA4, find the biggest drop-off in your main funnel
2. Go through that step yourself on a phone
3. Note every point of friction
4. Fix the most obvious one
5. Measure for a fortnight

That loop, run monthly, beats any testing tool you could buy.

The compounding argument

A 20% conversion rate improvement means every channel gets 20% cheaper per acquisition. Paid media, organic, referral, email — all of it, permanently, without additional spend.

That's why I'd fix the page before scaling the ads. Scaling a page that doesn't convert just buys the problem more traffic.