The first time a Malaysian merchant asks us to "improve conversion", roughly eight out of ten times they mean one thing: lift the headline conversion-rate number that GA4 prints on the homepage of their analytics. It is a fair instinct. It is also a slightly dangerous one.
Conversion rate is a useful metric, but it is famously easy to game. Cut your marketing spend, lose the casual traffic, and your conversion rate climbs without anyone actually buying more. Run a sale, lose your margin, and the rate goes up too. We have watched a client triple their stated conversion rate inside a quarter while quietly losing money on every order.
This is a short note on what we actually look at instead.
What conversion rate is for
Conversion rate is a leading indicator of friction. It is sensitive — too sensitive, really — to anything that slows a visitor down between landing and the action you care about. If conversion drops 25% week-over-week, something on the page changed, the traffic mix changed, or the offer changed. It is good at telling you "go look", not at telling you "here is the answer".
It is also useful as a comparison metric inside a single funnel — variant A versus variant B, page X versus page Y, last week versus this week — as long as you hold the traffic source roughly constant. Across different traffic sources, raw conversion rate becomes almost meaningless.
The three numbers we read alongside it
1. Cost per qualified conversion (CPQC)
The all-in cost — paid media, content, your team's hours — divided by the number of conversions that the sales team agrees are worth following up. CPQC drops when conversion is real; it spikes when conversion is junk traffic gaming the funnel. For most Malaysian B2B clients, this number cuts through a quarter of confusion in a single slide.
2. Twelve-week revenue per acquired customer
The revenue a customer actually generates in the twelve weeks after their first conversion. This rewards channels and pages that bring in customers who come back, not ones that maximise first-purchase noise. It also exposes one of the more popular kinds of self-deception in Malaysian eCommerce: a discount-heavy hero that fills the funnel with one-and-done buyers who never return.
3. Margin per converted session
Revenue minus cost-of-goods minus payment fees, divided by sessions. For merchants who run promotions to chase conversion-rate gains, this is the number that quietly tells the truth. We have seen merchants lift conversion rate by 40% while margin per session fell by half. The accounts team noticed first.
A conversion you cannot defend in the boardroom is just noise dressed up as a number.
What this looks like in practice
For an SME with a small finance team, you do not need a dashboard with forty widgets. You need four numbers, refreshed weekly, reviewed in a thirty-minute meeting on Monday morning: conversion rate, CPQC, twelve-week revenue per customer, margin per session.
Those four numbers, plotted as a small-multiple chart across thirteen weeks, will tell you whether your website is genuinely earning its keep, or whether you have been celebrating the wrong scoreboard. We build this exact one-page report for every CRO retainer client, in either Looker Studio or a small custom dashboard, depending on which their finance team prefers to live in.
The honest answer when someone asks "what is a good conversion rate"
Lower than you think, and it does not really matter. For a Malaysian B2B services site, between 1.5% and 3% is normal. For a mid-market eCommerce store, somewhere between 1.2% and 2.8% on mobile, slightly higher on desktop. For a fintech account-opening page, often under 1% — and entirely defensible if the customer lifetime value is high.
If your conversion rate is wildly out of these bands, the question is not "is the number good or bad" but "why is it where it is". The first answer is almost always traffic mix; the second is almost always page-level friction. Both are knowable. Both are fixable. Neither is fixable in a single Monday morning.
Want a small report like this for your business? The first one is free as part of our diagnostic call. Get in touch.
