Conversion rate optimisation agency · Australia
A conversion rate optimisation agency that scores tests on margin.
Our conversion rate optimisation services cover the parts of the funnel where buyers drop out: landing pages, A/B and split testing, user experience research, and the checkout or quote path. Most CRO programs report platform uplift and ignore the bank. We find where your funnel actually leaks, test the steps that move contribution margin, ship the winners, and hand your team the playbook.
30 minutes with Andy McMaster, founder. No pitch, and the written fix list is yours.
- Revenue influenced across attribution and conversion work since 2016
- $3.6B
- Standard CRO engagement length
- 12 wk
- Typical Meta-vs-CRM revenue gap, before and after. We close it before we test.
- 23% → <5%
CRO services
Four parts of the funnel, one rule for all of them.
CRO turns more of your existing traffic into customers or leads without buying more clicks. Every service below runs on the same rule: confirm the tracking counts conversions correctly, test only what your traffic can support, and score the result on margin.
01 / 04
Landing page optimisation
The page your ads and search results send people to. We check that it says what the ad or the search promised, that the offer is clear before anyone scrolls, that the form works on a phone, and that the page loads quickly on mobile. A high bounce rate from paid traffic is often a mismatch between the ad and the page rather than a traffic problem.
Landing page design for paid ads02 / 04
A/B and split testing
An A/B test runs one change against the original on the same page. A split test sends visitors to two different URLs, which suits a rebuilt page or a new checkout. Before we build a variant, we check your traffic can reach a reliable answer in a sensible time, and we run it in the testing tool you already have. If the sums say a test cannot finish, we fix the page instead of testing it.
03 / 04
User experience research
Where the hypotheses come from. Funnel reports in Google Analytics 4, session recordings and heatmaps in Hotjar or Microsoft Clarity, form-field drop-off, and the objections customers raise on sales calls and in enquiry emails. User behaviour shows where people stall; customers tell you why. The research feeds the UX design changes we test, and on a low-traffic site it is often worth more than any test.
04 / 04
Checkout and quote path
The last steps, where a visitor becomes a customer or a lead. For ecommerce that is the cart and checkout: delivery costs that appear late, forced account creation, payment options, slow mobile forms. For lead generation it is the quote or booking path: how many fields are required, whether calls are tracked, and how fast someone responds once the enquiry lands. It is where each lost visitor costs the most.
How we run CRO
Four principles. They constrain the work as much as they direct it.
Principle 01
Measurement before testing
Most CRO programs run experiments on a measurement layer that does not actually count conversions correctly. The first 30 percent of every engagement is reconciliation: the platform-reported conversion needs to match the bank before any test result is trustworthy.
Principle 02
Margin-aware uplift, not vanity uplift
A 12 percent lift in add-to-cart on a product range with 8 percent gross margin is not a win. We score every test against contribution margin and only report uplifts that actually clear margin. The number gets smaller; the number is also real.
Principle 03
Test sequencing tied to where the leak actually is
Funnel-leak diagnostics tell us where the worst conversion gap is, by step. That is where the first round of tests goes. Most agencies test the headline first because it is cheap to A/B; we test the step that is bleeding the most revenue first.
Principle 04
Permanent changes only
A test that wins gets shipped to default and the test infrastructure is removed. We do not run permanent multivariate frameworks that fragment future tests and degrade page speed. Speed itself is one of the largest CRO levers.
Deliverables
What an engagement actually delivers.
01
Funnel-leak diagnostic across the entire purchase or signup path. Step-by-step conversion rate, drop-off, and contribution-margin impact.
02
Hypothesis register with prioritisation against contribution margin and effort-to-implement. Top six tests scoped and ready to run.
03
Quantitative analytics rebuild: GA4, Meta CAPI, Hotjar or Microsoft Clarity, and the order of operations for properly running tests on a clean measurement stack.
04
Test execution in VWO, Optimizely, Convert, a Shopify testing app, or whatever your stack already runs. We work with what is in place rather than insisting on a tool change.
05
Margin-aware reporting: every test result reported as platform uplift, contribution-margin uplift, and statistical confidence. Three numbers, one decision.
06
Documented test playbook for the in-house team. After the engagement, your team operates the test cadence without us.
Who CRO is for
- Scaling on paid acquisition at meaningful spend ($20K+ a month)
- Working measurement layer (or willing to fix it first as part of the engagement)
- Funnel that converts somewhere; we are optimising it, not building it from scratch
- Owner or operator who wants the work explained, not delivered behind a curtain
Who it isn't
- Need a new website first (web design and builds are quoted separately, after the call)
- Funnel converts at zero (CRO compounds on a working baseline)
- Looking for ongoing managed-test retainer with no end (we ship and hand over)
- Sub-$2M revenue trying to optimise before the offer is proven
Pricing
Fixed-scope. Twelve weeks. Written number up front.
We scope and price the engagement on the call, tied to what you want the testing to achieve, then send a written, fixed-scope proposal with the number in it. A standalone CRO engagement is a fixed-scope twelve-week piece. Embedded inside a PROFIT framework engagement, CRO is the Optimise pillar and is part of that one engagement. No retainer drift.
What sets the number: how many funnel steps are in scope, the size of your stack, and whether the measurement needs rebuilding before a test result can be trusted. What does not: your ad spend. The fee is never a percentage of it, or of your revenue.
How we price every engagementWhere to go next
CRO is rarely the only fix.
If the funnel is leaky and the data is wrong, fix the data first or the testing won't be trustworthy. If you want the engagement to span measurement, paid acquisition and CRO together, the PROFIT framework is the route.
Related services
In their words
What owners say once the numbers add up.
Sales +140%, past $25M
“Profit Geeks rebuilt our whole sales engine, and not just the ad accounts. …”
−$1K/week → $10K+/day
“We were quietly bleeding about a thousand dollars a week and had no idea why. …”
2 → 15+ jobs/week, $10M+ turnover
“We went from scraping together two installs a week to running four crews and fifteen-plus jobs a week, north of $10M turnover. …”
$1.42M ad spend recovered, year one
“After iOS, our Meta numbers stopped matching the bank, and we'd basically been writing the gap off as “just tracking.” …”
+312% blended ROAS ($4.8M → $9.1M)
“We didn’t spend a dollar more on ads. …”
Reasonable questions
What you're probably thinking.
01We've been burned by an agency before.
02How do I know it'll actually work for my business?
03What if there's nothing worth fixing?
04What does it cost, and what am I signing up for?
Frequently asked
What buyers ask a CRO agency.
What does a CRO agency do?
It works out why visitors who could buy or enquire do not, then changes the site until more of them do. The work runs in a loop: research into how people use the site, a written hypothesis for each problem, a test or a straight fix, and a decision once the result is in. A good agency checks the tracking before any of that, because a test is only as reliable as the conversions it counts. We add two rules of our own: every result is scored on contribution margin rather than conversion count, and the winners are shipped and handed to your team with a written playbook, so you are not renting the program.
Is there a CRO agency in Australia?
Yes. Profit Geeks is an Australian conversion rate optimisation agency, the trading name of AM Marketing Group Pty Ltd (ABN 37 613 053 286), with a Sydney head office and a second office in Hamilton, Brisbane. We run CRO for businesses anywhere in the country, mostly on weekly calls and shared documents, so you do not need to be near either office. There are other Australian CRO agencies too. Whoever you talk to, ask how they score a winning test, who builds the variants, and what you keep when the engagement ends.
How much do CRO agencies charge?
Most charge in one of two ways: a monthly retainer for an ongoing testing program, or a fixed fee for a defined project. Either way, the number is driven by the same things: how many pages and funnel steps are in scope, how much research sits behind each test, who designs and builds the variants, whether the tracking has to be fixed before testing can start, and how long the work runs. Testing tool licences are usually billed separately by the tool vendor. Ours is a fixed fee for a twelve-week engagement, quoted in a written proposal after the free audit call. It is never a percentage of your ad spend or your revenue, and we do not publish a rate card.
How we price every engagementHow much does CRO cost?
More than the agency fee. Add a testing tool licence if you need one, developer time to build variants and ship the winners, and the weeks of traffic each test needs before it can be called. Then set the total against what a lift is worth, using your own numbers. Say a page gets 20,000 sessions a month, converts at 2 percent, and each conversion is worth $150 in margin: it earns $60,000 a month. A 10 percent relative lift, from 2 to 2.2 percent, adds $6,000 a month. If the program costs more than that over the time it takes to find and ship the win, it does not pay. Those figures are arithmetic, not a benchmark. We run the same sum with your numbers on the audit call, before anything is quoted.
What are the benefits of CRO?
There are four. More sales or enquiries from the traffic you already pay for, because the media bill stays the same while more visitors convert. A lower cost per acquisition, which raises the price you can afford to pay for a click, so paid channels can scale further before they stop making money. Better knowledge of your buyers, because the research behind each test tells you why people hesitate, and that feeds the ads, the offer and the sales script. And a faster, simpler customer experience, since much of the work is removing friction and fixing defects. None of it counts unless the tracking records conversions correctly, which is why we check that first.
What is the difference between SEO and CRO?
SEO brings more of the right people to the site from organic search. CRO gets more of the people who arrive, from any channel, to buy or enquire. SEO is judged on rankings, organic traffic and the enquiries it produces; CRO on conversion rate and revenue per visitor. They overlap more than most people expect: page speed, headings that answer what someone searched for, and content that matches intent help both. Run together, each makes the other worth more, because extra traffic lands on pages that convert better. We do both, on the same tracking, so organic traffic and conversion rate are read from one set of numbers.
Which agencies offer the best conversion rate optimisation services?
Rankings and award lists tell you little about how an agency would handle your site, so judge your shortlist on how each one works. Whether you call it a CRO agency or a conversion rate optimisation company, ask four things. How do you check the tracking before a test starts? How do you decide a test has won, on conversion count or on margin? Who designs and builds the variants, and will those people be on our calls? What do we own when the engagement ends? Ask for the track record in a funnel like yours, too. An agency that answers plainly and in writing is worth a proposal. One that quotes uplift percentages before it has seen your numbers is selling the forecast, not the work.
What is conversion rate optimisation?
Conversion rate optimisation (CRO) is the practice of improving the percentage of visitors who complete a desired action on a website (purchase, signup, lead). The work spans funnel diagnostics, hypothesis development, A/B and multivariate testing, and shipping the winning variant. Done well, CRO is one of the strongest levers in paid acquisition because every uplift compounds across the existing media spend.
What conversion rate optimisation is, in fullWhat is a good conversion rate in 2026?
There is no universal good conversion rate. The number that matters is whether each step of your funnel is converting at the rate the platform's audience and your offer should support. Direct-to-consumer ecommerce on warm traffic typically converts at 2 to 4 percent; lead-gen on high-intent paid search typically converts at 5 to 15 percent. The funnel-leak diagnostic returns your number against the band for your category.
What a good conversion rate is in AustraliaHow long does a CRO engagement take to show results?
First test results: weeks four to six. Compounding uplift across the entire funnel: months three to six. CRO is iterative; the velocity comes from running the right tests in the right order, not from running more tests. Most engagements ship six to twelve material tests in the first quarter.
How the testing works, in detail.
Do you do CRO without rebuilding our tracking first?
Only if the existing measurement is genuinely clean. If platform-reported conversions match the CRM within five percent, we go straight to testing. If they do not, we rebuild measurement first; otherwise every test result is unreliable. About two-thirds of operators we audit need the rebuild first.
What CRO tools do you work with?
VWO, Optimizely, Convert, Google Optimize successors (Optimize sunset in 2023), Shopify-native testing apps, and custom test infrastructure for headless commerce or custom apps. We work with whatever is in place. Forcing a tool change is rarely the highest-impact move.
Is CRO still worth doing if our paid acquisition is broken?
Usually no. CRO compounds on top of working acquisition; it does not fix broken acquisition. If your paid spend is below break-even ROAS, fix attribution and acquisition first. The CRO engagement is the second move, not the first.
How much traffic do we need before an A/B test actually means anything?
Work backwards from the maths, not the tool. To detect a 10 percent relative lift on a 3 percent baseline at 95 percent confidence and 80 percent power, you need roughly 50,000 visitors per variant. Move the baseline to 8 percent and the target lift to 20 percent and it drops to about 5,000 per variant. The rule that survives contact with reality: you want a few hundred conversions per variant, not a few hundred visitors. Below roughly 1,000 conversions a month across the whole funnel, most tests will run for months and still return noise. Calculate the required sample before you build the variant, then decide whether the test is worth the calendar time.
What do we do if we do not have that much traffic?
Stop testing and start fixing. Three things pay at low volume. First, defects: forms that fail validation on mobile, a phone number that is not tappable, a page that takes six seconds on 4G, a thank-you page that never fires. Those do not need a test, they need a fix. Second, qualitative work: session recordings, form-field drop-off analytics, and listening to twenty recorded sales calls will tell you more than an underpowered experiment. Third, test upstream where the volume already is. Your ads collect far more impressions than your site collects sessions, so creative and offer tests reach a decision weeks earlier than anything on-site.
Why does most CRO advice not work for lead-generation businesses?
Nearly all published CRO material is written from ecommerce data, where the conversion is the sale, the value is known the moment it happens, and traffic is high enough to test weekly. Lead generation breaks all three. The on-site conversion is an enquiry, not revenue, so a variant that lifts form-fills 20 percent can lower booked revenue if the extra enquiries are unqualified. Value is not known for days or weeks. Volumes are usually an order of magnitude lower, so the test never reaches power. The outcome is ecommerce tactics applied to funnels where the real constraint sits after the form, in qualification and response time.
Our funnel is a form and a phone call. What should we test first?
In this order. One, what happens after submission: response time, routing, and whether the enquiry reaches a human at all. Fix that before you touch the page, because no page change recovers a lead nobody rang back. Two, the form itself, specifically field count and which fields are required. Every field you make required trades volume for qualification, and that is a business decision, not a design one. Three, the offer on the page. Free quote attracts everyone, including people you cannot make money on. Four, page weight and load time on mobile 4G. Five, and last, headline and hero copy, which is cheap to test and rarely where the money is.
How do you score a conversion-rate change against contribution margin instead of conversion count?
Convert the test result into dollars before you decide anything. For each variant, take the conversions, multiply by that variant's average order or deal value, subtract cost of goods, delivery, payment fees, and any variable fulfilment or sales cost, then divide by sessions. You now have contribution margin per session, which is the only figure comparable across variants. Two things fall out of doing it this way. A variant can win on conversion rate and lose on margin per session, usually because it shifted the mix toward cheaper products or lower-value enquiries. And a variant can lose on conversion rate and still be the right one to ship, because it filtered out buyers who never cleared margin.
Can the testing tool itself hurt our conversion rate?
It can, and almost nobody measures it. Client-side testing tools inject a script that has to run before the page paints, which is why you get flicker, the original showing briefly before the variant swaps in. The anti-flicker snippet used to hide that works by hiding the page until the script responds, so a slow tool means a blank screen. Both cost conversions and both hit mobile hardest. Measure it rather than assume it: serve one copy of the page with the tool's script and one with it removed, split at the server or the edge so the tool is not doing the splitting, then compare Largest Contentful Paint and Interaction to Next Paint on mobile across the two. If the overhead costs more than the tests win, take the tool off the pages carrying your traffic and ship changes server-side.
Next step
Limited engagement intake. Bring the funnel diagnostic.
If your conversion rate has flatlined and the agency is testing headlines, the next step is a free 30-minute profit audit. We will look at your funnel and tell you, in writing, where the worst leak is and what testing it would cost.