A founder asked me last week what "profit optimisation" actually means when an agency uses the phrase. I told her the honest answer. Most of the time it means split tests on the product page.
That's it. The whole thing dressed up in CFO-flattering language. Split tests on the PDP, maybe a checkout tweak, maybe a post-purchase upsell flow. The agency calls it profit optimisation. The brand pays the invoice. The AOV ticks up 6%. Eight months later the gross margin is 200 basis points lower than it started because nobody touched the levers that actually move profit.
I want to break down what profit optimisation should look like if it's being done seriously. Not five things to add to your to-do list. Five levers, ranked by how much profit they can move at a £10M-£100M ARR brand. Pick two. Run them deep. Ignore the rest until you have results on the first two.
The five levers of real Shopify profit optimisation
Most engagements fail because they pick three or four levers and run them shallow. Two levers run deep produce more contribution margin than four run shallow, every time we have measured it. Here they are, in order of ceiling.
1. Gross margin engineering (highest ceiling, lowest agency interest)
Of every brand I have audited at PMD, the one with the largest unrealised profit opportunity was on the cost side, not the revenue side. Their packaging cost was 23% of COGS. Their fulfilment cost was 18%. Their merchant fees were 2.4%. Every one of those was a profit lever.
Agencies do not sell this work because it is not sexy. There is no split test result to put in a deck. But a 2-point gross margin lift on a £20M brand is £400k of contribution margin per year, and that work compounds. Every order from that day on is more profitable.
The agencies that genuinely optimise profit start with a margin teardown by SKU before they touch the funnel.
2. Subscription retention (the LTV multiplier)
If a Shopify brand has a subscription program, month-one-to-month-three churn is the most consequential metric in the business. A 10-point reduction in month-three churn typically lifts 12-month LTV by 40-70%. That LTV improvement raises the maximum allowable CAC, which loosens the top-of-funnel constraint, which lets the paid team be more aggressive. Cadence — a subscription brand we work with — moved their checkout subscription take rate to roughly 70% on the back of exactly this kind of work.
We rebuilt the subscription experience for a £22M apparel client and dropped month-three churn from 38% to 19% in 90 days. The 12-month LTV grew 64% on the next cohort. None of that work touched the PDP or the checkout. It was customer-portal architecture and welcome-series sequencing.
If you have a subscription program and your agency is selling you split tests instead of cohort-level churn analysis, you are being optimised on the wrong axis.
3. The funnel diagnostic (not the funnel redesign)
Everyone wants the funnel redesign because the deliverable is visible. The diagnostic is more valuable.
A real profit-optimisation diagnostic answers two questions. Which buyer paths are generating the most profit per visitor? And which paths are leaking? Those two questions get answered with attribution data and cohort behaviour, not with a heatmap. We covered this in detail in our Shopify CRO Audit Framework — the seven-part diagnostic is the prerequisite to every redesign decision we make. Skip it and you redesign the wrong thing 80% of the time.
The agencies that ship redesigns without the diagnostic are betting their reputation on aesthetics. We bet on the diagnosis.
4. Checkout and cart messaging (smaller ceiling, faster to ship)
This is the lever most agencies oversell because it produces visible split-test wins. The reality is the ceiling is modest. A 6-8% conversion lift on a well-optimised checkout is a good result. On a £15M brand that is £900k-£1.2M of incremental revenue, which is real money but a fraction of what gross-margin engineering and subscription retention can produce.
The reason we still run this work is that it is fast. We documented one example where strategic cart thresholds drove £25,535 in additional monthly revenue from a single split test. Six weeks of work, five-figure monthly lift, ongoing. Worth doing, but never as the headline play in a profit-optimisation engagement.
5. Post-purchase architecture (highest margin per touch, modest ceiling)
The post-purchase moment is the highest-margin revenue surface in the entire store. The customer is in a buying state, the payment friction is gone, and you pay no acquisition cost on the incremental order value. Most brands run a take rate of 3-7% on post-purchase upsells when 15-25% is achievable with offer-architecture work.
The reason this lever sits at five despite being the highest-margin surface is the ceiling. Even a great post-purchase flow on a £15M brand caps out around £500k of incremental annualised revenue. Worth doing aggressively, but not the single biggest lever in the business.
What a serious profit-optimisation engagement actually looks like
A real profit-optimisation engagement spends the first three weeks on the diagnostic. Gross margin teardown by SKU. Cohort retention curves. Path-level revenue attribution. Then it commits to the two highest-ceiling levers from the five above. Not all five. Two. Done seriously.
The diagnostic costs about three to five weeks of work. The first two implementation cycles ship within eight weeks. Meaningful profit lift typically shows in the second monthly P&L after implementation begins, which means realistic expectation is 90-120 days from kickoff to first measurable contribution-margin uplift.
Brands that compare this to a "we ship 20 tests per month" agency are comparing different things. One is profit optimisation. The other is test volume sold as profit optimisation.
The test that tells you whether you are being sold real profit optimisation
If your agency's profit-optimisation pitch consists entirely of split tests, you are buying CRO with a more flattering label on the tin. That is fine if you know that is what you are buying. It is a problem if you have been told it is something more.
The diagnostic test for any pitch is simple. Does the agency know your gross margin by SKU? Have they asked about your subscription churn cohort by cohort? Have they modelled your CAC ceiling? If the answer to all three is no and they are already presenting test ideas, they are selling you CRO with a paint job.
PMD's full-funnel CRO and profit optimisation engagement opens with all three of those questions. We do this because the founders we work with are tired of being optimised on the wrong axis. Eight months of work that lifts revenue but compresses margin is a result the CFO has to defend. We would rather give them a result that defends itself.
What I tell every founder who asks me this
Pick two levers. Not five. Not three.
If you have a subscription program, lever two is non-negotiable. If your gross margin is below 60%, lever one is non-negotiable. If neither applies, levers three and five are your starting point and four is what you do once the diagnosis is finished.
And do not let an agency tell you they are doing profit optimisation if their first deliverable is a list of split tests. Ask them what your gross margin looks like by SKU. If they cannot answer, they are not optimising profit. They are optimising convenience. If you'd rather skip the pitch deck and just talk through your numbers, grab a 30-minute call with Paddy McLarnon.
FAQs
What is the difference between CRO and profit optimisation?
CRO is one input to profit optimisation. Profit optimisation includes gross-margin engineering, subscription retention, cost-of-acquisition modelling, and post-purchase architecture in addition to traditional conversion-rate work. An engagement that calls itself profit optimisation but only does split tests is mislabelled.
How long does a profit-optimisation engagement take to produce results?
The diagnostic phase takes three to five weeks. The first two implementation cycles ship within eight weeks. Meaningful profit lift typically shows in the second monthly P&L after implementation begins. Realistic expectation is 90-120 days from kickoff to first measurable contribution-margin uplift.
Can a Shopify brand below £5M ARR run this kind of engagement?
The data volume below £5M ARR is not usually sufficient to make cohort analyses statistically meaningful. Brands below this threshold are better served by simpler funnel work and acquisition optimisation. The five-lever framework above is built for brands at £10M-£100M.
What is the typical fee structure for a profit-optimisation engagement?
Either fixed-fee per quarter with a defined scope, or a hybrid model with a retainer plus a performance-linked component tied to verifiable contribution-margin lift. We avoid pure performance-fee structures because they incentivise short-term tactics that compress long-term LTV.
How does this differ from a growth marketing engagement?
Growth marketing focuses on acquisition: paid media, conversion rate optimisation, lifecycle. Profit optimisation focuses on contribution margin per customer once they are acquired. The two engagements often run in parallel. Confusing them is how brands end up with growing revenue and shrinking profit. For deeper material, the PMD CRO learning hub collects our video tutorials and longer-form CRO breakdowns.
Should we pick the levers ourselves or have an agency choose?
The agency should propose the two levers, but they should propose them with the diagnostic data in hand. An agency that proposes levers before completing the diagnostic is guessing. An agency that defers the choice to you entirely is abdicating responsibility for the engagement outcome.
Full-funnel CRO. Profit obsessed.
Want this on your P&L?
We help subscription and high-LTV Shopify brands — including Cadence, Routine, Maelove, MyoMaster and others — turn cold traffic into post-click profit. Strategy, copy, design, development and CRO under one roof.