Conversion Rate Optimisation
Most Shopify stores don't have a traffic problem — they have a conversion problem nobody has measured properly. We run a steady testing programme that turns the visitors you already pay for into more revenue, and we show our working the whole way.
Storefronts designed around the buying decision.
Usually a monthly retainer. Plenty of people start with a one-off audit and decide from there.
You probably don't have a traffic problem.
Doubling your ad spend to fix a conversion problem is the most expensive mistake in ecommerce, and it's an easy one to make. Here's what it tends to feel like from the inside.
Traffic's up. Revenue isn't.
You're paying more for the same visitors and the conversion rate hasn't moved in eighteen months. The problem isn't at the top of the funnel.
Everyone's got an opinion on the product page.
The last three changes were somebody's preference. Nobody measured whether any of them worked, so nobody can say what to try next.
You've been quoted for a redesign.
A full rebuild is a very expensive way to find out your checkout was the problem. We'd rather find that out in week two.
Look, guess, test, keep the winners.
The same loop every month. It isn't glamorous — it's just the thing that reliably works.
- 01Weeks 1–3
Have a proper look
Analytics, session replays, heatmaps, a walkthrough of every step from ad click to thank-you page — and a chat with your support team, who already know exactly what’s broken.
A prioritised list of leaks, sized by what they're costing you.
- 02Week 4
Write down what we think
Every test starts as a sentence: we think this will do that, because of this. If we can't write it down, it isn't ready to run.
A ranked roadmap you've signed off.
- 03Ongoing
Run it properly
Right sample size, agreed duration, no peeking on day three and calling it a win. Losers get reported as loudly as winners — roughly half of tests lose, and that's normal.
Two to four live experiments at any time.
- 04Ongoing
Keep the winners
Winners get built into the theme properly rather than left running in a testing tool forever. Then what we learned feeds the next idea, which is where the compounding comes from.
A monthly catch-up in plain English, and a revenue number.
Every step between the ad click and the thank-you page.
Ordered by what they're worth, not by what's fun to work on.
- Product detail pages
- Collection & search
- Cart & checkout
- Homepage & landing pages
- Bundling & upsell
- Site speed
- Mobile navigation
- Trust & social proof
- Subscription journeys
- International & currency
- Post-purchase
- Forms & lead capture
One team, one number, every month.
No account manager relaying messages — you talk to the people running your tests.
- A full conversion audit in month one
- A roadmap prioritised by revenue, not effort
- Two to four live experiments at any time
- Test build, QA and deployment handled by us
- Winners implemented natively in your theme
- A monthly session with the people doing the work
- A shared Slack channel with your team
- No lock-in — 30 days' notice, always
The questions we always get asked
How much traffic do we need for this to work?
Around 20,000 sessions and 400+ orders a month is where classic A/B testing starts giving reliable answers in a sensible timeframe. Below that we still run a programme, it just leans more on research and changes we're confident enough to ship outright. We'll tell you which camp you're in before you sign anything.
Do you need to redesign our store first?
Almost never. Most of our biggest wins have been on themes the client already had. If we do think a rebuild is justified, we’ll show you the test data that says so first.
Which testing tools do you use?
Whatever you've already got, happily. If you're starting fresh we'll suggest something that fits your traffic and budget rather than defaulting to the priciest option.
How quickly will we see something?
The audit lands in month one and usually turns up a few things worth fixing straight away. Meaningful test results tend to arrive in months two and three. Anyone promising a lift in week one is guessing.
What happens to tests that lose?
They get reported, explained and fed into the next idea. About half of all tests lose or come out flat — that's simply how it works, and only ever showing you winners would mean hiding something.
Is this a project or a retainer?
CRO only compounds if it keeps going, so it’s monthly. Most people start with a one-off audit and roll into the programme once they’ve seen the list of what we found.
