More revenue from the
traffic you already pay for.

A design agency is paid to produce a new experience; we are paid to find out whether it performs. Both are legitimate, but they answer different questions. If you already know what you want built, hire the design agency. If the honest position is that nobody can say why the current site underperforms, research and testing will get you there for less than a redesign — and will tell you which parts of the redesign to keep.

Research takes three to four weeks, and the first experiments go live shortly after. Individual tests usually need two to four weeks to reach a decision, so the first shipped winner typically lands in month two or three. Anyone promising a lift in week one is not measuring properly.

No, and we would be cautious of anyone who does. We guarantee the process — research before hypotheses, metrics and sample sizes agreed before launch, honest analysis afterwards — and we report the programme's contribution in revenue so you can judge it yourself. Guaranteeing outcomes on a statistical process only works if you are willing to misreport results.

Then we will be the ones to tell you, at the quarterly review, in revenue net of our fee. It happens — usually because the constraint was never the site. If that is where we land, the useful outcome is knowing where the real bottleneck is, and we would rather end an engagement honestly than keep invoicing against a number nobody is checking.

Because anything shorter cannot produce a trustworthy result. A single test needs two to four weeks to reach significance, roughly three in four fail to beat control, and winners need time to be built and validated. A three-month engagement would let us run a handful of tests and report noise. The minimum protects the quality of the answer, not our revenue.

As a rough guide, around 8,000 sessions and 300 orders a month per tested template makes a normal A/B programme viable. Below that we lean on research, higher-impact changes tested one at a time, sequential analysis, and painted-door tests rather than pretending we can detect a 3% lift. If your traffic will not support a programme yet, we will say so on the first call.

For research and most of the site, no. It matters for checkout: Checkout Extensibility and script-level customisation are Plus features, so on standard Shopify we focus on everything up to checkout — which is usually where the larger opportunity sits anyway.

Usually, yes. Most of our clients have an in-house team and often a creative or media agency. We tend to own research and the experiment programme, and either build the winners ourselves or hand your developers a specification and the test code. We are not looking to displace people who are already doing good work.

You do, entirely. Everything runs in your Shopify store, your testing tool and your analytics accounts. At the end of an engagement you keep the research archive, the full test history including the losers, and the code — so the next team starts from what you have already learned instead of retesting it.