Doubling traffic is expensive. Improving what happens to the traffic already arriving is usually cheaper, faster and compounds with every other channel you run.
What we look at
- where people abandon a form or checkout
- whether the page answers the question that brought them
- how many steps sit between intent and completion
- what raises doubt at the moment of deciding
- whether the thing actually works on the devices people use
Evidence, not opinion
Redesigning based on what the loudest person in the room dislikes is how sites get worse. We start with data: which pages lose people, at which step, on which devices.
Sometimes the answer is unglamorous. A form failing silently on mobile Safari. A price only visible after three clicks. A phone number that is not a link.
Forms
Every field costs completions. We remove what is not genuinely needed, make optional fields visibly optional, and validate in a way that helps rather than scolds.
Then we test that submission actually works end to end — including what happens on failure, which is usually untested and usually bad.
This site is a live example: its form once returned success while sending nothing, because a hidden anti-spam field named website was being autofilled by browsers. Every enquiry was silently discarded and the page said thank you.
Trust at the moment of deciding
Doubt is specific: who am I dealing with, what will this cost, what happens after I send this. Reviews, clear pricing ranges, named people, a stated response time and a visible privacy policy each remove a reason to hesitate.
Testing honestly
A/B testing needs enough traffic to reach significance. Below that, running tests produces noise interpreted as insight. For lower-traffic sites we recommend qualitative evidence and fixing obvious defects rather than pretending to statistical rigour.
FAQ
What conversion rate should we expect?
It varies far too much by industry and traffic mix for a benchmark to be useful. Your own trend is the meaningful comparison.
Do we have enough traffic to test?
Often not, for A/B testing. That does not stop the diagnostic work being valuable.
How long before we see change?
Fixing a broken form is immediate. Structural improvements show over weeks.