Smile White had three beautiful new landing pages ready to go live. We tested them first. All three lost, and that is the win.
A redesign that would have halved bookings, stopped before launchSmile White had three new landing pages built to replace their main clear aligners page. All three were full redesigns, all three looked the part, and the plan was to pick a winner and roll it out. With the new testing discipline in the business we ran them as a split test against the existing page.
The strongest of the new pages converted 51% lower than the page it was built to replace, and the result was conclusive at real scale.
Had they shipped on instinct, which they would have done before we started working together, they would have roughly halved their bookings. That test cost them a fortnight. Launching blind would have cost them nearly half a year of demand.
The new pages drove considerably more engagement and more visits to the booking page, and still produced fewer bookings. That drove us to where the real problem was, and it was not the landing page.
It was the final step of the booking questionnaire, where genuinely interested people were giving up.
A reworked booking hero with sale-led copy beat the old version on bookings, 13.1% against 12.5%. That sat at 85% confidence rather than the 95% bar we hold ourselves to, so we called it directional, kept the better version live, and did not dress it up.
A testimonials test came back honestly inconclusive. The effect was real but small, and the test was never powerful enough to prove a change that size.
We learned, folded that learning into a bigger, braver test. That one is running now.
Optimisation is more than landing pages and split tests. Sometimes the change you feel most proud of is making the whole site kinder.
Try to leave and this jumped in. The dismiss link, “No thanks, I'd rather pay full price”, is textbook confirmshaming, aimed at people still trying to understand a treatment that takes weeks to choose.
Nobody clicks it feeling persuaded, and the VIP discount it dangled was never really explained, so a visitor could not tell whether it meant five pounds or a hundred.
We made the case against it. The client agreed and took it down, Bookings ticked up in the fortnight after, though with other tests live we can't pin that on the popup alone.
“If you're looking for someone to sharpen your paid landers and make CRO work that compounds, Jon is the person you want on your side.”
Joe Whiteman, PPC Manager, Smile White and Yorkshire Dental Suite
Test it before it goes live. Better to find out on a slice of traffic than on all of it.
Get in touch