Improving conversion by 82%

The problem

Only 10% of users reached the valuation step, with drop-off concentrated early in the funnel. Analytics, session recordings, interviews, and usability testing pointed to three consistent issues:

• High effort upfront. Users were asked for detailed technical inputs before seeing any value.
• Low trust in pricing. Users were unsure how valuations worked or whether they’d get a fair deal.
• Cognitive overload. The flow felt complex and the next step was rarely clear.

Approach

Working with product, engineering, and data, I translated those findings into testable hypotheses:

• Reduce effort to increase completion
• Increase trust to improve continuation
• Show value earlier to improve activation

Each hypothesis was tied to funnel metrics and tested through A/B and multivariate experiments, iterated weekly.

Experimentation

One experiment illustrates the approach well. The control showed a valuation with minimal context. Variant 1 added a comparison table showing Carwow versus a private sale. Variant 2 introduced social proof through ratings and reviews. Variants with trust signals consistently outperformed the control, confirming that confidence in pricing was a real conversion lever.

Design solutions

Reduced friction in car entry. Simplified language, reduced complexity, and introduced progressive disclosure and smart defaults. Users completed the early steps at a higher rate.

Reframed the entry point. Explained the process upfront and communicated value before asking anything of the user. More users chose to start the journey.

Built trust throughout. Social proof, pricing explanations, and contextual tooltips addressed anxiety at the moments it surfaced most.

Results

Conversion moved from 10% to 18.2%, an 82% uplift in one quarter, resulting in thousands of additional completed sales.

The clearest lesson: clarity and trust are conversion levers just as much as flow or copy. The speed of iteration, weekly experiments and incremental improvements, compounded quickly.