Conversion optimisation

Conversion optimisation is an experimental discipline. Without discipline it becomes a series of confidently wrong redesigns.

The order of operations

Instrument, then observe, then hypothesise, then change one thing, then measure. Skipping the first step is why so many optimisation programmes cannot tell you whether they worked.

Finding the real drop-off

Funnel step analysis, session recordings and form-field analytics together usually locate the problem within a week. The intuition of the team that built the page is the least reliable input, and it is almost always the first one offered.

High-yield areas, most of the time

• Form length and field order
• Message match between ad or search result and landing page
• Mobile layout: most abandonment is on phones
• Trust signals at the point of decision, not in the footer
• Error handling and validation messaging
• Page speed on the decision path specifically

Avoiding false wins

Short test windows, tiny sample sizes and seasonal effects manufacture wins that do not survive contact with the next month. Decide the sample size and duration before you start, and write the result down whether it was positive or not.