The website was already drawing healthy top-of-funnel traffic, so acquisition wasn't the constraint — but there was real ambiguity about what visitors actually did once they landed. The client didn't have a clear picture of where people lost momentum on the way to converting, particularly around the forms page that sat between arrival and conversion.
Working with an enterprise technology client, I analyzed website journeys using SQL in BigQuery, Python, and Excel — mapping how visitors moved through the site, where they dropped off, and how that varied across segments and traffic sources. That behavioral data, not assumption, was what pointed to the forms page as the highest-leverage place to intervene.
Rather than a single-variable tweak, I worked with design, product, and client stakeholders across multiple rounds to reconstruct the forms page from the ground up — aligning on expectations, iterating on design, and narrowing the field down to three distinct final designs worth testing head-to-head.
SQL and BigQuery for the underlying journey and funnel analysis, Python and Excel for data cleaning and reporting, and Google Analytics for tracking on-site behavior.
We ran a multivariate test across the three finalist designs simultaneously, rather than testing them one at a time, so all three could be compared against each other under the same traffic conditions. I analyzed the results, presented them to stakeholders, and identified a clear winning design.
The winning design increased primary-funnel conversions by up to 40% — turning a funnel that already had healthy traffic into one that converted a meaningfully higher share of it.
Even an already-successful pipeline can hide real optimization opportunities. Healthy top-of-funnel traffic doesn't guarantee an efficient path to conversion — mapping the full user journey is what surfaced the weak point here, and it's a lesson I carry into every funnel I look at now.