Funnel work pays before channel work because you cannot judge a channel through a broken funnel. The highest-leverage fixes are usually structural: the order of questions in a form, one pain per landing page, moving a funnel from an app store to the web.
These articles show the fixes and the lifts behind them, including the two tests that produced $2 billion in incremental loan volume at Rocket Mortgage.
Start here
Conversion & funnel · July 14, 2026 · 4 min read
Two experiments at Rocket Mortgage lifted close rates 12% and 18.8% by reordering questions. How to design form tests that find real friction.
Read the article →
Conversion & funnel · August 15, 2026 · 8 min read
A three-part landing page framework: the customer's pain, the ad theme it continues, the feature that answers it. Plus the hypothesis you write first.
Read the article →
Conversion & funnel · July 15, 2026 · 4 min read
Apple and Google take up to 30% of in-app subscriptions and hide your conversion data. How YouMail moved signups to the web and cut CAC 77%.
Read the article →All articles on conversion & funnel
Conversion & funnel · August 6, 2026 · 8 min read
No customer access? Sales and support hear objections all day. The structured async method that turned their notes into voice-of-customer research.
Read the article →
Conversion & funnel · August 5, 2026 · 9 min read
A lender replaced its contact form with a branching quiz that qualifies investors and brokers without hurting conversion. The build, questions and results.
Read the article →
Case study · July 28, 2026 · 6 min read
One Rocket Mortgage upsell tile, tested three times: copy first, then a trust badge. Clicks tripled, adding an estimated 150 loans a month. Full breakdown.
Read the case study →
Conversion & funnel · July 26, 2026 · 8 min read
A platform shipped your core feature for free. The calm playbook: treat it as validation, find the depth gap, and reframe where the problem gets solved.
Read the article →The step with the largest drop closest to revenue. Fixing the top of the funnel pushes more people into the leak below it.
Yes. Reordering the questions in Rocket Mortgage's application raised completion 4 percent and closed loans 12 percent, worth an estimated $1.5 billion in loan volume.
One. One pain, one proposition, one hypothesis you wrote down before the test.
I write about what actually moved the numbers. If any of it maps to what you're facing, let's talk.
Book an intro call