Conversion, compounded.
Turned the purchase funnel into a disciplined operating system: diagnose the leak, intervene with a controlled test, protect conversion, then scale the winner.
- +54.7%Homepage conversion
- +77%Plans-page conversion
- +10% AOVFlat conversion
02 / Operating system
Conversion operating loop
- 01Observe
Instrument homepage and plans-page drop-off.
- 02Test
Run one controlled hypothesis at a time.
- 03Guardrail
Protect conversion, order value, and quality.
- 04Scale
Ship the winner and compound the learning.
03 / Problem
The constraint behind the outcome.
The purchase journey had multiple conversion surfaces, competing hypotheses, and no single operating loop connecting diagnosis to scale.
04 / Decisions
The product choices that shaped the system.
- 01
Treat the funnel as a product system, not a queue of page requests.
- 02
Instrument the decision before designing the intervention.
- 03
Require a conversion and quality guardrail before scaling a winner.
05 / Tradeoffs
The boundaries that made it operable.
- Incremental tests improve attribution but slow broad redesign.
- A shared measurement spine adds up-front instrumentation work.
06 / Evidence
What the evidence can—and cannot—support.
The displayed results are owner-supplied portfolio outcomes and describe measured changes over the stated tenure; they do not attribute every point of lift to one experiment.
- Owner-reported outcomeHomepage conversion
Homepage conversion increased 54.7% over the measured tenure.
- Neel Jaiswal
- Owner-reported outcomePlans-page conversion
Plans-page conversion increased 77% over the measured tenure.
- Neel Jaiswal
- Owner-reported outcomeFlat conversion
An upsell increased average order value 10% while conversion remained flat.
- Neel Jaiswal