Skip to content

Case study 02 · Self-serve purchase funnel and analytics console

Buy Online: quote to checkout without a phone call 

Buy Online is the self-serve path from “what do you drive?” to a paid protection plan. I designed the purchase flows for auto and home, redesigning the quote step twice, and colleagues built most of them. Alongside the funnel, I designed and built most of an internal console, front end and back end, that shows the team conversion, drop-off and the performance of leads from the AI assistant.

Company
CoverageX
Role
Software engineer and product designer
Year
2025–2026
Disciplines
Design · Frontend · Backend · AI
Landing page with the choice of auto or home, beside a trust strip of claims paid and BBB, Trustpilot and Google ratings.
01 Landing page with the choice of auto or home, beside a trust strip of claims paid and BBB, Trustpilot and Google ratings.

01 The brief

The goal was to make buying a plan online as trustworthy as buying one by phone, and to let the team see where visitors left the funnel.

  • Self-servequote to checkout without a call
  • One consolefunnel, drop-off and A/B test management
02

Who buys, and what earns their trust

Buyers arrive choosing between two products, auto protection and home warranty, and each has its own funnel: six steps for auto and five for home. The landing page puts that choice first, beside a trust strip showing claims paid and ratings from BBB, Trustpilot and Google, because the visitor is deciding whether to trust the company before sharing any details.

03

One topic per screen

The quote step is where people decide, and I redesigned it twice. The version that shipped keeps each label visible once a field is filled, and ends with an explicit review step.

A persistent step rail shows progress throughout, and I also designed the error and session-expired screens.

02 Auto funnel at step 1 of 6, with a persistent step rail and one topic per screen.
04

A console that shows where buyers leave

I designed and built most of an internal console for the team. It shows conversion and drop-off for each step and how leads from the AI assistant perform, and it manages A/B tests.

An AI chat panel lets staff ask questions about the numbers in plain language.

05

Saves that cannot overwrite newer changes

The console saves with a version check. If a save is based on an out-of-date version, the server refuses it, and the console keeps the person’s edits and asks them to try again.

06

Resume by text message: a fix the data found

The drop-off view showed visitors leaving part-way through a quote. In response, I proposed and specified a way back that needs no account: a returning visitor enters their phone number, receives a one-time code by text message and picks up the saved quote.

The specification also covers what happens after purchase. Once a plan is bought, the “back to your quote” link stops working, so a purchased plan can never be quoted again.

07

How I work

The same process runs on every project. I own the design throughout; the build is a team effort in which I build much of the front end. This is what each stage looked like here.

In general

I review what exists and talk to the people closest to the customer. Then I write the brief, before opening Figma.

  • Brief
  • Audit
  • Journey map
Google AnalyticsLighthouseaxe DevTools

On this project

I mapped where the phone sales conversation lost people and where the web funnel lost their trust.

1 / 5
08

Highlights

  • Error and session-expired screens are designed as part of the funnel.
  • The AI assistant opens beside a step without leaving the funnel, and progress stays saved.
  • On a phone, the step rail collapses into a single progress line.
09

Outcome

Auto and home buyers can go from quote to a paid plan without a phone call. The team sees conversion and drop-off for each step. Resume by text message is proposed and specified.

10

What I’d do differently

I redesigned the quote step twice. Putting the first version in front of a few real buyers before building it would probably have saved one of those rounds. Next time I would build the drop-off view first, so the redesign has a baseline to beat.