Apryse · Senior Product Designer · 2026

One purchase flow, every product

Every product in the suite sold through the same flow, which makes it one point of failure and one point of leverage. I rebuilt it, and the share of people who finished a purchase more than tripled.

Skip to the numbers

  • ~10%

    Purchase conversion, before

  • 32%

    Purchase conversion, after

  • 1 month

    Time to reach it

  • 1

    Flow, shared by the whole suite

Role
Senior Product Designer
Product
Purchase Flow
Duration
2026
Shipped
Shipped

Context

One flow, and the whole suite sells through it

Xodo is a suite of PDF products — desktop, mobile and web — and every one of them sold through the same purchase flow. That makes the flow a single point of failure and a single point of leverage: one bad step costs every product at once, and one good fix pays out everywhere. It is also trust-critical work. A purchase flow is where a product stops being free and starts asking for a card, and it gets one attempt to look like it deserves one.

The mess

People were not refusing to pay. They were refusing to sign up first.

The analytics said people wanted to buy. Clicks into the purchase journey were high and completions were not, and the distance between those two numbers was the entire problem.

Following the drop-off led to one step: account creation. People who had already decided to pay were stopping at the point where the product asked them to become a user first.

What I did

The design that shipped was the one people got through

Turned the analytics into a document three teams could argue with

Engineering, leadership and marketing all had a stake in this flow and none of them were looking at the same picture of it. I pulled together the analytics, the market analysis, the existing user flows and what we were actually trying to move, and used AI to organise that pile into one scoping document rather than four opinions.

That document is what got the three teams aligned before a single screen was designed. The wireframes came afterwards and came by hand — mid-fidelity, then high-fidelity, then a prototype stakeholders could click. By the time it reached approval there was very little left to disagree about, which is the only reason it moved as fast as it did.

Treated it as one system, not one flow per product

The suite shared a single purchase flow, so I designed it as shared infrastructure rather than as a feature of any one product. That cuts both ways: every improvement lands everywhere at once, and so does every mistake.

Prototyped the variations instead of arguing about them

I used Figma Make to build flow variations fast and put them in front of real users. Purchase flows attract opinion — everyone has a theory about the step that loses people — and prototypes settle that faster than a review meeting does. The design that shipped was the one people actually got through.

Evidence

It more than tripled — and here is what I cannot show you

Purchase conversion moved from about 10% to 32%, and it got there within a month.

Instrumentation available to me in this role: Firebase, Heap, HotJar, Sprig, StatSig.

What the number does not tell you

  • I have not published the sample size, the split, or the test design behind that figure.
  • I measured a month, so I claim a month. I am not claiming 32% held beyond it.
  • I designed the flow and I claim the flow. I am not claiming it was the only thing that changed in the product that month.

What shipped

It replaced the old flow across the suite

The streamlined flow replaced the old one across the Xodo suite, and it more than tripled the rate at which people who started a purchase finished one. It is the largest number on this site. What is still open is durability: a month is enough to know something changed and not enough to know it stuck.