Warranty Keeper
A redesign of a cloud app for storing warranties and receipts, shipped on iOS and Android.
- Contribution
- Sole designer, research through UI
- Team
- Working directly with the founder
- Product area
- Consumer document storage
- Research sample
- 5 years of live users + 4 competing apps
- Delivery
- 6 months, iOS and Android
- Outcome
- Shipped
The problem
Five years live, and no clear order on any screen
Warranty Keeper had been running for five years with tens of thousands of people using it. They all came for the same thing: proof of purchase that would not be lost, damaged or unfindable on the day it mattered.
Three things stood between them and that. Uploading ran long with no sign of how many steps were left, so people gave up partway. Nothing on a screen had a clear order, so it was hard to tell one element from another. And too much of what someone needed sat several taps deeper than it had any reason to.
I created a semi-long title for my first item and notice that the texts runs into the “Back” menu option… there should be a character limit.
Play Store review
There were not many reviews to read at ten thousand downloads. The ones there were told me where people got stuck, and they were specific in a way I could act on straight away. That is the whole reason I started with them rather than with the founder’s feature list.
Final experience
The three features that carry the app
The three features that carry the app, running on the shipped build.
The brief
Same product, easier to finish
Give the app a clear navigation hierarchy and an intuitive experience, without changing what it does.
Two numbers to move: onboarding friction for a new customer, and taps for an existing one.
I had a goal of my own. I wanted to own every role on a product, research through to interface, and find out what I was good at.
Overall the app is great I would like to ask for two features. 1: being able to upload the photo at full-scale and not having to resize the photo. 2nd: having the ability to add a personal note on the warranty.
Play Store review
User personas
Four people, two of whom never finish
Four people behind the reviews and the funnel data. Two of them never finish a first upload.
The new arrival
Downloaded it today- Need
- To store one receipt and feel that the app worked, inside a couple of minutes.
- Constraint
- The upload gave no sense of how long it would take, so the funnel lost them mid-flow.
The long-term keeper
Years of documents stored- Need
- To pull one document out of a full account without remembering where they filed it.
- Constraint
- A fixed category list stops matching how anyone actually thinks about their own things.
The person mid-claim
Something broke this morning- Need
- The purchase date, the retailer and the document, on one screen, right now.
- Constraint
- That detail was buried several taps down, which is the worst possible moment to hide it.
The founder
One-person product decisions- Need
- Design calls he could argue with, backed by something other than taste.
- Constraint
- A five-year-old live product, so nothing could be redrawn from scratch.
What I designed
Cutting steps, and putting the deadline first
Solution 1 of 4
An upload that shows its own length
- Problem
- The upload ran long and said nothing about how much was left, so people abandoned it partway through.
- Design decision
- My first answer was to make the steps prettier. That was the wrong problem: the funnel showed people leaving, not disliking. So I cut the steps to the ones a claim genuinely needs, and made what remained visible from the first screen.
- Result
- Storing the first item is short enough to finish in one sitting.
Solution 2 of 4
Categories people already think in
- Problem
- A fixed category list works for the first ten documents and stops working somewhere after that.
- Design decision
- Let people name their own, and pair that with a search bar so an account that outgrows its own filing is still one query deep.
- Result
- A full account stays navigable without the app deciding how someone thinks.
Solution 3 of 4
Search, for the account that outgrew its filing
- Problem
- Categories work until an account has enough in it that the owner forgets which one they used.
- Design decision
- Search sits in the tab bar rather than behind a menu, and it returns the item card itself, expiry and all, rather than a row you then have to open.
- Result
- One query, and the document is on screen with its dates.
Solution 4 of 4
The expiry date, on the screen they open first
- Problem
- The one thing a warranty has that a receipt does not is a deadline, and it was sitting several taps down.
- Design decision
- Every item on the home list carries how long is left, as a bar and as a number, and expired items get their own group below. The reviewer who hit a title colliding with the Back button was describing the same failure from the other side: the screen had no order, so nothing could give way.
- Result
- You open the app and know what is about to run out.
The research behind it
What the funnel and the reviews said
I started with the in-app journey, the funnel and the behaviour patterns. I generated widely against what that turned up, then narrowed to what was worth testing with beta groups and A/B.
Two ideas had to land before any feature did: this is secure, and it is in the cloud. People hand this app the only copy of a receipt. The mood boards chased both, and the palette and components that came out of them carry that reassurance into every screen rather than into a badge.
User flows
The first upload, and the day it matters
First warranty stored The new arrival
- Install and sign up
- Onboarding, shortened
- Photograph the receipt
- Confirm the details
- Stored, and visible on the home screen
Find it under pressure The person mid-claim
- Open the app
- Search, or open a category
- Open the item
- Date, retailer and document on one screen
Where it landed
Shipped on both stores
onboarding friction for a new customer. The first of the brief’s two target numbers.
taps for an existing customer. The second target.
features carry the app by design: upload, search, custom categories.
stores it shipped to and runs on: App Store and Google Play.
The brief named two numbers to move. No post-launch measurement of either reached this write-up, so they are recorded as the targets they were rather than as results.
What this project changed
The app kept doing exactly what it did before and got easier to finish. Fewer steps to store something, a way to organise it that the owner chose, and the claim-critical details on the screen a person opens first.
Warranty Keeper does one thing, and every idea I had wanted a screen. There is more I would build around saving and organising documents. None of it is worth having if it pulls the app away from the reason somebody opened it.