Problem
Members could see their score, but not why it changed
Members with scores of 640 or below out of 900 were leaving fastest. They already paid more to borrow, and found it harder to get credit or housing.
Score changes weren’t explained
“Tell me why my credit fell cause even I don’t know the reason.”
There was no clear next step
“I don’t see how your app is helping me increase my score other than tell me my credit score.”
Members didn’t know why they were declined
The records behind it were buried in the report, and Support had to send members on to TransUnion.
We couldn’t bring them back
Fewer turned on notifications, so we had no way to tell them when something changed.
Opportunity
The brief was to bring them back
Drive retention and engagement in Canada, focused on subprime members with scores of 640 or below, and give us new reasons to reach members by email and push.
What we measured
- Subprime members returning
- Churned members reactivated
- Email and push engagement
- Discovery and use of the feature
Research
For most members, credit was a black box
Members wanted to improve their credit, but few understood how it worked, younger adults least of all.
of Canadians were trying to improve their credit score
NerdWalletof Canadians understood their credit score
NerdWalletof active members used Credit Karma to improve their score
Member survey
Approach
Members didn’t need a bigger report. They needed to know where to look first.
Early concepts added more information to the existing report. Research pointed to something simpler, so I organised Credit Build around three jobs:
What changed?
Explain score movement in plain language.
What matters?
Bring the highest-impact credit factors to the front.
What can I do?
Connect each explanation to a useful next step.
That model shaped the hierarchy, the screen flow and the walkthrough below.
A method carried over
A three-second health check, read three ways
The first answer had to land in seconds. On BBC Weather, research showed people read a forecast three ways: the symbol, the number or the words. So every factor card answers all three ways.
BBC Weather
Credit Build
- VisualA colour and a status: Needs work, Going OK or Going great.Weather: the symbol
- NumberOne measure to track over time: 43% of your limit used.Weather: the temperature
- WordsOne plain sentence, with your own figures in bold.Weather: the summary
Design exploration
What I explored before anything shipped
The two things every member sees first, the score header and the factor card, were explored before we built them. Each set below names the designs, then shows the one that shipped and why.
Exploration 1
Score dial header experiments
Credit Build needed a way in from the home, and room to promote new features, without pushing the score down. The score was the most-used element on the page.
The business development slot was the top-performing placement on the home, so it had to stay in the same position.
Removing the dial on mobile web and native freed up space. That space became a second promo slot for Credit Build, the business development slot stayed exactly where it was, and the score stayed first.
Score header experiments
11 designs
Exploration 2
Factor card experiments
Members stopped coming back because they couldn’t see what mattered. Each card had to show, at a glance, how a factor was doing and how much it counted.
Mobile first, so compact: as many cards as possible above the fold. Every state had to work in English and French.
The shipped card grouped the factors by impact and named the status on each, so the legend could go. The colour went down the side because it was the most compact.
Factor card experiments
12 designs
What shipped
From a number to a next step
Credit Build answers the three questions in order: what changed, what matters, then what members can do about it.
What changed? · At a glance
Now the first screen says what changed
Members saw their score move with no reason, so the score and how it has moved come first.
What matters? · By impact
Now what matters most comes first
Members couldn’t tell which factors mattered, so they’re grouped by impact, each with a status, a number and one plain sentence.
5× more people clicked through to Credit Build
What matters? · The detail
Now every factor explains itself, in one place
Members had to search the whole report, so opening a factor shows what’s driving it: the missed payments, the account and what each one means.
64% of those who used it said it helped them understand their score
What can I do? · Next step
Now every factor ends with a next step
Members didn’t know where to start, so each factor ends with a Karma Pro Tip matched to their situation.
48% of those who found it helpful said they now knew what to do
What can I do? · Offers
Now members can check their chances before they apply
A declined application could cost members points, so hard inquiries point to Approval Odds and offers are matched to their credit.
Marketplace engagement rose 17%
The hard call
We put the worst news at the front
Derogatory marks, such as bankruptcies and court judgments, could decide whether a member was approved for credit, but they were buried in the full report and poorly explained. Members were declined without knowing why, then contacted Credit Karma for answers that sat with TransUnion.
I partnered with Product, Legal and Member Support to agree what we could show, how to explain it and where members should go next. Derogatory marks became their own factor, explained and linked to official help and dispute routes.
We expected more complaints. They fell 40%.
BeforeTwo-thirds of the way down the full report, below personal details, every account and closed accounts. No explanation and no next step.
AfterIts own factor, with the others. The tip says what the options are, links to official help, and leads to a dispute.
Launch
A launch designed to build momentum
Credit Karma had never launched a feature to members, and its emails looked like phishing. With Marketing, I designed the launch to build momentum: sign-ups before release, members moved into the app, a beta for feedback, then one factor at a time as a reason to come back.
Before
Our emails looked like phishing
Every email used the old brand and read like a fraud alert, so members didn’t trust them. The launch had to look like Credit Karma, and say something useful.
Week of 21 March
Goal: sign-ups before release
Early April
Goal: app downloads and feedback
Late April
Goal: return visits
Members with the appAlready came back more often
Sees “Launching soon” and taps I’m interested.
Goes straight to their factors, marked as beta.
One factor at a time, a reason to come back.
Web-only membersMoved into the app
Gets early access, and is told it’s coming to the app first.
Asked to download the app to see their factors.
Now in the app, gets the same emails as app members.
I defined the MVP: error states were designed but marked “not required for MVP”.
Results
Members found it useful, and the UK copied it
More members came back, complaints fell and members said it helped them understand their score. The UK later adopted the improvements.
extra returning members a month in April and May
fewer complaints about derogatory marks, each one avoided saving around £100
more people clicked through to Credit Build
Complaints
We expected more complaints. They fell by 40%.
Complaints about derogatory marks had been in our top three, around 70k a year. With the worst news now up front, we expected more. Instead they fell by 40%, and each one avoided saved around £100.
Member survey
Members were hungry to understand their reports
We surveyed 1,490 active members after launch. 66% of those trying to improve their score found it helpful. Adding offers didn’t cost revenue: monetisation stayed neutral, and the card offers entry point lifted marketplace engagement by 17% in May.
said improving their score was their most common goal
of those who’d used it said it helped them understand their score
of those who found it helpful said they now knew what to do
Reflection
What I learned
The most important design work was deciding what members needed first. Credit Build worked because it did not treat every piece of credit information equally. It prioritised change, impact and action.
The derogatory-marks work was the strongest lesson. Sometimes the best member experience is not softening difficult information. It is making it visible early enough for someone to do something about it.
Thanks to
Patrick Ross Content Designer and Fadi Dagher Head of Design.
