Honest About Loss. Built to Convert.
Compliance mandate, turned into a calculator built to convert.
In deployment across Saxo's markets.
Client:
Saxo Bank
Role:
Lead Product Designer
Date:
2026
Context
I inherited a calculator built to help people understand what their money could actually become: the place where curiosity turns into the decision to press "Start saving" and open a saving account. Every change here carried weight beyond the chart.
How It Started
An AFM (Netherlands) mandate required investment tools to show realistic outcomes, not growth-only projections. The calculator only modeled gains. One regulation, one chart, and a visual language built entirely around going up now had to learn to tell the truth.
Before proposing anything, I reviewed the engineering handoff already locked for the positive-only version: spacing, breakpoints, panel widths already committed. I designed inside that constraint, not around it.
Four directions, three rejected
I tested a below-zero split, a deposit-consuming overlap, and a single fill with lines drawn over it, before landing on the final approach: two lines, Deposit and Total Savings, with the space between them filled to show the gain or loss directly. A later gradient treatment, pulled from an external reference, went through the same test and failed for the same reason: it looked good in one screenshot and broke down at real chart proportions.
The most technically honest of the rejected options showed loss as a visibly shrinking deposit region. I rejected it anyway. Stakeholders read a shrinking deposit as alarming, and an alarmed user doesn't convert. The two-line approach kept the loss just as visible, without making the deposit itself look like it was disappearing.
Loss and gain, encoded twice
The returns line switches between green and orange depending on sign, but the color is a headline, not the only source of truth. Every point on the chart carries its exact number on hover: deposit, net return, total return, broken out. Nobody has to infer a loss from a hue; the number says it plainly.
Color tells you the shape. The number tells you the truth.
A tooltip that never traps itself
The tooltip repositions based on which half of the chart is hovered, centers between the two compared values, never collides with the axis. A user modeling a loss is already anxious: if the number clips at the exact moment they're checking how bad it gets, I lose them. This wasn't polish. It was retention.
A slider that tells the truth slowly
I compressed the negative range of the return-rate slider into a small fraction of the track and gave the rest to positive scenarios. The full range stays one drag away; I wasn't hiding -10% to serve conversion. The interaction itself nudges exploration toward realistic-to-optimistic outcomes.
Compliance got the honest range. The product kept its optimism.
Delivery
Shipped as a stakeholder deck, developer handoff notes, and a feasibility brief, including an honest flag on the one tradeoff this introduces (uneven keyboard-step behavior). I'd rather hand over a known limitation than a surprise.
The Persuasion Layer
The calculator now tells the truth about loss. The next layer is designed to make that truth work harder, not soften it.
My hypothesis: a conditional messaging layer that reads the calculated outcome and responds with copy tailored to it. Positive projections close with an invitation: keep going, here's what waiting costs you. Negative ones close with a counterfactual: here's what the last five years would have looked like if you'd started. Untested, but the bet is that a truthful negative scenario converts better than a hidden one.
Compliance made the chart honest. This is designed to make the honesty do the selling.
Outcome
I went in thinking this was a compliance fix. I came out with a case for honesty and curiosity sharing a chart instead of fighting over it. The calculator's job was always to show someone where their money could actually go (good scenario or bad) clearly enough to press "Start saving." A regulation forced it to tell the truth about the bad scenario too. I made it do both, and I expect the persuasion layer to prove that telling the truth was the better invitation all along.
S E E O T H E R P R O J E C T S
