Portfolio Detail Page Redesign

Founding Product Designer

2025

The QUESTION

Dub is the first regulated copy-trading app in the US. Users pick a portfolio someone else has built and mirror its trades, dollar-for-dollar, in their own account—a real-money decision dressed as a tap.

The CEO asked me to redesign the discover page so users would view more portfolios. We knew from product analytics that users who view 5–6 portfolios are far more likely to convert to copying — but the existing card model showed only an All-Time return on each portfolio, and that wasn't doing enough work to earn the click.

The obvious answer was to add more stats to the discover card. I sketched it. It became a wall of text. Cards turned illegible faster than they got informative. So I stepped back and asked a different question: if the bottleneck wasn't really the discover card, what was it?

Treating discovery as a system, not a page.

The PREMISE

Discover gets the click. The portfolio detail page closes the comparison. They're two halves of the same job — and you can't optimize one without the other. I'd been briefed on the first half. The second half was where the breakdown actually lived.

I started looking at how shopping sites handle the same problem: high-volume browsing, fast comparison, low-stakes commitment. Three patterns kept showing up:

The product card on the listing page does the bare minimum to earn a tap. It doesn't try to be the product page.
The product page leads with social proof and a clear "what is this" — then exposes detail progressively. Specs are below the fold for the small subset of buyers who care about thread count.
Visual variety comes through a carousel of angles, not a wall of facts. Amazon shows you the product worn, in scale, from the back.

The WORK

Old PDP (left) led with a generic line chart; new PDP (right) leads with social proof and a smaller chart.

Move 1 — Promote social proof, demote the chart

The old PDP led with a generic line chart. The new PDP leads with three social-proof stats — Copiers, Copying volume, and Activity — and demotes the chart to supporting evidence below.

This isn't only a hierarchy change. It's editorial. By making the All-Time return not the largest number on the page, the design encodes a position: don't decide on past performance alone; trust the community signal. That's a values choice as much as a UX choice, and it aligns with where dub wants users to anchor their evaluation.

Move 2 — Restructure the body as a carousel of angles, not a wall of stats

The old PDP body used "Summary / Activity / Breakdown" tabs — categories built around what an owner of the asset wanted to inspect.

The new PDP uses "Performance / Holdings / Sectors" — categories built around what someone considering the portfolio needs to compare. Each tab is a different angle of the same product. The key is that users switch between angles, not see them all at once. Cognitive load stays low; comparison stays fast.

Move 3 — Push detail below the fold for the small audience that needs it

For the diversifiers — users who want to drill into the risk profile — the detail lives below the fold: Volatility, Market Sensitivity, Biggest Loss, Cash Income.

It's the design equivalent of "92% polyester, 8% cotton." Not the lede; not absent.

The OUTCOME

We ran a four-arm test starting January 2026. Funnel: Viewed Portfolio Details → Started Copy Portfolio → Copied Portfolio.

Headline result: the redesign converted 6.9% better than control, and it led at both steps of the funnel rather than only in the total — more of its users started a copy, and more of those who started went on to finish.

The more interesting result is the one between adding and subtracting. The redesign converted 15% ahead of the two additive variants pooled. Two independent attempts to put more in front of users, both finishing behind the version that took things away:

The presence of risk metrics, even below the fold, dragged hesitancy upward into the hero zone. Knowing that more numbers were waiting downstream apparently caused users to defer — and then not return.
Substituting a proprietary score for community signal in the hero hurt the most. Users transfer trust through community more readily than through algorithm. "412 people copying" is a lighter cognitive lift than "Alpha Score: 8.4."

The Alpha Score variant came out of leadership and the quant team. The experiment settled an internal debate that wouldn't have been settled by argument.

Worth saying plainly: this was a UX decision as much as a test outcome. The experiment told us which direction to move and settled an argument that wasn't going to be settled any other way — but the reduction-driven design was the right call before the numbers came in, and it would still have been the right call if they'd come back flat.

Home Activation v3 ›

J.CHEN

Brooklyn, NY
joyhelenachen@gmail.com

Set in Sorts Mill Goudy and Satoshi. Built in Framer, 2026. The logotype is borrowed from J.Crew, with affection.

Ask for the Soundtrack: "Joy," composed for the proprietor by Steven Mudd, 2026.

J.CHEN

Brooklyn, NY
joyhelenachen@gmail.com

Set in Sorts Mill Goudy and Satoshi. Built in Framer, 2026. The logotype is borrowed from J.Crew, with affection.

Ask for the Soundtrack: "Joy," composed for the proprietor by Steven Mudd, 2026.

J.CHEN

Brooklyn, NY
joyhelenachen@gmail.com

Set in Sorts Mill Goudy and Satoshi. Built in Framer, 2026. The logotype is borrowed from J.Crew, with affection.

Ask for the Soundtrack: "Joy," composed for the proprietor by Steven Mudd, 2026.