Overview
The real cost of fragmented context
A growth marketer's day looks something like this: check one platform for who's talking, another for campaign performance, a third for analytics, and a handful of tabs for competitors and community signals nobody bothered to centralise.
None of that was built to work together. So the marketer does the stitching — manually, every time — and by the time they've pulled enough context to feel confident, the moment they were trying to act on has usually passed.
I joined as Dopamyn's founding designer to close that gap, owning the product from the first sketch through what's shipping today.
Problem
Rejecting the obvious brief
The easy version of this project was "build a better campaign manager, with sharper analytics." I turned that down.
The actual problem wasn't that people struggled to launch campaigns.
It was that they didn't know who deserved their attention in the first place — and no amount of dashboard polish fixes that.
So I set one filter for every idea that came up in planning:
That filter is the reason Dopamyn looks nothing like a typical growth tool.
Solution
Once the problem was reframed, the structure followed naturally.
I mapped how a marketeer's thinking actually moves
Discover
Understand
Plan
Act
and used that sequence as the product's information architecture, instead of the usual research/campaigns/analytics split most tools default to.
Discover
Organise
Execute
A session starts with audience intelligence and AI-assisted insight,
Who is influential,
Where conversations are happening &
Why a particular audience is worth paying attention to.
As soon as something looks promising, it turns into strategy without a hard handoff
Similar audiences can be expanded automatically.
Lists become dynamic
rather than static.
Competitor research provides additional context, while AI helps users identify opportunities they may have otherwise overlooked.
With a strategy in place, the next step is execution.
Whether users want to
Launch creator campaigns
Reach out directly
Activate existing relationships,
The context built during discovery remains available throughout the process.
None of this is a dead end.
so the product compounds instead of resetting each time someone opens it.
Interfaces that get out of the way
I kept the visual language deliberately restrained: typography does most of the work, colour is used only when it signals something, and space exists to separate decisions rather than decorate a layout.
Predictability was the goal, not polish for its own sake.
Every time an interface has to be relearned, it's competing with the thing the user actually came to do.
Impact
Owned the full 0→1 process — research, IA, interaction design, visual system.
Replaced what would've been three disconnected tools with one continuous workflow.
Cut context-switching by keeping user intent intact across every stage of the growth process.
Left engineering a design system they now extend independently as new features ship.
