114 local governments, one way to make the case
LogoWeb app

North Carolina's 100 counties and 14 municipalities are receiving opioid settlement money across eighteen years. The decisions about where it goes are not made by the commissioners who sign them off: they are made by working groups of coordinators, task force members and health department staff, who have to arrive with a recommendation they can defend.
FOCUS is the tool those groups work in, built for the UNC Center for Thriving Communities and funded by the William T. Grant Foundation. Five objectives, eighteen steps, and a working session that runs for months. I designed the identity and built the application. A funder's deadline shapes most of the university work I take.
Identity
A mark about the people, not the search
Other brands with "focus" in their name include a magnifying glass, but we wanted something different. FOCUS is not a search engine, it is a way to get the right voices into a room and keep them there, so the mark is points drawn around a center. The type is Sora, chosen from ten candidates judged on their capitals: credible to a public health audience without going neutral, and with enough character to be memorable. (And note the forward arrow hidden in the mark!)
Wayfinding
Start anywhere, leave, come back in March
The thing the client asked for above everything else. A working group might go straight through, or jump to the step their next meeting is about, or pick this up again after a two-month gap. One menu holds all five objectives, every step inside the one you are in, and where you currently are, so none of those three ways of working is the awkward one.

The hard part
A systems map, drawn by the people in the room
The center of the whole tool, and the piece that could not be a form. A group puts the populations, services and policies of its own county on a canvas, draws the arrows between them, and the tool reads the result back: where the hubs are, which loops feed themselves, what to carry into the next step. Undo, redo, tidy the layout, go full screen. It is a drawing tool living inside a guided process, and the output is structured enough for the next step to use.
Who is in the room
Fifty-two questions nobody enjoys being asked
Rate how involved each group currently is: commissioners, the health department, the school board, housing, the people with lived experience. The discomfort is the point. A group that rates law enforcement High and lived experience None has learned something about itself before it has spent a dollar.

Reading it back
The score keeps itself, in the margin
Beside the questions, a live map of which voices are in and which are missing, filling in as the group answers. An unrated row counts as absent, which is the honest default and the uncomfortable one. Under it, every section of the step with its own count.

The chain
Each step is built out of the one before it
Objective 2 opens holding the goal statement the group wrote in Objective 1 and the impact areas it marked, each labeled with where it came from. That is what makes this a framework rather than eighteen forms: the output of one step is the input of the next, and the group can see the join.

Deciding
Vote on it or talk it out, and the tool adapts
Some groups want every voice reflected in the outcome. Others need a defensible ranking to show commissioners. The step offers both, says which fits which situation, cites the SAMHSA guidance behind it, and reshapes the rest of the step around the choice. You can switch later.

Keeping things
A library that follows the group across the whole tool
Save a video, a guide or a dataset from any step and it lands in the county's own library, carrying the step it came from and a note the group can type on it. It travels across every objective, wherever it was found. Remove something and it stays where it came from, so nothing is ever actually lost.

Versioning
Named drafts, and a safety net under the restore
A group can save the state of a step under its own name, before the March steering committee or after it went to the county manager, and come back to it. Opening one replaces the current work, so the tool saves the current version first, automatically, as a restore point. Going back cannot cost you anything, which is the only reason people use a feature like this.

Saving
Nobody has to remember to save
Work is written as it is typed, and the header says so in plain words. For a group meeting monthly for a year, a lost session is not an inconvenience, it is a meeting that has to happen again.
Taking it out
Every page prints as something you can hand over
The recommendation eventually leaves the tool and goes into a meeting. So any step exports as a document with the navigation stripped, the interactive pieces redrawn as readable lists, and the funders credited at the foot. Not a screenshot of an app: a page that makes sense to somebody who has never opened it.

Under the hood
One account per county
A county signs in as itself. Personal logins would have recreated individual accounts without anyone deciding to, inside a tool whose unit of work is a committee. No county can see another's work.
Built for months, not a sitting
Progress, drafts and library all survive the gap between meetings, because the gap between meetings is where this tool actually lives.
Every save keeps the one before it
Section-level history behind the drafts, so a restore is a move somebody can make in a meeting.
No vendor lock on the way out
Standard Postgres, and auth that lives in the app. A research tool that outlives its grant has to be portable, and the exit here is a connection string.
Failures are visible
Errors report themselves with the context to act on, before a county has to mention that something did not work.
Reviewed inside the thing itself
The research team comments directly on the screens, in place, and threads get resolved and reopened there. Hundreds of them have gone through it.
Check out these other projects
Adrial is my go-to designer for anything my business needs … which is why I have used him for three different companies over the last 20+ years.
Craig WoodCEO, DefinianYour great work deserves to look great.
Twenty-five minutes. Tell me what you're trying to do and who needs to see it, and I'll tell you what I'd build first and roughly what it costs.




