Healthy
Together

MMXX–MMXXIV

A farmer’s screen in the app, titled MD FMNP Authorized Farmer, with buttons to start a transaction, view history and check a balance

From paper checks to digital benefits at Maryland’s farmers markets, in half the quoted time

Summer to fall 2021

Maryland’s farmers market benefits ran on paper checks, and the state needed a digital service for residents, farmers and market managers in time for the season. I designed the mobile experience and the design system under it, and changed the product wherever real use proved us wrong: manual entry for seniors who struggled with QR codes, and sales that work with no signal. We delivered everything the RFP asked for in half the quoted time, and the same system later became two more state programs with little more than a rebrand.

12weeks for the full scope, against 24 quoted
18markets in week one, against a target of 10
8market manager offices, against 3 expected
2more state programs on the same system
Role
Principal Product Designer. Design lead for the launch and owner of the mobile experience
Program
Maryland’s Farmers Market Nutrition Program: produce benefits for WIC participants and seniors
Team
3 designers (me on mobile, one on admin, one on marketing and sales); 5 iOS, 3 Android, 3 backend engineers; 2 QA; the state’s program partners
Outcome
Full RFP scope in 12 weeks against 24 quoted. 18 markets in week one against a target of 10

The start

Benefits on paper, and one market season to change it

Maryland’s farmers market benefits ran on paper. Residents got checks and coupons, farmers took them at their stalls, and market managers reconciled everything by hand.

That meant slow processing, a lot of administrative work, and benefits that went unspent. The state needed a digital service for all three groups in time for the 2021 market season.

The RFP quoted 24 weeks.

Who was involved

Me
Principal Product Designer and design lead for the launch. I owned the mobile experience for residents and farmers on iOS, Android and web, and the design system the whole platform was built on.
Two designers
One on desktop and admin, one on marketing and sales.
Sales and the RFP writers
We shaped the engagement together before it existed.
Product, engineering and QA
Product, 5 iOS, 3 Android and 3 backend engineers, and 2 in QA.
Maryland’s program partners
Their feedback on clickable journeys shaped the scope before engineering started.

The service

Three people have to succeed for one payment to work

Residents

Seniors and WIC participants, with very mixed confidence with phones.

They needed to see their balance and pay without learning a complicated system.

Farmers

Selling outdoors, often with little digital experience.

They needed to take a payment and know they’d been paid.

Market managers

They issued benefits and cards and supported participants.

They reconciled every transaction.

  1. Issue

    A market manager issues benefits and cards.

  2. Access

    A resident sees what they have to spend.

  3. Pay

    The resident pays a farmer at the stall.

  4. Reconcile

    Every sale is matched back to the program.

The process

How it unfolded

  1. Before the contract

    With the RFP writers and sales, I used production-ready work and fast mockups to make the service concrete and shape what we committed to.

  2. The service, mapped

    Once we won it, I mapped the whole service, from issuing benefits to reconciling them.

  3. Clickable journeys

    For residents and administrators, so the state’s feedback could shape the scope before engineering started.

  4. One design system

    Components, tokens and flows for every group and every platform.

  5. Three tracks at once

    The work for farmers, residents and market managers overlapped instead of running one after another.

  6. In the field

    We tested with real seniors and farmers, and changed the product where it failed them.

  7. Weekly walkthroughs

    Design and engineering went through the builds together every week.

  8. Launch

    The 2021 season. 18 markets were live in the first week.

Decision 1

Design the speed in

The RFP quoted 24 weeks. We delivered all of it in 12, and we didn’t cut anything to get there.

The speed came from how the platform was set up, and above all from the design system I built: one set of components, tokens and flows that gave residents, farmers and market managers each their own experience.

Engineering reused patterns instead of waiting for new screens, and the work for each group ran at the same time.

  • iOS
  • Android
  • Web
  • Phone
  • Desktop
Quoted
24 weeks
Delivered
12 weeks

One system, three experiences

Decision 2

Manual entry, because seniors struggled with QR codes

The ideal flow was scan and pay. In the field, scanning was exactly where people got stuck.

We tested the app with real seniors and farmers, and scanning a QR code was where they struggled. Other users told us the same.

So residents got manual entry alongside the QR code, with large tap targets and the balance always in view.

A resident’s benefits card in the app: a QR code with its number under it, the remaining balance of $30.00, the expiry date and the last transactions
  1. PlannedScan a QR code to pay
  2. ShippedScan a QR code, or enter it by hand

Decision 3

A sale has to work with no signal

Farmers markets are outdoors, often with a poor connection. A design that had to be online would have failed at the stall.

Farmers could complete a sale and see balances with no signal. The app reported back when service returned, so every transaction reconciled.

The farmer’s flow has one clear action per step and says plainly what happened, for outdoor use and limited dexterity.

At the market
The app’s scan screen: the camera pointed at a benefits card, with a frame around its QR code

Working together

In it until it shipped

Discovery, prototyping and engineering ran at the same time, so I stayed in the work through QA and the pilot.

  1. Walk through

    Every week, design and engineering went through the builds together.

  2. Review

    Behaviour, accessibility and edge cases.

  3. Rank

    Pilot feedback, ordered by its effect on users.

  4. Ship

    Into the next iteration.

Result

Everything the RFP asked for, in half the time

“I’m going to scan it. You have $33 left.”A farmer, at a first digital sale

It was the first time these markets went fully digital.

One farmer over 70, who had resisted digital tools, processed more than $400 in benefits in her first week.

The 2021 season

41,617participants
$999,780in benefits issued
$401,524redeemed

What it unlocked

The same system became two more state programs

When Oklahoma and Missouri came on, most of the design work was already done. The system had the pieces, so each new program was mainly a rebrand.

  1. MarylandFarmers Market Nutrition ProgramBuilt here first
  2. OklahomaWICMainly a rebrand
  3. MissouriSEBTMainly a rebrand
Looking back

What went right, and what went wrong

Four things that worked and four that didn’t.

Went right

  1. The design system came first

    One set of components, tokens and flows meant engineering reused patterns instead of waiting for new screens.

  2. Clickable journeys before engineering

    The state could react to something real, so the scope changed while changing it was still cheap.

  3. Three tracks at once

    The work for farmers, residents and market managers overlapped. That is where half the quoted time went.

  4. Weekly walkthroughs

    Design and engineering reviewed behaviour, accessibility and edge cases together, every week, through the pilot.

Went wrong

  1. We assumed everyone could scan a QR code

    Seniors struggled with it in field testing, and other users reported the same.

    What I changedManual entry alongside the QR code, with large tap targets.

  2. We assumed there would be signal

    Farmers markets are outdoors, often with a poor connection.

    What I changedThe whole sale works offline and reports back when service returns.

  3. We found the daylight problem late

    Some colours didn’t hold up in bright sun at a market stall, and dark mode came late too. We caught both in QA.

    What I’d changeTest outdoors from the first week. It would have been cheaper than fixing it in QA.

  4. I didn’t sit with users enough

    What we learned in the field changed the product every time.

    What I’d changeMore time directly with seniors and farmers, earlier.

For your team

What I’d bring to your team

I set the system up first, test where the product is really used, and stay in it until it ships.

See all my workNext: Topology Eyewear

  1. 1A design system that lets engineering move without waiting
  2. 2Scope agreed on clickable journeys before anyone builds
  3. 3Testing where people really use the product