UX Research Report & Redesign Concept

Petal Supply Co. — the research behind the redesign.

Three months embedded as an accountant — not a researcher. Eight findings, five interviews, and a redesign that was never commissioned.

113years in business
5interviews conducted
8findings documented
5screens redesigned
Read the research

Full report

PDF UX Research Report
Fragmented truthNo single source of truth — inventory was an estimate, not a number
Analog ordering100+ item orders texted to a rep's personal phone
Invisible pricingA manual 2–2.5x markup, quotable only for the day it's asked
Cosmetic refreshThe "2011 upgrade" sat on 1980s infrastructure underneath
The redesign

Five screens, one connected system

Homepage

Homepage

One login for every vendor-licensed buyer — not four mismatched portals.

Public · Responsive
View live →
Buyer Portal

Buyer portal

Live catalog, transparent pricing, midnight fresh-stock incentive. Fully interactive.

Public · Responsive
View live →
Sales Rep

Sales rep mobile

Order entry with a live price calculator, replacing the bubble terminal.

Internal · Mobile only
View live →
Dispatch

Driver dispatch

Dispatcher map + driver mobile view, with a working delivery confirmation.

Internal · Desktop + Mobile
View live →
Accounting

Accounting dashboard

Auto-billing, house accounts, a searchable billing table.

Internal · Desktop only
View live →
Design System

Design system

Colors, typography, and the tokens behind every screen in this project.

Reference
View live →
Learnings

Proximity beats a mandate. It also has a limit.

Nobody asked, so nobody can un-ask.No commissioning client means no political cover and no obligation for anyone at the real company to act on this. The upside — total honesty — is also the ceiling on its impact.
Being embedded surfaces what interviews miss.The stretched landline cord, the hand-written catalog, reps colliding over the same phantom stock — nobody complained about these directly. Three months on the floor caught what a single interview round wouldn't have.
A real prototype still isn't a validated one.Five interactive screens prove the interaction design holds together. They don't prove a buyer would actually adopt this over a phone call to a rep she's ordered from for a decade. Those are different claims, and only one of them is answered here.
What's next

Built to inform. Not yet validated.

Every screen here runs on realistic, hand-placed data — not a connected backend. Checkout and camera capture say so directly instead of pretending to work. In priority order, here's what would close the gap between concept and something a real buyer could use:

01 — Fill in the competitor worksheet

Five B2B floral suppliers, six questions each, currently a template with no completed answers. This is the fastest, lowest-effort item on this list — and it's the one most likely to reveal whether "pricing hidden until checkout" is an industry norm worth calling out explicitly.

02 — Interview a buyer directly

Every other role in this project — reps, a manager, drivers — spoke for themselves. The buyer's journey map is still built from sales-staff accounts and observed behavior, not the buyer's own words.

03 — Wire the two core flows end to end

Buyer: discover → login → order → confirm. Rep: order comes in → check stock → enter → sync. Right now each screen works on its own; neither flow is clickable start to finish across screens yet.

04 — Kept: phone and text ordering stay

The obvious move would be forcing every order through the portal. It doesn't. Any real next step keeps this channel alive rather than treating it as legacy behavior to be designed away.