Backhaul Direct, LLC
ShipDeez — End-to-End Shipping Platform
TL;DR
End-to-end UX for a shipping startup, from brand identity to live product in 10 business days.
ShipDeez — End-to-End Shipping Platform
Role: Lead Product Designer (UX, UI, Brand, Analytics, Webflow) Company: Backhaul Direct, LLC Timeline: 2-week sprint (10 business days) Tools: Figma, Webflow, EasyPost API, ShipBlink, Claude, GPT
The Problem
Small shippers consistently overpay for shipping. Without the volume that enterprise accounts have, a small business or casual shipper pays full retail rates — often 50–85% more than necessary — because they can't access negotiated carrier programs.
Backhaul Direct had the carrier relationships to change that. The plan: aggregate buying power across thousands of small shippers and pass the savings through a clean, consumer-friendly platform. That platform became ShipDeez.
The hard part wasn't the idea. It was the timeline: launch in 10 business days. The underlying platform contracts activated March 1, 2026 whether or not the product was designed.
The Constraints
Hard deadline. 10 business days. Contractual, not aspirational. The platform activates March 1 with or without a complete UX.
A three-platform stack. ShipDeez runs on EasyPost (carrier connectivity, label generation), ShipBlink (the merchant-facing white-label interface), and EasyPost Forge (branding customization). Every design decision had to work within what these platforms could actually do in v1 — not what I wanted them to do.
No custom backend. Designs had to operate within existing API capabilities. Server-side development was out of scope.
Two completely different users. A casual shipper who needs one label right now, and a regular shipper who needs a dashboard, saved addresses, and shipment history. These users have almost nothing in common.
Brand from scratch. No existing identity. Color, typography, tone, component library — all built alongside the product.
Wallet-based billing. EasyPost uses pre-funded wallets. Most shipping tools don't have to design for insufficient funds, failed wallet charges, or mid-payment abandonment. We did.
What I Was Responsible For
I was the sole designer on this project. I owned everything from strategy to delivery:
- UX architecture and user flows for both paths
- Wireframes and high-fidelity UI across all core screens
- Brand identity from scratch
- Design system and component library
- Landing page design and Webflow build
- Platform implementation mapping (which screen lives where in the stack)
- Developer handoff documentation
- Analytics instrumentation plan
I worked directly with the CEO, project lead, and technical lead across daily check-ins and weekly demos.
Research
Platform walkthroughs. I spent serious time inside ShipBlink's white-label screens, EasyPost's API docs, and Forge's customization limits before designing anything. You can't design within constraints you don't fully understand.
Competitive analysis. I mapped Pirate Ship, Shippo, and EasyShip to understand what's table-stakes and where there's room to differentiate. The main takeaway: most shipping tools are either confusing, ugly, or both. There's more room than you'd think.
Two personas. Working from the client's customer knowledge, I mapped two primary users:
The Guest Shipper — Ships once or rarely. No interest in creating an account. Wants the fastest path from "I have a package" to "I have a label." Cares most about price and speed.
The Regular Shipper — Small business owner or e-commerce seller. Needs saved addresses, shipment history, tracking dashboards, and potentially store connections. Cares about reliability, cost savings over time, and not having to re-enter the same address every time.
The Biggest Design Call: Two Separate Entry Points
The most important decision I made on this project: don't funnel both user types through the same path.
If a casual shipper hits a sign-up form before they've seen any value, most of them are gone. They came to print a label, not set up an account. Forcing that creates friction at exactly the wrong moment.
So I designed two clearly separated paths from the landing page itself:
- "Just Print Me a Label" — Guest path. No account required. Enter addresses and package details, compare rates, pay, download. Under 2 minutes start to finish.
- "Create My Account" — Full onboarding into a dashboard with wallet funding, saved addresses, history, and store connections.
The success screen for guest users includes a conversion prompt: "Ship Often? Save Even More. Create a Free Account." That's where we convert casual shippers into regulars — after they've already experienced the savings, not before.
Platform Mapping
Before designing screens, I mapped every interaction to its execution layer — which platform actually owns it:
- ShipDeez-owned UI: Landing page, guest quick-label flow, account creation, payment
- ShipBlink (via Forge white-label): Authenticated dashboard, shipment management, store connections
- EasyPost API: Rate retrieval, address verification, label generation, tracking
This mapping answered "who builds what?" for every screen before we got into detailed design. Without it, the dev team would have built things in the wrong system and we'd have lost days undoing it.
The Guest Flow
Quick Label Form
The most critical conversion screen. A casual shipper clicked "Just Print Me a Label" — now they need to enter package details without getting bogged down.
Key decisions:
- Single-page form, grouped clearly: From Address, To Address, Package Details
- Service Preference dropdown with a "Balanced" default to pre-filter results by speed vs. cost
- Optional fields are actually optional (company name, second address line)
- Full-width "Get Rates" CTA at the bottom so the next step is obvious
Rate Results
This is where ShipDeez's value proposition becomes tangible. Users see their actual discounted rates side by side.
- Card-based layout: carrier logo, service name, delivery estimate, price
- Lowest price pre-selected with a green "Selected" badge to reduce decision fatigue
- Promo code field for affiliate and marketing campaigns
- Persistent order summary so the running total is always visible
Payment
A first-time user entering card info on a brand they just discovered. Trust is the design problem here.
- Order summary sidebar showing exactly what they're paying for
- Explicit fraud protection callout: "Secured by AVS, CVC, and Stripe"
- 3D Secure note for high-risk transactions
- Guest checkouts capped at $500/day — sets expectations and limits fraud exposure
Success and Label Download
The payoff.
- Green checkmark, "You're All Set!" — warm and clear
- "Download Label (PDF)" is the dominant action
- Tracking info shown immediately: tracking number, carrier, estimated delivery
- The conversion moment at the bottom: "Ship Often? Save Even More." with a free account CTA
The Authenticated Flow
Create Account
Dashboard
For regular shippers, the dashboard is home base. Design decisions:
- Wallet balance front and center with "New Shipment" and "Add Funds" as primary actions
- Quick access to Account Settings and ShipBlink invite status
- Recent Shipments table with status pills: Delivered, In Transit, Pending
- Navigation organized around task frequency, not system architecture
Design System
Built from scratch alongside the product. Not after — alongside.
Core colors:
- Primary Orange (#C8641A) — CTAs, brand accent, interactive elements
- Brand Dark (#2B1408) — Hero backgrounds, navigation
- Cream (#F6E4C2) — Section backgrounds
- Surface White / Light Gray — Cards and page backgrounds
Semantic colors: Success green (#2FBF71), warning gold (#F3C969), error red (#E85B4F)
Type scale: H1 36/44 Bold → H2 24/32 Bold → Body 16/24 Regular → Label 14/20 Medium
Spacing: 4px base unit — scale: 4, 8, 12, 16, 24, 32, 48
One constraint worth noting: the system had to work across both ShipDeez-owned screens and ShipBlink's white-label environment. Not every token translated cleanly — some choices were deliberate compromises between ideal and feasible.
Component library: Buttons (primary, secondary, disabled), inputs with error states and validation messaging, selectable rate cards, status pills, shipment history table rows, navigation bar with wallet display, service preference dropdown.
AI in My Workflow
AI tools were part of the process from day one — not as a shortcut, but as a way to compress the time between idea and reviewable artifact.
Claude for architecture planning. I used Claude to map the platform integration — which screens live in ShipDeez-owned UI vs. ShipBlink vs. EasyPost. That mapping normally requires multiple rounds of back-and-forth with engineering. I showed up to the first stakeholder review with a draft the team could validate, not start from scratch.
GPT for edge case analysis. What happens when the wallet has insufficient funds? When address verification fails? When the user abandons mid-payment? GPT helped me stress-test flows for failure states before I'd designed the failure states themselves.
Claude for design system documentation. Initial token specs and component documentation were drafted with Claude, then refined for the actual developer handoff package.
Copy and microcopy. Landing page headlines, CTAs, error messages, empty states — all drafted with AI, refined for voice and brand tone.
Analytics taxonomy. The full funnel event tracking plan (20+ events from landing to label download) was architected with Claude to make sure there were no gaps.
What I didn't use AI for: the actual design judgment. Deciding what the product should be, what to cut, what users actually need — that's still the work.
Analytics Foundation
I delivered a complete instrumentation plan for the full funnel before handoff — not as an afterthought.
Funnel events: Landing page view → business/consumer selection → Ship Now CTA → signup start → signup complete → rate quote retrieved → label created → label purchased → label downloaded → tracking viewed
Nurture and brand events: Email opt-in, swag store clicks, storefront conversion, return visits
Implementation: Google Analytics 4 with custom event taxonomy, Microsoft Clarity for session recordings and heatmaps
The goal: drop-off visibility from day one, not bolted on later after the team realizes they can't answer basic funnel questions.
What Was Delivered in 10 Days
- End-to-end UX for both guest and authenticated journeys
- High-fidelity mockups for all core v1 screens
- Design system: color tokens, typography, spacing, full component library
- Live landing page built in Webflow
- Analytics instrumentation plan covering 20+ funnel events
- Platform implementation mapping for EasyPost / ShipBlink / Forge
- Wallet and failure-state UX rules document
- Developer handoff package
Key outcome: 85% average savings vs. retail shipping rates for platform users, based on negotiated carrier programs.
What I'd Do Differently
Test with real carrier data earlier. Rate comparison cards were designed with placeholder data. I couldn't validate the real UX of carrier response variation and actual rate formatting until post-handoff. That's a risk on the most important user-facing screen.
At least one usability test on the guest form. There are a lot of fields. Given more time, I'd run a session to measure completion rate and find friction before shipping.
Stepped form for mobile. The single-page form works on desktop but gets long on mobile. A stepped flow — address → package → rates — would likely improve mobile conversion. That's v2.
Reflection
ShipDeez required thinking like a product strategist and a systems designer, not just a UI designer. Every pixel had to live within a box defined by EasyPost's API capabilities, ShipBlink's white-label limits, and Forge's branding rules. Understanding those constraints before designing anything was what made the 10-day timeline survivable.
The clock also forced radical prioritization. Every decision got evaluated against one question: does this help someone get from landing page to first label? If not, it's out. That discipline is what produces a focused, shippable product instead of a spec document full of features that may never get built.
