Convoy

Redesigned last-mile package delivery after research proved users wouldn't pay to protect packages they barely think about.

Context UXDG 415 — Junior Studio II
Role Research Lead
Methods Semi-structured interviews, survey design & analysis, qualitative coding, affinity mapping, system flow mapping
Convoy ecosystem overview showing app, locker, and robot delivery

Problem Space

As package theft and loss becomes increasingly common, consumers are left without effective ways to protect their deliveries or prevent recurring losses. But our research — 109 survey respondents, 13 interview participants — revealed that people don't change their ordering behavior because of theft. Convenience wins. The deeper friction was unreliable tracking, inconsistent notifications, and a delivery system that gives recipients no control over when or how packages arrive.

Research Framing

The team assumed we'd design a home security solution like a smart lockbox or porch safe to prevent package theft. Our HMW framed the scope: how does package theft affect customers' experience ordering and receiving a product for delivery?

Our primary research pursued two questions: "What is the current package delivery process for delivery drivers and customers?" and "How do consumers experience package delivery and theft?" These questions guided our generative mixed-methods approach — building the frame we needed to understand not just whether theft happened, but whether it actually changed how people ordered, tracked, and received packages.

My Role & Constraints

My Role

Research Lead — wrote all interview and survey questions, conducted 3 of 13 user interviews, built the qualitative coding sheet (432 coded entries), synthesized findings into the five research themes that redirected the project, and designed the comprehensive system flowchart mapping all delivery paths.

Team

Brynn Jaratsonkit (Visual Lead), Giuseppe Mollo (Interaction & Industrial Design Lead), Max Clardie-Hoo (Project Manager). Task flows were collaborative and the UI screens and visual design were not mine.

Constraints

One-quarter timeline, no incentive budget. The quarter ended before we could run usability tests on the final designs.

Research Methods

Mixed-Methods Approach

I sequenced the methods deliberately: survey first, interviews second. The 17-question survey was designed to uncover customers' ordering frequency, theft experience, spending tolerance, and value-dependent behavior. The results surfaced patterns across 109 respondents and helped us understand what the problem looked like at scale.

But spending tolerance, behavior change, and intrinsic values data is difficult to surface in closed-ended questions. I selected semi-structured interviews specifically to allow flexibility for participants to provide deeper context into the issues they actually faced. The interview guide moved from warm-up ordering behavior through theft experience and response, into delivery notifications and tracking, and closed on trust, trade-offs, and willingness to pay.

Participant Demographics

109 survey respondents across age groups, housing types, and ordering frequencies.

Age Distribution

18–24
77
19–30
12
25–34
11
55–64
3
Other
6

Housing Type

Single Family
53
Apartment
38
Multi-Family
11
Townhouse
6

Ordering Frequency

1–3x / month
51
1–2x / week
23
< 1x / month
19
2x / month
7
3x+ / week
3

Willingness to Pay

Monthly Spending Tolerance

Nothing
4
Under $5
12
$5 – $10
12
$11 – $25
3
Multiple ranges
3

The survey asked respondents how much they'd spend per month to protect packages. The results were stark: the majority weren't willing to spend more than $10, and a significant group wouldn't spend anything at all. This data was central to redirecting the project away from a premium home-security product.

Delivery Priorities

Delivery Priority (Forced Choice)

0%
Cheapest
0%
Fastest
0%
Most Secure

When forced to choose between cheapest, fastest, or most secure delivery, respondents overwhelmingly chose cost over security — reinforcing that users wouldn't adopt a solution that added expense.

Interview Data Analysis

Across 432 coded entries, three thematic clusters emerged: 100-series (problems), 200-series (current experience), and 300-series (desired experience).

Problems (100-series)

176 entries

Inconsistencies
28
Security Concerns
19
Unreliable Notifications
15
Inconsistent Tracking
17
Liability / Stolen Pkgs
15
Driver Delays
13
Driver Ignores Instructions
12

Current Experience (200-series)

135 entries

Efficient Delivery
29
Risk ~ Package Value
14
Caution
12
Price Matters
12
Convenience
16
Safety Networks
8
Surveillance
8

Desired Experience (300-series)

105 entries

Up-to-date Tracking
26
F2F Interactions
17
Trust
16
Proof of Delivery
14
Customer Service
9
Adaptable Shipping
7
Scheduling Control
2
Key Finding

Every top desired-experience code pointed toward system-level transparency — not toward any product placed at the door.

Key Insight & Pivot

The picture that emerged from triangulating all three clusters revealed a home security product would fail to solve the actual pain points participants expressed. Package theft wasn't frequent, severe, or costly enough to justify added cost or complexity — users simply didn't value their packages enough to pay more to protect them.

The real problem was a broken delivery system that's inconsistent, unaccountable, and indifferent to the recipient. The driver research added another dimension: because drivers prioritize efficiency over careful handling, we needed a solution that worked with driver incentives rather than against them.

That reframe suggested a different design space entirely: rather than securing the home, we could redesign the delivery ecosystem itself. The infrastructure already existed in adjacent industries — Uber Eats and DoorDash had proven that real-time tracking and delivery windows were solvable at scale, and autonomous sidewalk robots were already navigating last-mile routes in LA. We weren't inventing new technology. We were taking the real-time tracking, scheduled delivery windows, and autonomous last-mile infrastructure proven by food delivery and pointing it to package delivery.

Outcome

Convoy app ecosystem showing delivery tracking and robot dispatch

The research redirected the project entirely.

What started as a home security concept became a connected delivery ecosystem because the data proved users didn't have a security problem — they had a system problem. The five findings I synthesized became the design brief: build around tracking transparency, secure-by-default delivery, customizable scheduling, and zero added cost to the user.

The final ecosystem — app-based delivery scheduling, neighborhood lockers, and 24-hour autonomous robot delivery — was presented at the end-of-quarter review, received well by both the class and professor, and nominated for submission to two design awards.

Delivery Ecosystem — System Flow

Full system flowchart mapping all delivery paths from hub to recipient.

Traditional Delivery
Assign to Driver Push: Delivery Scheduled Select Time Window Route Optimization Push: Driver En Route Deliver to Door Recipient Home?
Yes Sign / Accept Delivery Complete
No Photo Confirmation Push: Left at Door Delivery Complete
Locker Pickup
Route to Nearest Locker Scan & Deposit Package Push: Ready for Pickup SMS: Retrieval Code Travel to Locker Kiosk or App?
Kiosk Enter Code on Screen Locker Opens Package Retrieved
App Tap Unlock in App Bluetooth Unlock Package Retrieved
Robot Delivery
Dispatch Autonomous Robot Push: Robot Dispatched Track Live on Map Push: Robot Arriving Robot Arrives at Address User Present?
Yes Scan QR to Open Retrieve Package Delivery Complete
No Secure Hold Mode Push: Waiting for You Retry / Reroute

Final Design: The App

The Convoy app gives recipients control over their delivery experience — scheduling windows, choosing delivery methods, and real-time tracking across all carriers.

Delivery home screen
Delivery home
Compare delivery methods
Compare delivery methods
Choose robot delivery
Choose robot delivery

Final Design: The Locker

Neighborhood lockers provide secure-by-default delivery. The kiosk interface and NFC-enabled retrieval eliminate porch exposure entirely.

Locker home screen
Locker home
Enter retrieval code
Enter retrieval code
Kiosk home screen
Kiosk home screen

Reflection

The final concept depends on user trust in autonomous delivery robots that we never validated. I'd research that perception early, before the solution took shape, since a concept that hinges on an unproven behavioral assumption is fragile regardless of how strong the surrounding research is.

I'd also stratify recruitment: 77% of survey respondents were 18–24, and the "theft doesn't equal behavior change" pattern may not hold for older demographics who order more frequently and carry different theft risk profiles.

Finally, I'd protect time for at least one round of usability testing before the final presentation — the multi-path delivery model was never tested for actual comprehension.