← Back to Work
UX Research · Service Design

Retail store representatives were bypassing the arrival kiosk. Customers were avoiding it.
That was the real research brief.

A major U.S. telecom retailer mandated a kiosk-first ID verification system across its stores. Store Customer support agents were quietly routing customers around it. Customers were hesitating, refusing, or walking straight up to a person instead. Understanding why both groups were doing this changed everything about how we framed the problem.

My Role
UX Research Consultant
Timeline
Nov 2024 – Jan 2025
Research Scope
20 Pilot Stores
Methods
Interviews · Observation · Shadowing
Industry
Telecom Retail
Read as
Telecom retail store interior with ID verification kiosk and queue

A mandatory ID kiosk was placed at the entrance of hundreds of U.S. telecom retail stores. The goal: speed up service by having customers self-identify before speaking to a customer support agent. What was actually happening: Customer support agents were quietly skipping the kiosk during busy hours — and customers were confused, hesitant, or refusing to use it entirely.

The problem wasn't the interface. It was trust.

Before & After — at a glance
Before
Original bare CHECK IN screen

No context. No reason. Just a demand to comply before anyone had explained why.

After
New entry screen: Let's Get You Helped Quickly and Securely

Check-in reframed as help, not enforcement. Customers given a choice from the first screen.

What the research revealed
🔄
Customer support agents bypassed the system
During peak hours, Customer support agents skipped the kiosk to keep queues moving. This was the first signal that the system didn't fit real-world behaviour.
🔒
Customers didn't feel safe
Entering personal data on a shared public device — with no explanation of why — created hesitation, avoidance, and drop-off.
🏛️
Three teams, three KPIs
Retail Ops measured throughput. Compliance measured accuracy. Product measured kiosk adoption. None measured customer trust. Nobody was aligned.
Core Insight
Identification is not a task.
It is a trust transaction.
What changed across 20 pilot stores
−50%
Queue wait time
Average peak wait dropped from 2 minutes to 1 minute
20
Pilot stores
Observation studies conducted across 20 stores; initial 2-store pilot was already underway when I joined
~100%
Self-completion rate
From 3–4 reluctant customers/day to almost every customer completing independently
💡
What I'd do differently: I would have included lower-literacy users and non-English speakers from week one. Accessibility gaps emerged late in the process — a gap I now build into research planning from the start.