Design Sprint

Sr. UX Designer & Sprint Facilitator·Amazon·1 week·Aug 2024

Nearly 500 upvotes. A clear feature request. A team ready to build. Everything said: build template sync so updates flow from templates to change requests (CRs). We almost did.

I flew to Sydney to run a 5-day design sprint with the engineering team. By Day 5, the feature was dead.

What we found

We put the prototype in front of users and discovered three groups who all wanted the opposite of what we were building.

Executors

“Fork it, finish it, forget it.”

Use templates to create CRs, then never look back. Don't want upstream changes disrupting an in-flight process.

Creators

“My template is the source of truth.”

See published templates as stable, vetted artifacts. Don't want to be responsible for disrupting active deployments.

The System

“Updating breaks trust.”

Certification is tied to a specific template version. Any update revokes it. Re-certification is slow and manual.

The catch-22

109 people asked for it. Nobody would use it.

Creator updates template

Fix error or improve process

Bar-raised status revoked

All downstream CRs lose certification

CRs can't execute

Teams blocked on re-certification

Re-certification is slow

Manual review by senior engineers

Nobody updates templates

Templates ossify to avoid disruption

cycle repeats

The feature was technically sound but behaviorally dead on arrival.

“In five days, we learned something that would have taken us months to discover after shipping.”

— Vineeth Varghese, Software Development Manager

Impact

What we invested
5 days

7 people, 1 flight, research, prototype, test

Week 1Month 6
What we avoided
3-6 months

4 engineers, build, ship, discover low adoption, pivot

Week 1Month 6

Prevented the wrong feature

The team paused work on template sync, avoiding 3-6 months of engineering investment in a feature users wouldn't adopt.

Validated the sprint format

One week of structured research prevented months of misdirected effort — especially valuable for teams without dedicated PMs.

Methodology adoption

Shared the sprint process across the org, inspiring adoption of design sprints for strategic planning.


Process

How we got there

Why a sprint

The engineering team was in Sydney. I was in Seattle. A 2-hour daily overlap made iterative collaboration nearly impossible. The team had no PM — they were self-prioritizing based on upvote counts. Without the sprint, the next step would have been engineering tickets, not research.

Designer

Seattle

PST (UTC-8)

12,800 km
Engineering

Sydney

AEDT (UTC+11)

2-hour overlapfor meetings

The sprint

Day 1Understand

Mapped the template lifecycle, interviewed experts, generated How Might We questions.

Day 2Ideate

Lightning demos, Crazy 8s, independent solution sketches.

Day 3Decide

Structured critique, dot-vote prioritization, storyboard for prototype.

Day 4Prototype

High-fidelity interactive Figma prototype.

Day 5Test

User feedback sessions. The twist.

Manly Beach, Sydney — after the sprint.