Design Sprint Changelog
Four versions, four teams, one lineage — from GV to Zebra.
Design Sprint Changelog
Four versions, four teams, one lineage. Each generation learned from the last.
Last Client — Research Tech
3 sprints / January 2026
The first full Zebra engagement. Client: "I'm amazed. Unbelievable." "I can't wait to give it to users." The methodology held — the retrospective shaped everything that follows in these docs.
4.0 — Zebra Design Sprint
Charlie Ellington / Zebra Design — 2025
The design sprint rebuilt for an era where you ship code, not Figma.
If AI lets you build working software in days instead of months, what happens to the design sprint? It gets closer to the real thing. Sketches become unnecessary. Prototypes become production code. User testing becomes the actual differentiator.
What Changed from 3.0
Shipped code as default. Every sprint ends with working code in the client's codebase. Production-ready, not demo-ready.
Sketching workshops replaced with scope-locking workshops. 4+ hours of sketching compressed to 2.5 hours of alignment and scope lock. The insight from 50+ Hypersprints I facilitated at Deep Work: facilitator work got upvoted anyway. Client sketches added "small details, if any."
Scope Lock Workshop as standard. Workshop 1 aligns direction. The Scope Lock Workshop locks scope. These are different things. The 1-hour workshop prevents building in the wrong direction.
Code before consensus. We build working software, show it, iterate on reality — not imagination. Sprint 2 workshops happen on working code, not mockups.
Validation sprint in the 3-sprint transform. Build + Build + Validate instead of Build + Build + Build. Testing was the most valuable part of the first engagement — it deserves its own sprint.
Supervote for prioritisation. Straw poll, present, decider votes. Under 15 minutes. No endless debate.
Also standardised: mandatory recovery weeks (1 week after 2 sprints, 2 weeks after a transform), pre-sprint onboarding for all workshop attendees, and a sprint retrospective call one week post-delivery.
Why It Matters
Design and code are being commoditised. The sustainable differentiator is knowing what to build and proving it works. First engagement: 3 sprints, €15k, 55k lines of code, 173 components, 25+ screens. 9 user interviews. Visual design scored 9/10. Client: "I'm amazed. Unbelievable." "A bargain."
3.0 — The Deep Work Hypersprint
Charlie Ellington (co-founder) + Andrej Danko / Deep Work Studio — 2019
The design sprint that ships code, not just prototypes.
I co-founded Deep Work Studio in 2019 with Andrej Danko. We took the AJ&Smart model and asked: what if the sprint output was shipped, production-ready code in the client's codebase — not a throwaway prototype?
What Changed from 2.0
Implementation added to the sprint. Workshop days produced aligned direction, then implementation days produced real code. The client walked away with deployed software, not a Figma file.
The Hypersprint format. Workshop followed by intensive build, delivered to the client's production environment. I facilitated and led 50+ of these across Ethereum Foundation, Ramp Network, ConsenSys, dYdX, and dozens more. The structure created clear boundaries that protected both parties.
Validation as endpoint. The sprint didn't end when the design looked good. It ended when users validated it. Data decided, not opinion.
What It Proved
Deep Work scaled to 38+ designers, $800k peak revenue, and official recognition from the Ethereum Foundation. Shipped code delivery works as a productised process. Clients want working software, not Figma files they need to rebuild.
What It Taught
Over-optimisation kills. We built the team and structure the team wanted, not what the market required. When the market shifted, the company couldn't adapt. It became too expensive, too rigid. The company died.
The lesson Zebra carries forward: build an organism, not a machine.
2.0 — The AJ&Smart Design Sprint
Jonathan Courtney + team / AJ&Smart (Berlin) — ~2016
The design sprint made practical and accessible.
AJ&Smart took Jake Knapp's 5-day sprint and made it work as a commercial service. They condensed the format to 4 days, made facilitation itself the product, and proved the methodology could be a repeatable business.
What It Didn't Solve
Output was still Figma. The sprint produced aligned direction and tested prototypes — but the client still needed developers to build the real thing. The handoff gap remained.
1.0 — The Google Ventures Design Sprint
Jake Knapp / Google Ventures — 2010
The design sprint that started it all.
A 5-day process for answering critical business questions through design, prototyping, and testing with customers. Time-bounded design with a Decider role — one person has final authority, the sprint doesn't get stuck in consensus. The prototype gets tested on Friday with real users.
What It Left Open
Prototypes were throwaway. After the sprint, the team still needed to build the actual thing — which often took months and diverged from what was tested. The gap between "validated in a sprint" and "shipped to users" was where value leaked.
The Pattern
| Version | Solved | Left Open |
|---|---|---|
| 1.0 GV | Compressed validation into 5 days | Prototype was throwaway |
| 2.0 AJ&Smart | Made it a commercial service | Output was still Figma |
| 3.0 Deep Work | Shipped code to client codebase | Over-optimised the business model |
| 4.0 Zebra | AI-era execution + validation as differentiator | Distribution (the current challenge) |
Each generation gets closer to the real thing. Zebra 4.0 is where the methodology always wanted to go — one flow, shipped in code, validated with real users.
Updated: 10 February 2026 Next update: After Sprint 4