Case Study: Wake - Annotations
Wake lets design teams share in-progress work at full fidelity and discuss it in a live stream. Teams commonly connect Wake to Slack and Sketch, so feedback flows through the same stream everyone already watches.Wake lets design teams share in-progress work at full fidelity and discuss it in a live stream. Teams commonly connect Wake to Slack and Sketch, so feedback flows through the same stream everyone already watches.
What I Did
Product Leadership
UI/UX
Research & Interviews
0→1
5x
Fewer Duplicate Comments
70%
Higher Clarity in Communication
12hr
Time to First Actionable Comment
Annotations turned Wake’s running design stream into a place where feedback could be precise, fast, and easy to act on. I led the end-to-end product design for on-canvas comments that link directly into Wake’s conversation stream. The work focused on removing ambiguity in reviews, reducing duplicate feedback, and shortening the time from post to resolution.
Beyond shipping a feature, this project was about strengthening a core promise of Wake: that sharing in-progress work should feel lightweight, and that feedback should help teams move forward instead of getting stuck in long, ambiguous threads. Annotations needed to feel native to how teams already used Wake, while quietly raising the quality and traceability of every design conversation.
Problem
Wake’s stream captured the story of a project but not always where feedback applied on the canvas.
This led to:
Vague comments that required follow‑up (“this nav,” “the blue button”)
Designers exporting circled screenshots just to clarify a single point
Duplicate feedback on the same UI element
Teams bouncing between Wake and other tools that did support on‑canvas comments, fracturing the single-stream experience
Design challenge: Make feedback unambiguous and spatially anchored, while keeping the stream as the single source of truth.
Goal
Bring precise, on-canvas feedback into the existing Wake stream so teams keep context without introducing a new mental model or extra workflow.
Make feedback unambiguous. A reader should know exactly what a comment refers to.
Preserve the narrative. Annotations must show up in the same stream so teams can skim chronologically, search, and share to Slack like they already do.
Reduce review time. Shorten time to first actionable comment and time to resolution.
Avoid visual noise. Keep mocks readable at any density and any zoom level.
Ship on Mac app, responsive web, and iOS with consistent behavior.
Approach
I combined qualitative research with product and technical constraints to shape the solution.
Research and discovery:
Interviewed designers, PMs, and engineers on 6 Wake‑heavy teams
Observed real critiques and async reviews using Wake
Collected lightweight diary notes on when feedback felt slow or brittle
Key insights:
Feedback needed to live where eyes already were: on the image/artboard.
Teams relied on the stream as the shared narrative and memory.
People were wary of clutter; pins had to scale to dense mocks.
Constraints:
Feedback must remain in the existing stream (no separate “review mode” tool).
Performance had to stay strong with large images and long threads.
Core behavior needed to work on Mac, web, and iOS.
Annotations had to be discoverable to first‑time reviewers (often arriving from Slack).
Existing posts and comments had to keep working with no migration.
Solution
Concept: Pins as First‑Class Stream Objects
Annotations became on‑canvas pins tightly linked to the conversation stream:
Create: Click/tap on the canvas to drop a pin, type a comment, hit return.
Link: Wake creates a corresponding stream entry with a thumbnail, author, and deep link.
Jump: Clicking the stream entry smoothly pans/zooms to the pinned area, highlights the pin, and opens the thread.
Resolve: Threads can be resolved; pins collapse into a subtle state to reduce clutter while preserving history.
This kept two complementary entry points into the same object:
People who live in the stream keep skimming posts chronologically.
People focused on a screen work directly off pins on the canvas.
Interaction Details
To balance clarity and readability:
Pin states: Idle, hover, focused, resolved. Idle pins are quiet and minimal. Only the focused pin expands to show text.
Density handling: Overlapping pins group with a count badge; zooming or stepping via keyboard disambiguates dense regions.
Keyboard and accessibility: J/K or arrow keys move between unresolved pins; logical tab order and focus rings support keyboard‑only review.
Deep links: From the stream, Slack, and notifications, links land reviewers on the exact pinned region.
Design Tradeoffs and Exploration
Marker style: Tried avatars, numbers, and simple dots. Avatars gave social context but felt heavy on busy mocks. Numbers were great for linear walkthroughs but failed with asynchronous threads. Final: lightweight pins with optional count badges; avatars remain in hover tooltips.
Motion: Hard jumps to a pin felt disorienting. Switched to short, eased pan and zoom for spatial orientation.
Thread placement: Tested a floating panel anchored to the canvas; it hid content on smaller screens and fought the “single stream” model. Final: keep conversation in the main stream, with highlight and inline interaction on the canvas.
Discoverability: Combined an explicit “Add annotation” affordance for new users with direct click‑on‑canvas for power users.
Outcomes
Annotations strengthened Wake’s core promise: lightweight sharing with actionable, trustworthy feedback.
Impact observed in pilots with internal teams and early customers:
Clearer, faster reviews
Shorter time to first actionable comment on new design posts, and fewer duplicate comments on the same region.
Fewer clarification loops
Designers reported fewer “What did you mean here?” threads and less need for marked‑up screenshots.
Higher engagement without extra tools
Because annotations lived in the existing stream and propagated to Slack, more teammates engaged in critique without adopting a new tool.
Stronger accountability
Each comment was anchored to a specific region and resolvable thread, making design decisions easier to trace.
Next Opportunities
If I extended this work, I’d explore:
Batch triage: Tools to select multiple pins and assign, label, or resolve them in one sweep, especially for design leads reviewing large critiques.
Smarter anchors for responsive layouts: Better heuristics (beyond coordinates) to keep pins meaningful as artboards resize or layouts reflow.
Review analytics: Heatmaps and alerts for clusters of unresolved feedback to highlight risk areas before handoff.