best overall — design interactive cohort, fall 2025

jolt

Learning to brew coffee at home is lonely and full of guesswork. Our team of four designed Jolt, an app where brewers learn from each other instead of chasing likes. It won best overall out of six teams.

4 min read

team3 designers + 1 pm
role
  • product designer
  • onboarding & posting
platformiOS app (figma prototype)
durationoct – dec 2025
01

overview

the two flows i designed: onboarding, and posting a brew
#1best overall, out of 6 teams
43brewers heard from: 33 surveys, 10 interviews
9testers across 2 rounds of usability testing
5features kept, out of three feature matrices
Eric Leproduct designerOnboarding, and the vision for posting a brew.
Kaitlyn Changproduct designerHome page and circles page.
Anjali Raviproduct designerGrowth page and account page.
Edyn Steplerproject managerProject management.
02

the problem

Brewers want to get better, and they learn best from people they know. Most picked up brewing from family, friends or social media, then got stuck with no clear idea what to adjust.

94%
are open to growing their coffee knowledge
73%
would most like to learn from friends and family
33 survey responses · designed for brewers aged 18–30

"People wanted to improve their brew knowledge through people in their community."

— the pattern across our 10 interviews

So Jolt is built on circles, not tools. From three feature matrices we kept five areas: home, learn, posts, saved and community.

three feature matrices: user value against effort
03

key decisions

no likes, just learning

finding — in round one, like counts and post analytics pushed testers toward chasing attention, and the feed started to feel like another influencer app.

  • Counts removed — no like totals and no post analytics in the final prototype.

  • Best advice rises — replies are ranked by the community, Reddit-style, so the most useful feedback sits on top.

  • Circles — every brewer gets a small group to learn with, matched to their level.

trade-off — likes are the easiest engagement loop to build. Without them, Jolt has to earn return visits with useful answers.

onboarding that ends in a community

finding — brewers learn from people, so an empty feed on day one is the worst first impression.

  • Account — Google, Apple or email. Password rules check off as you type (8+ characters, a number and a symbol, passwords match).

  • Location, explained — the permission prompt says why: “We only use your location to personalize your experience. No sharing, no tracking.”

  • Level before circles — New, Casual, Everyday Enthusiast or Professional. The next screen recommends circles for that level (e.g. Espresso Experts, Cold Brew Crew), with Select all.

  • Never trapped — a progress bar, Back and Skip on every step.

trade-off — asking for location and level adds two steps, so each one says what it’s for.

watch: experience level, then circles picked for that level

a post is a brew log, not a photo

finding — the useful part of a post is what you’d change next time, not how it looks.

  • Posted to a circle — “Posting to Morning Brew Club” sits at the top, so you know who will answer.

  • Rate your own brew — one to five bolts. The only score on a post is the brewer’s own.

  • Tags, then a story — Cold Brew, Homemade, New Brewer; then a title and “walk us through your process, technique, any tweaks”.

  • Safe exits — Save draft, a visibility setting, and Post drops you back in the feed while it uploads.

watch: rate, tag, add a photo, then Post returns you to the feed

growth as badges, not caffeine stats

finding — the mid-fi stats page (cups a day, caffeine, a “coffee quality score”) measured brewing, not learning.

  • Personal insights up top — earned badges, community helpers, suggested circles.

  • Badges tied to actions — Starting Sip (first brew), Weekly Grind (7-day streak), Insight Adopter (used feedback).

  • Next to unlock — locked badges show the next step, like Join a circle community.

growth: mid-fi stats vs. final badges

trade-off — badges are less precise than numbers, but they reward the behaviour Jolt is for: posting, helping, joining.

04

what testing changed

Two rounds, five testers then four. Every row below is a change we made between them.

feed“My Posts” became “Recommended For You”.Round one’s home showed only your own posts.changed
postsAdded a brew rating and the poster’s profile picture.Round one’s posts had neither.changed
commentsLarger type and a clear reply hierarchy.Round one’s comments were small and flat.changed
likesCounts and post analytics removed.They pulled attention away from learning.removed
home: round one vs. final
05

visual system

  • Coffee-brown actions — a dark brown palette for buttons and icons: warm, and high-contrast on the cream background.

  • One typeface — DM Sans, set as Figma text styles so all three designers’ screens match.

  • Rule of 4s — every spacing value is a multiple of 4, and every page shares the same margins.

design system
06

constraints & reflection

  • Scope — one cohort sprint and a Figma prototype; nothing was built or measured live.

  • Not yet designed — a café map, post sharing and direct messages.

  • Before scaling — dedicated usability and accessibility testing, and a revenue model (café partnerships, sponsorships).

The biggest call was taking likes out. It made the app harder to keep people coming back to, and much easier to learn in. Shifting from technical tracking to learning from each other turned a fiddly hobby into a shared one.