PROJECT 01　·　2020 – 2025　·　TAIPEI

One Ten

An independent restaurant in Taipei serving indigenous Taiwanese ingredients in a Western format. Two people ran all of it. I was co-founder, front of house, and everything that wasn't the kitchen.

This is a summary. The full nine-chapter case study is available in Chinese — use the 中文 link above.

677
original social posts, 5 years
15
international delegations hosted
25
documented collaborations
114
design source files
2
people

Figures taken from the website archive, official Meta data exports and a file-by-file design audit (Feb 2020 – Apr 2025). Not estimates.

SCOPE

What I owned

Kitchen was one person. Everything else was me.

Brand
Full identity system, designed in-house 114 source files: logo suite, 11 signage production drawings, 9 business card versions, menus reissued by season, coasters, stickers, aprons. Nothing outsourced.
Digital
Website, social, and the booking system 677 original posts across Facebook and Instagram over five years, plus 619 stories and 44 short-form videos — planned, shot, designed and written solo.
Press &
partnerships
Inbound only — no agency 13 press features and interviews, 25 collaborations with complete project files, 6 universities requesting research interviews.
Front
of house
Service, hosting, and introducing every dish Including 15 diplomatic delegations across four continents, each dish introduced in English by me, translated live by the accompanying officials.
Compliance
Green Restaurant certification, self-assessed The application required an ingredient-provenance disclosure sheet: every ingredient, its origin and supplier, logged and proportioned. I compiled it.
SYSTEM

We bought one first. It solved the wrong problem.

An off-the-shelf online booking system was in place from early on. It handled "how many people, what time" without trouble. What it had no concept of was that the food here had to be ordered in advance — without the order there is nothing to prep. So every booking turned into another email asking what they wanted, another check for a reply, another note about who hadn't finished. The tool wasn't broken; its assumptions just weren't ours.

In mid-2023 I built a replacement around that requirement using Fillout, Notion, n8n and the Gmail API, folding the repeated manual confirmation into the process itself. It is structured in four layers.

Forms
6 forms, 2 languages, 122 submissions through the main one The live opening calendar is embedded in the form itself, so closures are visible while booking. A separate form lets guests check their own reservation status instead of calling.
Logic
33 calculations, branching on party size and need Parties of six or more route to prepayment. Guests who want help ordering route to a follow-up path; guests who don't complete their order inline — because food is prepped to order, ordering has to close at booking time.
Database
One dataset, seven views Each view answers one question: not yet confirmed, needs help ordering, ordered, confirmed, calendar, board. A second table links each guest to every visit date — a return-visit history, not a booking log. It also joins in-store POS records and Google reviews against the same person.
Automation
Three reply types, drafted automatically Reads the record, expands the fields, evaluates conditions, and drafts the matching email. The workflow is numbered 1-2, because it is one branch of a larger set.

The guest table holds names, phone numbers and email addresses, so no screenshot of it appears anywhere in this portfolio.

Built 2023 – 2025, alongside running front of house.

CLOSED

May 2025

The restaurant stopped trading after four years and eight months at the same door. 114 design files, 15,112 photographs and videos, 677 posts and one working system came out of it.

Running a business at that size teaches something specific: with two people and no budget for headcount, the only way to absorb more work is to change the structure of the work itself. That is the same instinct I bring to systems implementation.

Back to
Portfolio