ProblemPeople who want company and good food, and no easy way to find both at once BuiltAn app where a home cook hosts a dinner and a stranger books a seat

Why

Two years of eating alone, and a restaurant that doesn’t fix it.

The pandemic made eating a solitary habit. The want for company came back before the habit broke.

We surveyed 55 people in Los Angeles about how they eat and who with. The pattern was individualism by default: delivery, desk lunches, dinners alone, and underneath it a wish for the table. A restaurant sells food, not company. What was missing was the dinner party, for people who don’t know anyone to invite. So the brief became: make the home kitchen a place a stranger can book.

Two people, one table

Jeremy has cooked in a fine-dining kitchen for nine years and wants to open a French-Korean place with a communal feel, but everyone he knows is too polite to give him real feedback. Joe is new in town, just lost a job and a girlfriend, and his favourite Korean restaurant closed. One has a kitchen; the other has an appetite.

How it was made

One semester, from a survey to a prototype people could tap.

1. ISurvey 55 peopleHow they eat, who with, and what they thought the app was for when shown lo-fi mockups.
2. IITwo personasJeremy the host and Joe the guest: the two sides of one dinner, written from the survey.
3. IIIStoryboards and flowsJoe finds Jeremy’s dinner; Jeremy sets the menu. Two flows, drawn before any screen.
4. IVWireframesEvery screen in grey boxes, so the team argued about the order of things before the look.
5. VThe logo and the design systemA mark, a green, type and components, so three designers produced one app.
6. VIPrototype and testA clickable Figma prototype, shown to the survey group again; the feed and the follow came out of their feedback.

The design

One green, a glass in a speech bubble, and the food doing the talking.

The brand stays out of the way of the photographs: a single green for the buttons and the mark, black and white for everything else, and the hosts’ own pictures of their dishes as the colour. The logo is a cocktail glass inside a speech bubble: dinner and conversation in one shape.

• Green#097A57 · The brand, the buttons, the ratings
• Ink#1C1C1E · Text
• Paper#FFFFFF · The screens, so the food photographs carry the colour

The app

What a host does, and what a guest does.

1 2 3

01The host writes the dinner.

1. 1One form makes a dinner: a name, a few photos and a short video of the cooking, a night and the number of seats.
2. 2An address and a date: a dinner is a place and an evening, not a listing that is always open. The description is where the host says who it is for.
3. 3The menu is its own page, dish by dish with a rating each. A host builds a reputation on the food, not on the venue.

1 2 3

02The guest books a seat.

1. 1Browse by cuisine, name or location. Each card is a dinner, not a restaurant: a picture of the food, the host, their rating and how far it is.
2. 2The reservation page repeats the date and the seats and asks for allergens, so the host cooks for the people who are actually coming.
3. 3A note to the host. The first message is written before the two have met, which is the point: the conversation starts with the booking.

1 2 3

03The table doesn’t end at dessert.

1. 1After the dinner, a rating and a review. The review is the host’s feedback and the next guest’s reason to book.
2. 2Recommendations: people you ate with, people with the same taste. Following is how a one-night table becomes a circle.
3. 3The feed grew out of the survey: people were excited about the food and wanted the community-building part to be clearer, so it got its own tab.

1 2 3

04A profile that is about taste.

1. 1A profile leads with a bio and favourite food, so a stranger has something to say before the first course.
2. 2Friends and a gallery: the dinners you have been to, in pictures, as the proof that you are a good guest or a good host.
3. 3A calendar of upcoming dinners with a reminder to the host to set the menu. Hosting is a plan, and the app keeps it.

Try it

The prototype, as it was tested.

Outcome

A prototype people understood, and one thing they asked for.

• 55people surveyed, twiceOnce about how they eat, once with the lo-fi mockups: what did they think the app was for?
• 2flows in one appThe host’s and the guest’s, sharing a profile, a calendar and a feed.
• 1change from testingPeople liked the dinners and wanted the community part to be clearer. The feed, following and recommendations were added, and the tagline became “Exchanging food for thought”.

1. 1Three designers, one app. The design system came before the screens so that three hands produced one look, and I owned it as the lead.
2. 2The same person on both sides. A host is also a guest. Keeping one account and one navigation for two jobs was the hardest structural problem.
3. 3Trust between strangers. Ratings, reviews, the menu and the allergen field do the work a restaurant’s sign would do.
4. 4A concept, not a company. It ended as a tested prototype. Safety, payment and liability are the questions a real launch would start with.