Skip to main content
For Home

The Last Supper Club

The bottleneck on end-of-life conversations is not paperwork, it is starting them, so almost no one talks about their wishes and the people they love are left to guess, argue, and carry guilt during grief. Only 31% of US adults have any advance directive.

Launch The Last Supper Club

What it does

Break the seal on the conversation everyone is avoiding. Run a warm, guided dinner that turns a heavy topic into connection, gets real wishes on the record, and leaves your people closer instead of dreading it.

Currently

Live MVP. Free, browser-only, no signup, nothing stored on a server. A guided three-round dinner flow (the Vibe, the Heart, the Keepsakes) walks a group through end-of-life wishes easy-first, with a skip on every question, and produces a per-guest Wishes card you can copy or print. The honest gaps are real. Answers live only in browser memory and vanish on refresh, there is no durable or shareable keepsake yet, and the tool is uninstrumented, so the North Star is modeled, not measured. Instrumentation plus a durable, shareable keepsake are the next things in.

Roadmap

/now-next-later-roadmapNow/Next/Later Roadmap
Now
  • Instrument (anonymous, privacy-safe, answer-content-free)
  • Durable keepsake (local persist + re-openable card)
  • Shareable keepsake link (the K test)
  • Send-it-forward prompt
  • Occasion templates (Friendsgiving, milestone, memorial)
  • Grief-safety framing + keep/strengthen easy-first and skip
  • Distribution: warm-content moment + 2-3 partnerships (non-build)
Next
  • Warm host onboarding + shareable pre-dinner invite
  • Annual come-back-next-year nudge
  • Solo mode (validate demand first)
  • Practitioner host mode + co-brand (interviews first)
  • Speak-your-answer (voice)
  • Emailed / social-ready keepsake
Later
  • Physical companion deck
  • Culture / faith question packs
  • SEO / landing for the trigger
  • Practitioner directory
  • Adjustable depth toggle

Recently shipped

  • Mar 24, 2026Shipped the MVP: a guided three-round dinner conversation for end-of-life wishes, with copy-to-clipboard and print-to-PDF on a per-guest Wishes card.

How I know it's working

instrumented Β· awaiting traffic
North star
Monthly Completed-and-Kept Dinnersβ€”

The count, per month, of dinners where a group finishes all three rounds and at least one guest saves or shares a keepsake. Completed captures delight (they reached the end), kept captures reach and durable value (a saved or shared keepsake is the product's only growth loop). It scores a bounced visitor, an abandoned setup, or a finished-but-forgotten session at zero. By design it is not revenue and not engagement-minutes; this is an episodic product, not a daily habit.

The events are wired into a privacy-first store. Counts populate here as people use it. This is the honest part of the work: instrument first, claim later.

How it's built

Browser-onlySolo

How I built it

Last Supper Club started from a number that stuck with me: only 31% of US adults have any advance directive, and almost no one has said their wishes out loud. The bottleneck is not paperwork. It is that the conversation feels too morbid to start, so people wait for a crisis, then their families are left guessing, arguing, and carrying guilt during grief. The app is a guided dinner that lowers the activation energy: three rounds, easy-first, a skip on every question, that turn a heavy topic into a night that brings people closer. So I ran the full product process on it, the same one I would run for a client.

What the discovery surfaced

A jobs-to-be-done pass made the differentiated job sharp: it is the feeling, not the file. Every incumbent (Cake, Lantern, Everplans) leads with documents and logistics. Breaking the seal and turning dread into connection is the white space, and almost nothing targets it. Three personas fell out of that, and the strategic order matters. Maya the host is primary because she finds, runs, and shares the product. Gene the reluctant guest is the subject, never the buyer, so the warmth and the skip option are built to retain him, not to market to him. Robin the death-positive practitioner is the leverage: one Robin can put the tool in front of hundreds of Mayas, and she is the only plausible revenue path.

Because the triggers are episodic (a death, a diagnosis, a milestone) this is not a habit product. The North Star reflects that. Monthly Completed-and-Kept Dinners combines the two things that matter: completed captures delight, kept captures reach, since a saved or shared keepsake is the only growth loop the product has. It scores a bounced visitor or a forgotten session at zero on purpose.

What's shipped, and what's honest about it

The MVP went live in March: the three-round flow, the per-guest Wishes card, copy-to-clipboard and print-to-PDF. RICE ranked the backlog, but I did not follow it literally. The strategically central build, the durable plus shareable keepsake, scores below trivial items only because its growth assumption is honestly unproven, so the roadmap keeps it as the centerpiece and elevates occasions and distribution above their RICE rank. The pre-mortem named the real risks plainly: the keepsake might get shared without creating new hosts, dinners might never start at all, and measuring a death-topic app could feel like surveillance and break the exact trust that differentiates it. So the honest next bet is not new surface area. It is privacy-safe instrumentation, then a keepsake worth keeping and sharing, then getting the tool in front of people at the moment a trigger hits. Today every metric in the strategy is modeled, not measured, and that is the first thing to fix.

The strategy behind it

A snapshot of the product process this app was built from. Each artifact comes from a k8 skill you can run yourself. It reflects the original plan, not the live state above.

Jobs to be done

/jtbd-analysisJTBD Analysis
When I… I want to…ImportanceFrequency
When I know my family should talk about end-of-life wishes but every attempt feels too morbid or too sudden, I want a low-stakes, structured way to start, so I can stop avoiding it and finally get something on the record.highlow
When I think about my own death, I want the people closest to me to know what I actually want, so they are not forced to guess, argue, or carry guilt during grief.highlow
When my people gather, I want our hardest conversations to bring us closer rather than scare us, so the night feels like a gift, not a chore.highmedium
When I want to bring my family or friends through this conversation, I want a script that does the facilitation for me, so I can take part instead of playing teacher.mediummedium
When the dinner is over and real things were said, I want to hold onto them, so the answers can actually be honored later instead of half-remembered.mediumlow

Persona

/persona-createPersona Create
M
Maya, The Host
Primary persona. The recently-bereaved family connector, 35-50, who runs the group chats and refuses to let her family be caught unprepared again. She finds, runs, and shares the product.
Goals
  • Get her mother's and her own wishes on the record before another crisis forces it
  • Make the conversation feel loving, not clinical, so her family will actually go through with it
  • Have a keepsake to hold onto afterward
Frustrations
  • Every time she tries to raise it, it lands wrong: too heavy, wrong moment, someone changes the subject
  • She does not want to be the morbid one who makes everyone sad at dinner
  • The tools she has found feel like tax prep, not a conversation

Persona

/persona-createPersona Create
G
Gene, The Reluctant Guest
The parent, 60-75, who would never bring it up but will play along once someone else does. The subject of the conversation as much as a participant, and never the one who acquires the product.
Goals
  • Spare his kids the chaos he saw with his own parents
  • Be remembered as himself: his music, his jokes, his particular requests
  • Not turn a nice dinner into a downer
Frustrations
  • Would never initiate; bringing it up feels like burdening the kids or tempting fate
  • Worries the conversation will get heavier than he wants
  • Distrusts apps that ask for personal data, especially about death

Persona

/persona-createPersona Create
R
Robin, The Death-Positive Organizer
Growth and distribution persona, not volume. The death doula, hospice volunteer, or death-cafe host who runs these conversations for others and can put the product in front of hundreds of Mayas. Also the only plausible revenue path.
Goals
  • Facilitate meaningful conversations without reinventing a script every time
  • Give participants something to take home
  • Find products good enough to recommend to their whole network
Frustrations
  • Existing tools are either too clinical (advance-directive forms) or too unstructured (open Death Cafe with no prompts)
  • Nothing produces a keepsake the participant values
  • Hard to use a solo-built tool with a group
Prioritization and risks(the internal detail)

RICE prioritization

/rice-scoreRICE Score
FeatureReachImpactConf.EffortScore
Keep + strengthen easy-first / skip (validate)5000.580%0.21000
Send-it-forward prompt4001.550%0.5600
Grief-safety framing5000.570%0.3583
Warm-content moment (Kate's story)5001.550%1375
Shareable keepsake link400345%1.5360
Durable keepsake (local persist + re-openable card)500270%2350
Occasion templates600240%2240
Death-positive partnerships800245%3240

Pre-mortem risks

/pre-mortemPre-mortem

The keepsake loop fails: people share a keepsake but no new hosts arrive, so K stays near zero and the whole reach thesis collapses.

medium

Mitigation: Front-load the cheap K test: instrument the share-to-new-host path first and treat the keepsake as delight and retention value regardless. If K stays low, reweight reach to practitioners and content.

Dinners still do not start: even a perfect keepsake does not make a group sit down, and non-consumption is the category norm.

medium

Mitigation: Pull occasion templates and distribution into Now. Lead with concrete moments (Friendsgiving, milestone birthdays) and partner where gatherings already happen, then watch whether starts actually rise.

Instrumentation or persistence reads as surveillance on a death app, breaking the no-signup, nothing-stored trust that differentiates it (Gene bounces).

medium

Mitigation: Anonymous, answer-content-free events only; local-first persistence with clear user control; plain-language transparency. This privacy design is launch-blocking. Fallback: ship persistence with zero analytics rather than analytics that erode trust.

Solo-builder capacity: the Now lane is too big for one part-time person and slips, so nothing lands.

high

Mitigation: Protect the unblocking core (instrument plus durable and shareable keepsake) and let occasion templates, onboarding, and quick wins slide. Shipping the keepsake pair alone still proves the central hypothesis.

Grief-ambush: a recently bereaved guest is blindsided when round two goes deep, causing real human harm and reputational damage.

medium

Mitigation: Ship grief-safety framing: a gentle upfront note, the existing skip, and an optional lighter-tonight mode. At minimum, a one-line content warning before The Heart.

localStorage on a shared family computer leaks sensitive answers to the next person who opens the browser.

medium

Mitigation: Do not persist by default without intent; offer clear-after; warn on shared devices; never store on a server without an explicit action from the user.

Built with these skills