Emilie Berthiaume
Co-founder, go-to-market
Fifteen years in marketing, including brand and growth work at Cirque du Soleil, Fairmont and Dialogue. She studied dance and finance before that, which is a wider range than most résumés admit to.
LinkedInAbout
Built by two people, for one brain at a time.
Mustard has a small team of two, and that's a design constraint as much as a fact about payroll. A product nobody has time to over-engineer tends to stay the size it should be.
Get early accessWhy it exists
Two kinds of task app. Both failed me.
Hey, I'm Dave. I write software, I play guitar for Quebec artists on the side, and I have the kind of brain that needs a fence around the week or the week eats me.
I tried the simple apps. Loved them for a month. Then I asked “is next week already full?” and got silence. Answering that would need structure they had decided, on purpose, not to have.
I tried the platforms. Statuses, workflows, custom fields, automations, a permissions matrix, a settings page with its own search box. They could answer, right after I spent a weekend teaching them my life, and then kept me teaching them forever. They didn't fail me by doing too little. They failed me by becoming a second job.
So Emilie and I built the thing in the middle: enough structure to answer “does this fit?”, not enough to need an administrator. Cycles, points, velocity. That's most of it.
Built to run its own build
The weeks that make Mustard are planned in Mustard
This isn't a planner designed in a spreadsheet and then sold to you as a planner. The product's own build weeks run on cycles and points, with the velocity number being unflattering on a fairly regular basis.
That has visible consequences. Anything annoying to use every single week doesn't survive to a release. Anything that needs setup you'd have to redo gets cut. And the numbers you see in this copy are the shape of numbers you get from actually running a week: planned thirteen, finished eleven, roll two forward.
It also means the first person inconvenienced by a bad decision is the one who has to fix it.
Boundaries
What it refuses to become
Knowing what a product won't be is worth more than a roadmap. These four are load-bearing.
A platform
No app marketplace, no automation builder, no custom-field system quietly waiting for you to design a schema. Every extension point is a place where the tool asks you to do work on its behalf.
A second job
If keeping the tool current takes more than a few minutes a week, the tool is broken. Setup is not a rite of passage and configuration is not a feature.
A ceremony engine
Cycles, points and velocity came from software teams; the meetings did not have to come with them. Gantt and an Eisenhower matrix are on the roadmap: the ritual that usually rides along with them is not.
A judge
It nudges once, with a number, and then respects whatever you decide. Nothing here keeps score of your streaks or tells you at nine in the morning that you are behind.
Simplicity over completeness
The default answer is a slow no
Every feature request comes from a real person with a real problem, and most of them are entirely reasonable. Saying yes to all of them is exactly how you build the second kind of app: the complete one that nobody can use.
So the bar is a single question: does this help someone plan a week they can actually finish? Points cleared it. Cycles cleared it. Scoped project sharing cleared it. A great many sensible ideas haven't, and they're not sitting in a backlog pretending otherwise.
It's early. Mustard isn't public yet, there's no pricing to show you, and this page is not going to pretend there are thousands of happy users. If the narrow middle sounds like where you live, the waitlist is right below: one email when it opens.
Early access
Next Monday shows up either way. Show up to it planned.
Join the list and two things happen today, not at launch. You get the week-shaped method in your inbox (it works on paper, before the app exists). And you're first in when it opens, while we're still two people who read every reply.
One email now with the method. One at launch, maybe two if something big lands. No newsletter.