This page exists because someone is going to ask. We get asked often enough that it deserves a written answer: how do we test a recipe, how do we write the article, what do we throw away, what do we refuse to publish.
What follows is the working method behind every piece on this site. It is also the standard we hold ourselves to. We update this page when the method changes. The last update is at the bottom.
How a recipe gets tested
A recipe is tested at minimum three times before it is written, and is rewritten at minimum twice before it is published. Most pieces go through more iterations than that. The rule is not three rounds for the sake of three — it is three rounds because we have learned, repeatedly, that the third version is the one where the surprises stop and the dish becomes legible.
Each test is logged in writing. We keep a working notebook for every recipe in development — what we changed, what worked, what failed, what surprised us. By the time a recipe goes to print, we have a record of every variation we tried and a clear answer to the question why this version and not another.
What we test for, in order
- Reproducibility. Can another person, in another kitchen, with normal equipment and supermarket ingredients, get the same result? If the answer is "only if you have a wood-fired oven," we either rewrite for what most people have, or we name the limitation honestly in the article.
- Tolerance. What happens when something goes slightly wrong — a teaspoon over, a minute long, the oven running cool? A good recipe should not collapse from a small error. We test for the edge cases on purpose.
- Honesty about time. If we say a dish takes thirty minutes, it takes thirty minutes including reading the recipe. We do not count "active time" separately from "passive time" except where it genuinely changes the planning — and even then, we say so.
- Honesty about difficulty. "Easy" means anyone can do it. "Medium" means basic technique is required. "Hard" means you need to have done it before once or be willing to fail the first time. We don't grade on a curve.
How a recipe gets written
The first draft is written by the editor who tested the recipe. The second pass goes to one of the other two editors — peer review, with the freedom to question any sentence and any choice. The third pass goes to Morgan, who edits every piece for length, sentence rhythm, and whether each paragraph still earns its space.
An article is not finished until all three of us have agreed it is. This is also why our publishing rhythm is slow. We would rather miss a week than ship something one of us is unsure about.
What we won't publish
We do not publish:
- Recipes we have not personally tested. No exceptions. If we recommend a technique we learned from someone else, we credit them and link to their original work.
- AI-generated content presented as our own. We use AI tools the way other writers use spell-check or a calculator — for research support, fact-checking, occasional copy passes. The author of record is always one of us. The byline reflects who actually did the cooking and the writing.
- Sponsored or paid recipes. If a brand wants to send us a tool, we will use it — and if we end up recommending it in an article, we will say so plainly. We do not accept payment for product mentions, and we do not write recipes commissioned by a brand. This is a discipline we hold to even when it costs us money.
- Listicles, hot takes, or seasonal filler. Every published piece on this site is the result of testing, writing, and rewriting. We do not pad the calendar with quick content to keep traffic up.
- Recipes that don't work. If a recipe fails repeatedly during testing — not because we are doing it wrong, but because the recipe itself is structurally flawed — we mark it dead and move on. We do not publish it with a hopeful caveat.
How we photograph
We photograph in natural daylight, in the kitchens where the recipes were tested. No studio lighting, no food stylist, no plate arrangements that would not survive a real meal. The plate in the photo is the plate we ate at the end of the test. If the dish looked imperfect when we plated it, the photograph shows the imperfect plate.
Photographs are credited to the editor who took them. They are not stock images, and we do not buy generic food photography to illustrate articles.
How we handle errors
Every published recipe is open to correction. If a reader writes in with a substantive note — a measurement that doesn't work in their kitchen, an ingredient that behaved unexpectedly, a step that needs clarifying — we test it ourselves and update the article. Major updates are noted at the bottom of the article with the date.
If we made a clear factual error, we say so explicitly. We do not silently rewrite history.
Tools we trust
For transparency: a kitchen scale we believe in, a thermometer we believe in, and a single sharp knife per cook. We will publish a longer "tools" page eventually, but the short version is — invest in three things, ignore the rest.
Last updated: May 2026.