The Diet That Works Is the One You'll Actually Follow

If you flatten the last fifty years of nutrition research into one chart, here is the embarrassing summary: every halfway-reasonable diet works for the people who stick to it, and basically none of them work for the people who don't. The factor that predicts success is adherence. Adherence is rarely a function of macros. It is almost always a function of life — schedule, palate, social pressure, finances, willpower, and how the diet feels at month four when nobody is watching.

This is the boring truth the diet industry cannot sell. So instead, every six months, a new prophet appears with a new villain food and a new $40 hardcover, and the cycle resets. The villain is fat. The villain is carbs. The villain is seed oils. The villain is meat. The villain is fruit. The villain is breakfast. The villain is dinner. There is always a villain because there has to be a villain to sell the next book.

The actual data

The cleanest comparison study is still the DIETFITS trial out of Stanford (Gardner et al., JAMA 2018). Six hundred adults, twelve months, randomized to "healthy low-fat" or "healthy low-carb." Both groups lost weight. The mean difference between the two diets was statistically indistinguishable. Genetic profiles, insulin status — none of the much-hyped predictors mattered. What mattered: did you eat the way you said you'd eat?

Earlier and larger meta-analyses (Johnston, JAMA 2014, comparing eleven popular diets across 7,000+ participants) found the same. The differences between diets, at one year, were trivial — under five pounds, on average. The variance within each diet — between the people who stuck with it and the people who didn't — dwarfed the differences between diets.

If a study tells you which diet is best, ask whether the participants finished the study. Most of them didn't.

What actually drives adherence

Practically, the diets people stick to share boring features:

• The food is food they already like. A diet built around lentils when you hate lentils is going to fail. A diet built around chicken thighs when you eat chicken thighs anyway is going to work.
• The rules are few and unambiguous. "No food after 8pm" is followable. "Avoid pro-inflammatory foods" is a Rorschach test.
• The social cost is low. Diets that require you to refuse food at every wedding, every birthday, and every business dinner have a half-life measured in weeks.
• It works around your existing chaos. If your job is night shifts, intermittent fasting on a "no eating after 6pm" schedule will fail. The diet has to fit the life, not the other way.
• The hunger is bearable. Some diets keep people fuller per calorie (high-protein, high-fiber). Some don't. Bearable hunger is a real engineering input, not a moral failing.

The obvious counter

"But there are real metabolic differences between diets." There are some — keto creates a real metabolic shift, very-low-calorie produces faster initial loss, and there are edge cases (PCOS, severe insulin resistance, epilepsy) where specific dietary patterns are genuinely therapeutic. Granted. None of that contradicts the main point: for the median person trying to get healthier, the marginal benefit of picking the "metabolically optimal" diet is small, and the marginal benefit of picking one you can sustain for ten years is enormous. Compounding favors adherence the way compounding favors index funds.

"GLP-1 drugs make all this irrelevant." They change the calculus for some patients dramatically. Most people are still not on them, and most who are still benefit from a sustainable food framework around the medication. The drug isn't a substitute for figuring out what you'll eat the rest of your life.

The response

If you are trying to design a sustainable diet, run this checklist before adopting any plan from any influencer:

1. Could I eat this way at my mother-in-law's Thanksgiving without making it weird?
2. Will I be able to do this on a Tuesday at 9pm when I am tired and the kid is sick?
3. Is there a single rule I can repeat in one sentence?
4. Does the plan have a graceful failure mode — am I "off the wagon" if I miss one meal, or do I just resume?
5. If I look at this list of foods, are at least 70% of them foods I already enjoy?

If your plan passes those, the macronutrient ratio almost doesn't matter. If your plan fails any of them, you can hit the macros for six weeks and you will still be done by month four, exactly like the previous three plans.

The diet industry will keep yelling. The fight isn't about you. It's about the next book. Pick the boring one you'll actually follow, eat dinner, and stop reading articles about diets — including, possibly, this one.

← Back to all essays