Personalised nutrition: what works and what to test

By Dan Remon · Evidence checked October 2026

You’ve tried the diet your colleague swears by. It worked for them. It did very little for you.

That isn’t a self-control problem. The same food affects different people in different ways, and the research behind that is solid.

What’s less solid is everything sold on top of it. Gene tests, gut kits and apps all promise a diet written for you. Some of it holds up and some of it doesn’t. This is how I grade the evidence, what I test, and what I leave out.

What is personalised nutrition?

Personalised nutrition means building what you eat on data about you instead of population averages. That data can include your blood work, your body composition, how your blood sugar and blood fats respond to meals, and the week you actually live: travel, client dinners, late meetings and training. A personalised plan changes one thing at a time and measures what happens. A generic meal plan gives everyone the same answer. The research is clear that people respond differently to the same food. It’s much less clear that every test sold as personalised helps you decide what to eat.

Why the same meal hits people differently

Two large studies answered the first question. In 2015, researchers at the Weizmann Institute in Israel monitored blood sugar continuously for a week in 800 people and measured their responses to 46,898 meals. Responses to identical meals varied widely from person to person, which led the authors to suggest that universal dietary advice may have limited value.¹

In 2020, the PREDICT 1 study gave 1,002 twins and unrelated healthy adults in the UK identical meals and measured what happened next. The spread between people was large: a coefficient of variation of 103% for blood triglycerides, 68% for glucose and 59% for insulin.² Person-specific factors such as the gut microbiome had more influence on the rise in blood fats after a meal than the meal’s macronutrients did (7.1% against 3.6% of the variance). For blood sugar it was the other way round, and the meal mattered more (15.4% against 6.0%). Genetic variants had a modest impact on the predictions: 9.5% for glucose and 0.8% for triglycerides.

What I take from both: your response is partly you and partly what’s on the plate. Both are worth measuring, and your genes explain only a small part of it.

Does personalised nutrition work?

Showing that people differ is one thing. Showing that a personalised plan beats good general advice is harder. This is the evidence graded, strongest first.

1. Strong: people respond differently to the same meal. Two large studies with detailed measurement, each validated in a separate group of 100 people.^1,2
2. Moderate: responses can be predicted, glucose better than blood fats. The PREDICT 1 model predicted glucose responses with r = 0.77 and triglyceride responses with r = 0.47. In the Weizmann work, a blinded randomised trial of diets built on their prediction lowered blood sugar responses after meals.^1,2
3. Modest: personalised programs beat standard advice, by a little. In a 2024 randomised trial of 347 US adults aged 41 to 70, an 18-week app-based personalised program lowered triglycerides by 0.13 mmol/L more than standard dietary advice. LDL cholesterol didn’t change significantly, and glucose, insulin and blood pressure didn’t differ between groups. Weight, waist, HbA1c and diet quality improved, particularly in people who stuck closely to the program. Several of the authors work for the company behind it.³
4. Useful: personalised advice changes what people eat. In Food4Me, a 6-month randomised trial in 7 European countries with 1,269 completers, people given personalised advice ate less red meat, salt and saturated fat and scored higher on the Healthy Eating Index than people given conventional advice.⁴
5. Weak: genes as a diet selector. In Food4Me, adding body measurements and blood markers, or those plus 5 genetic variants, made the advice no more effective. In DIETFITS, a 12-month randomised trial of 609 adults, people lost an average of 5.3 kg on a healthy low-fat diet and 6.0 kg on a healthy low-carb diet. Neither genotype pattern nor baseline insulin response predicted who did better on which.⁵

So the grade is this. The biology is real, prediction is improving, and the benefit over good general advice is real but small in trials lasting 18 weeks to 12 months. The advice that worked in Food4Me was built on what people actually ate. That’s where I start.

What to test, and what to skip

A test earns its place when it changes a decision or shows you whether a change worked. On that standard, this is how I rank the options.

• Blood work. It runs through our partners at Orba Health, the health company I co-founded. The markers the trials above tracked make a sensible core: triglycerides, LDL cholesterol, glucose, insulin and HbA1c. Blood work is your baseline and your scoreboard. Anything that needs medical attention goes to your doctor.
• Body composition. Weight alone can’t tell you whether you’re losing fat or muscle. If fat loss is the goal and the scale has stopped, read why the scale stalls when you diet.
• Your actual week. 7 days of what you ate, when, and how you felt afterwards. It costs nothing, and in Food4Me, advice built on diet alone did as well as advice that added blood and gene data.
• Meal-level response, if you want detail. A week of continuous glucose monitoring is how the Weizmann team measured responses. It shows how your blood sugar reacts to your meals. It shows nothing about blood fats, which varied even more between people in PREDICT 1.
• Skip: DNA tests to choose a diet. DIETFITS and Food4Me both found genotype didn’t help. Gut microbiome tests are earlier still. The microbiome is one input in the personalised programs tested so far, and their benefit over standard advice was modest.

The Energy Audit: your first week of data

The W2 Energy Audit is the second of the 12 protocols in the Stealth OS. Its job is to find where your energy rises and drops across a normal week, and what’s behind it. Food is one input, alongside sleep, training and travel.

For nutrition, it starts with a 7-day log. Don’t change anything yet. You’re recording a baseline.

• What you ate and drank, and when. Every meal, snack and coffee.
• How you felt 2 hours later. Sharp, flat or hungry, in a few words.
• The context: how you slept, whether you trained, whether you travelled or ate out.

At the end of the week, read it back. Look for the meals that were followed by a dip, the ones that held you steady, and the days that fell apart and why. Pick the pattern that cost you the most energy. That’s your first change.

The log is how you see which change did it.

How I build a personalised nutrition plan

1. Test first. Blood work through our partners at Orba Health, plus body composition. We work from your numbers, not a template.
2. Look at how you actually eat. Travel, client dinners, late meetings. The plan has to fit the week you have. I’ve written about eating well at any restaurant on any day in a flexible approach to nutrition for performance.
3. Adjust one thing at a time. When I want to drop body fat, I don’t cut the amount I eat. I change the macros, or the order I eat them in. Then I watch what happens. One change per 28-day cycle, so you know what caused the result.
4. Measure again. Retest, compare, keep what worked and drop what didn’t.

If the goal is weight loss on a schedule that leaves little room for the gym, start with effective weight loss for busy executives.

Two things to install this week, before any test comes back: protein at every meal, starting with breakfast, and the 7-day log.

How to start

1. The 1% Scorecard. 8 questions, 2 minutes, free. You get a score out of 100, the system that’s leaking most and the one thing to install first.
2. The Quest. My performance membership, run in 28-day cycles. The Stealth OS, including the Energy Audit, installed one change per cycle, with a structure that keeps running on the weeks you don’t feel like it. Cancel anytime.

Related reading: Nutrition coaching: what a coach does and what the evidence says, Infused water: 8 ideas and the tips that keep you hydrated and Type 2 diabetes remission: what the research shows.

FAQ

Sources

1. Zeevi, D., Korem, T., Zmora, N., et al. (2015). Personalized nutrition by prediction of glycemic responses. Cell, 163(5), 1079-1094. doi.org/10.1016/j.cell.2015.11.001
2. Berry, S. E., Valdes, A. M., Drew, D. A., et al. (2020). Human postprandial responses to food and potential for precision nutrition. Nature Medicine, 26(6), 964-973. doi.org/10.1038/s41591-020-0934-0
3. Bermingham, K. M., Linenberg, I., Polidori, L., et al. (2024). Effects of a personalized nutrition program on cardiometabolic health: a randomized controlled trial. Nature Medicine, 30(7), 1888-1897. doi.org/10.1038/s41591-024-02951-6
4. Celis-Morales, C., Livingstone, K. M., Marsaux, C. F., et al. (2017). Effect of personalized nutrition on health-related behaviour change: evidence from the Food4Me European randomized controlled trial. International Journal of Epidemiology, 46(2), 578-588. doi.org/10.1093/ije/dyw186
5. Gardner, C. D., Trepanowski, J. F., Del Gobbo, L. C., et al. (2018). Effect of low-fat vs low-carbohydrate diet on 12-month weight loss in overweight adults and the association with genotype pattern or insulin secretion: the DIETFITS randomized clinical trial. JAMA, 319(7), 667-679. doi.org/10.1001/jama.2018.0245

This article is general information, not medical advice.