Do 95% of Diets Fail? Where the Zombie Statistic Came From
You’ve heard it from a personal trainer, an anti-diet infographic, and an ad for a pill that swears it’s the 5%. So do 95% of diets fail? The number traces back to a 1959 case series of 100 people at a New York hospital nutrition clinic. They got one counseling visit, a diet sheet, and a goodbye. The number 95 doesn’t appear anywhere in the paper. Its own author told The New York Times in 1999 that people were stretching his data way past what it could hold. Nobody listened.
This episode doesn’t argue for dieting or against it. It investigates a statistic both camps love, because a round number that confirms what you already believe is very hard to put down.
Have you heard that 95% of diets fail?
Zoë: Welcome to Your Diet Sucks, the podcast that just keeps coming back. I'm Zoë Rom.
Kylee: And I'm Kylee Van Horn.
Zoë: Kylee, have you heard the statistic, or this idea, that 95% of diets fail?
Kylee: Maybe in passing. Reading online, social media, maybe an ad from a company trying to sell something.
Zoë: Yes. What's crazy about this episode, and the reason I wanted to do it, is that I've used this statistic. I've definitely said it to convey the sense that trying to lose weight by severely curtailing the amount and types of food you eat doesn't work long term.
Kylee: So restrictive dieting, right? Because diet can mean a couple of things.
Zoë: Right, and we'll talk about this. I know you're already crawling out of your chair because you want to define terms, which is why you're the best co-host of this podcast ever. But I haven't been thinking in that sophisticated a way when I've repeated the statistic, and I don't think it's been received in that sophisticated a way. So I'm excited to disentangle all of this. The idea is that changing what you eat in these super restrictive ways does not lead to long-term weight loss.
Kylee: If somebody says the word diet, I feel like they think we're talking about a weight loss diet. But a diet can also just be the way that you eat.
Zoë: Which is literally why we named the podcast Your Diet Sucks. When I say your diet sucks, I mean your restrictive way of eating: your keto diet, your paleo diet, your ancestral diet, your gluten-free diet for no reason. But a lot of people hear it as "my pattern of eating sucks." So how do you tease out those definitions with people? Because this is your whole thing.
Kylee: I ask people what their style of eating is. What are your habits or thoughts around food? What are your patterns of eating? Are they serving you in a way that's productive? Or is this more of a restriction to lose weight or change your body size?
Zoë: Exactly. We're not a diet podcast for weight loss, but we're going to be talking a lot about diets today. So to bookend this up front: this is not an episode arguing for diets, and it's not an episode arguing against diets. This is an episode born out of genuine curiosity about a statistic I hear all the time and that I've repeated. Something recently clicked for me that it's way too simple to actually make sense, so I need to make an entire 90-minute podcast about it.
Zoë: We're not exactly neutral on the idea that you should severely restrict what goes into your body. We have a perspective on that, which is that a lot of ideas people are sold are harmful and don't make sense. But I still think it's important to hold this statistic up to the light and understand the ways it has bolstered both diet culture and anti-diet culture. No matter where we land, I want it to be informed by the science, and I don't want to repeat things that aren't true. And I want to understand why this thing I suspected wasn't true was so culturally sticky.
Kylee: I feel like this is a trend with a lot of wellness topics. A number shows that something is true, and it's very hyper-specific.
Zoë: Or it's so unspecific. Okay, 95% of diets fail. What do you mean by diet? What do you mean by fail? What do you mean by percent? You can tear apart any single aspect of it. So what I want to do is look honestly at the long-term data on whether diets work, not as an endorsement or a teardown of dieting, but as a question in and of itself. This is one of my favorite episodes because I've really been on a journey with it, and we also get to talk about Kevin Hall.
Kylee: And you've maybe changed some of your initial thoughts.
Zoë: That's every episode. The only one I haven't changed my mind on is Weight Watchers. It was dumb the whole time. That was just type-one fun of "this is stupid." This is not one of those episodes.
Where does the 95% statistic come from?
Kylee: I've heard the 95% thing in eating disorder recovery spaces, intuitive eating spaces, gyms, from personal trainers, and in ads. So where does it actually come from?
Zoë: People see it everywhere: diet books, anti-diet books, podcasts, Instagram infographics. It functions as a push-your-reading-glasses-up-your-nose "well, actually." It has the authority of a super specific number attached, so it's been thoroughly science-washed. One camp throws it out to say, see, nothing works, diets are bad. The other side, usually the diet industry, holds it up to say 95% of diets fail, but not ours.
Kylee: Our pill will work.
Zoë: Exactly. Any time a statistic is this useful to this many people, it's worth being suspicious of the work it's doing. When a stat like this goes viral, it hits three of our cultural erogenous zones. It's a round number. It's never 72.106%. It's always 95. It confirms something people already suspect, which is that diets are hard to sustain. We love research that backs up our priors. And it's just specific enough to sound true. At 99% you'd think that seems high. At 100% you'd think that doesn't feel right. 95 is round enough to remember and precise enough that it doesn't sound made up.
Kylee: If I heard 99, I'd be like, that seems fake.
Zoë: Twelve out of seven dentists recommend.
Kylee: So the number had to come from somewhere, right? Or is it just vibes?
The 1959 Stunkard study
Zoë: Patient zero is a 1959 paper by Stunkard and McLaren-Hume in the AMA Archives of Internal Medicine. It's called "The Results of Treatment for Obesity: A Review of the Literature and Report of a Series."
Kylee: Hell of a title.
Zoë: I see why we remember 95 and not "Report of a Series." Albert Stunkard was a psychiatrist who spent about 50 years on this. At the time he was at New York Hospital, and later he ended up at Penn. Mavis McLaren-Hume was a dietitian. These were mainstream figures, not grifter, influencer, John Harvey Kellogg types.
Zoë: Here's what Stunkard actually did, and I think it's pretty valid. He started with 30 years of published research on treating obesity and applied some minimal quality standards. Only eight studies made it through his filter.
Kylee: That's kind of small.
Zoë: It's not great. In those eight studies, about 25% of patients lost 20 pounds and about 5% lost 40. Those were all published by experts, so he figured regular practice was probably worse and went to check. That's where the 100 patients come from: 100 consecutive people referred to the nutrition clinic at New York Hospital. When you zoom in, the treatment was literally a diet sheet, a sheet of paper.
Kylee: Like with rules?
Zoë: Yeah. You got counseled once, at your first visit. One time. You were on a diet somewhere between 800 and 1,500 calories. Then you were released into the wild.
Kylee: Wow. 800. You can't even sit in a chair and really function.
Zoë: Is that three cigarettes and an animal cracker? There were no check-ins and no follow-up. To be fair, that's just what standard care looked like. 1959 was not exactly concierge medicine or Jenny Craig. This was before behavioral weight management existed as a field. The toolbox was: here's a piece of paper, good luck, chap.
Kylee: That sounds like clinical nutrition in the hospital these days. These people are the sickest they've ever been, you hand them a sheet as a dietitian, walk away, and expect them to make dietary changes.
Zoë: I'm imagining you walking down a hospital hall while "The Sound of Silence" plays.
Kylee: And this is why I do this podcast.
Zoë: Stunkard's hypothesis was correct. Instead of 25%, only 12 people out of the 100 lost 20 pounds. Instead of 5%, one person lost 40. Two years later, two people had kept a 20-pound loss off. Also worth knowing: the 20- and 40-pound cutoffs weren't even his idea. They were inherited from the literature, and a colleague at the Rockefeller Institute suggested scoring it that way. So these arbitrary round numbers got laundered through a multi-year discourse that amounted to a colleague at the water cooler saying, what if you made the cutoff 20 and 40 pounds?
Kylee: So if somebody lost 30 pounds, they're not counting it?
Zoë: They'd get rounded down into the 20-pound bucket. Your eyes just got so big. Not ideal science, even in the late 1950s. And even with that raw bucketing, it's still not 95%. Depending on which threshold you pick, the failure rate is 88% or 98% or 99%. No matter how you slice it, 95 isn't in the paper. It is straight made up.
Zoë: It's also worth acknowledging that this is a case series. It's barely even a study. It's an EP, not a full vinyl record. Researchers watched a group of patients and wrote down what happened. There was no control group, no randomization, nothing to compare against. Maybe a half-decent starting point, but not a conclusion we should draw meaningful information from, let alone still be using today.
How a 1959 case series became a "fact"
Kylee: So how does it go from 100 people in 1959 to this idea?
Zoë: Mostly repetition. It got cited, then it got cited again, then people cited the people who cited it. After enough rounds, the sample size, the study design, and the date fade into the background, and what survives this game of bad-science telephone is 95. Like dad jeans, it really took off in the '90s. National Institutes of Health panels put out statements that most weight loss gets regained within five years, and once something gets wrapped up in official government language, it stops sounding like a study and starts sounding like a fact.
Zoë: Then in 1999, a New York Times reporter named Jane Fritsch did the thing no one had bothered to do. She thought, huh, this feels weird, and went to read the original study.
Kylee: This is something you would do, Zoë.
Zoë: Shout out to Jane. The headline was "95% Regain Lost Weight. Or Do They?" Which, love. That is so YDS. To be clear, we are not the first debunkers here. We're standing on the debunking shoulders of giants. Fritsch talked to Stunkard, and he told her the number was being stretched way past what his data could hold. He pointed out that his patients had basically been handed a diet and sent on their way. The guy who wrote the study was like, y'all, please stop. That was published in the New York Times, and the cultural conversation basically didn't move.
Kylee: Wow.
Zoë: That's either extremely funny, a bit depressing, or job security, depending on how you look at it. And the statistic has mutated along with diet culture. The 1959 paper says roughly two out of 100 people at two years. Now you can find writing that says 97% of weight loss attempts end in regain or disordered eating within five years, citing a 2007 paper that doesn't say that anywhere.
Kylee: Check your citations.
What does it mean for a diet to "fail"?
Zoë: A lot of this depends on how you define failure. And this is how I know I've properly YDS-pilled you: when I first brought this idea to you, you were extremely annoying. You had questions about what diet means, what failure means, and what 95% means. Her transformation is complete.
Zoë: The trajectory that gets cited most comes from Hall and Kahan in 2018, summarizing a meta-analysis of 29 long-term studies.
Kylee: Which is more than eight. Love that.
Zoë: More than half of the weight lost in these studies comes back within two years. By five years, more than 80% of it is back. If you're zooming through, that sounds bad. But pause for a beat: that's 80% of the weight, on average. That's not 80% of people regaining everything. Those are different things. Say somebody loses 40 pounds and puts 16 back on over five years. They've regained 40% of what they lost, and they're also still down 24 pounds. Did that diet fail?
Kylee: It depends on where you draw the line.
Zoë: Which is why you don't go to the water cooler and let your buddy say, I don't know, 40. That's not how you do science. And the 1959 paper's bar for success was losing 20 pounds.
Kylee: That was the only thing they measured? Not feeling better, not moving better, not any real health outcome?
Zoë: Nothing but what size jeans you wear. Twenty pounds wasn't chosen because that's where metabolic biomarkers improve. It was just a number pulled out of a hat.
Kylee: Which shows how simplistically we view weight loss as the highest achievable outcome, rather than asking whether these people feel any better. Did they stop smoking? Are they walking? That's what I try to get people to understand: what are we looking at outside the number on the scale?
Zoë: So you can see how regular people get fixated on numbers that science has also been fixated on. If you lost 18 pounds, or 19.5, you went into the failure column, regardless of your starting weight. These are blunt tools. And more than a quarter of those patients never came back after the first appointment.
Kylee: Wait, what? That's kind of important.
Zoë: They just didn't show back up, and they immediately got grouped in as failures. I don't think that's an appropriate descriptor, because they straight up weren't on a diet.
Kylee: That's not how anything works.
Zoë: So when somebody says 95% of diets fail, the actual rebuttal is: define fail, define diet, and define over what period. Which is also why I'm zero fun at parties.
Kylee: Both of us now.
Zoë: This is why we are not popular.
Why diets fail: physiology, design, and environment
Zoë: Before we go anywhere near whether that number is right: diets do fail a lot. I came into this with the vibe that diets just kind of don't work, people regain the weight, and yet people keep doing them, stuck on a treadmill of the next trendy thing. What are you hearing from clients? Do people think dieting is effective? Do they think it's hopeless?
Kylee: My perspective is skewed because of the types of people I work with. A lot of them have been dieting their whole lives, it appears not to work, and they're stuck in a yo-yo cycle. And if somebody felt successful with their diet, would they be working with me?
Zoë: You're right to acknowledge the limits of your sample size. But are you hearing from folks who thought it would work, it didn't, and now they feel stuck?
Kylee: All the time. That's probably the biggest percentage of people I work with.
Zoë: To be fair, the failure rate for long-term weight loss is high. Pretty much everyone agrees. You won't find a mainstream expert saying weight loss is super easy and happens all the time. So let's talk about why. I'd file things in three buckets: physiology, design, and environment.
Zoë: Physiology first. After weight loss, your hunger hormones shift toward wanting to eat more. Ghrelin goes up and leptin comes down. Leptin is the fullness signaler. That's from a 2011 study in the New England Journal of Medicine where people did a 10-week low-calorie diet, and a year later those hormones were still measurably shifted. The metabolism side, which we'll get to with our favorite Biggest Loser study, is less settled. And the body defends much harder against losing weight than against gaining it. We've talked about set point theory: your body has a calibration for its happy place, and for survival reasons it would rather nudge you up than let you go down.
Zoë: The second bucket is design: how studies measure this stuff. In the research, weight loss typically peaks at about six months. A 2007 review in the Journal of the American Dietetic Association looked at 80 randomized trials and over 26,000 people across pretty much every approach: low carb, low fat, meal replacements, counseling. On average, weight loss plateaued around six months. That plateau happens to people who are still following the plan. It also arrives right around when a lot of study support ends, because these studies are hugely expensive. Imagine bankrolling a bajillion Kylees in the field with these people.
Kylee: For years.
Zoë: For years. So the higher-quality, more controlled studies tend to be shorter, and the longer studies tend to be observational, so the data gets shaggy.
Zoë: There's also research on rigidity versus flexibility. A rigid diet means foods are allowed or forbidden, and one deviation means the day is ruined. That predicts worse outcomes and more binge-restrict cycling.
Kylee: Which I would have guessed. Didn't need to read that study.
Zoë: Flexible means you can adjust without the whole thing collapsing. Flexible approaches predict more sustainable weight loss and less binge-restrict cycling, which nobody would argue against. So what predicts failure isn't how hard the plan is. It's how brittle it is. The Mediterranean diet is one of the better-studied diets, and it's very not brittle.
Kylee: I have people who tell me, I don't like that, I need one end or the other, that's how I operate best. Then we get to talking, and I ask, is this how you achieve your definition of success?
Zoë: Go back and listen to our OG episode about diet cults. People psychologically prefer rigidity. They want to be told how to live. But when it comes to actually embodying those rules while living in the world, it's a shit show. People are drawn to the thing that's significantly worse for sustaining body composition changes. And again, we're not saying that's the best outcome. We're asking whether that outcome fails 95% of the time.
Zoë: Then there's environment. You're eating in a food supply built by people who have better data on your behavior than you or most medical research institutions do. And the goal is often wrong from the start. Clinical success tends to be defined as losing 5% or 10% of body weight, because that's where blood pressure and glucose tend to move. That's from Wing and Phelan, 2005, in the American Journal of Clinical Nutrition. But humans don't start a diet thinking, I want to lose 5% of body weight and change my glucose markers. They think, I want to look good for spring break, or I need to lose 40 pounds because that's what I weighed in high school.
Kylee: It's a random number they made up or heard, or weighed 50 years ago. I ask people where it came from.
Zoë: These numbers are so much more related to feelings than facts. And then failure gets sold back to you as proof you need the next program. It's not that diets are hard and largely ineffective. It's that this one was, or something about you made it ineffective.
What long-term weight loss studies actually show
Kylee: I'm betting there's been a good bit of research since 1959. What has it shown?
Zoë: Thankfully, we have been doing research since 1959. I want to start with one of the bigger, longer studies, a trial called Look AHEAD, with over 5,000 adults with type 2 diabetes. It was randomized, with eight years of follow-up, and they still had 88% of people at the end, which is amazing retention for this kind of study. At eight years, the intensive lifestyle group averaged about 4.7% of body weight below where they started. Around half were still down at least 5%. About 27% were down at least 10%. And of the people who lost 10% or more in the first year, roughly 39% still had it off at year eight. So in the best-supported trial we have, the failure rate at the clinically meaningful threshold is somewhere around half.
Kylee: That's not 95.
Zoë: It's also worth acknowledging this was in people with type 2 diabetes, so there's motivation beyond looking good in a swimsuit. So much of this research is on people with metabolic conditions, and then it gets extrapolated to everyone. I don't think we throw it out, but if someone tells me I won't look hot by some arbitrary cultural standard, I'm pretty okay with that. If someone tells me my metabolic health is going to be really messed up and I might lose a foot, I'm going to be motivated.
Kylee: That was a long period of time. Did these people have ongoing support from professionals?
Zoë: They had ongoing support and counseling. It wasn't wham, bam, here's your sheet that tells you not to eat fructose. If I'm not mistaken, there were a couple of different methodologies, so it wasn't one study on 5,000 people but several trials repackaged into Look AHEAD. The caveat is that those people got years of counseling from actual professionals, which is not what most dieting looks like.
Zoë: Going back to that Franz 2007 review with 80 trials and over 26,000 people: people lost about 5 to 8.5 kilos in the first six months, then plateaued at almost exactly six months across basically every type of intervention. At four years out, an average of 3 to 6 kilos was still off. None of the intervention groups regained all the way back to baseline. Not a single one. The advice-only and exercise-alone groups barely lost anything at any time point, which is its own conversation: shockingly, when you tell people what to do and don't help them do it, nothing happens.
Zoë: The catch is in the methodology section. Those numbers describe people who completed the study. Attrition was 29% at one year and 31% overall. Roughly a third of people walked out of those trials and we have no idea what happened to them. Everyone who stayed in is what's reflected in that average. Attrition is a huge problem in this field.
Kylee: And most people in the general public aren't digging into the methodology to check the attrition rate.
Zoë: They have way better things to do. We are fundamentally broken in that we like to read the methodology section. My toxic trait is that whenever someone cites a study in normal conversation, I ask what the methodology was. And they're like, we're at a bar right now. Why are you like this?
Kylee: Let's pull up PubMed at the bar.
Zoë: Someone says half of sea turtles don't get the common cold, and I say, define common cold, define sea turtle, was it observational or a randomized controlled trial? And people are like, get away from me, I was trying to enjoy my beverage at a social function.
So what is the real diet success rate?
Kylee: After going through all that, is there an actual number we'd stand behind?
Zoë: If you define success the way a lot of researchers do, which tends to be keeping off 10% or more of your starting weight for at least a year, the studies land somewhere between 10% and 33%.
Kylee: Not 95.
Zoë: No. Dansinger's trial compared Atkins, Ornish, Weight Watchers, and the Zone diet.
Kylee: That sounds like a bingo game. Diet bingo. Anyone for the cabbage soup diet?
Zoë: You know what we never hear about anymore? South Beach. Where is the South Beach diet in all of this?
Kylee: Do we still hear about the Ornish diet?
Zoë: No, dude. I Googled it and was like, I don't have time. Those all averaged to about 10% keeping off 10% of their starting weight for at least a year. Wing and Phelan put it near 20%. Look AHEAD was around 26%. The stepped-care trial got about 33%. So the honest failure rate is somewhere between two-thirds and 90%. Which is not great, and also not 95. This is a fascinating debunk, because it's not "this is totally wrong and diets rock." It's "this is totally wrong, and diets still aren't great."
Zoë: Every one of those studies gave people counseling, education, ongoing support sessions, and usually an exercise program on top of the diet. That's intensive programming. So that range is closer to a ceiling than a representative average. Joe Schmo with an app, meal prepping from TikTok on Sundays, is going to have a much lower success rate without all that free support.
The Biggest Loser study and metabolic adaptation
Zoë: This brings me to Mr. Kevin Hall and his Biggest Loser study. What do you remember about it? Give me your elevator pitch.
Kylee: The main thing I remember is that a lot of people regained a lot of the weight. And didn't their metabolisms slow?
Zoë: Yes. Come with me. This one gets even more complicated, even by Kevin Hall's own assessment, the longer you look at it and the further you go beyond the headlines.
Kylee: Draw me in.
Zoë: This was a 2016 study from Fothergill, Hall, and colleagues, where 14 of the 16 original Biggest Loser contestants were measured six years after the show. During the competition, they lost about 58 kilos each on average. Six years later, they'd regained 41 of those kilos. What captured the headlines was that their resting metabolism was about 700 calories a day below where they started, and about 500 calories below what you'd predict for people their size. That's the metabolic adaptation we talked about in our calorie counting episode: metabolism isn't some fixed, immutable thing.
Kylee: And that was six years later?
Zoë: Yes, still changed in a measurable way. These people lost weight, regained a lot of it, and their metabolism slowed down. The question is how much those two facts are connected, and that's what the paper set out to test. The metabolic adaptation at the end of the show did not predict who regained the most weight. The paper found no significant correlation between how much weight people regained and how much their metabolism downshifted.
Zoë: Here's the part that gets overlooked, because I have no clue how you'd fit it in a headline. The researchers concluded that metabolic adaptation persists over time and is likely a proportional but incomplete response to contemporaneous efforts to reduce body weight. Proportional means the more weight-reduced you currently are, the more slowdown you have. Incomplete means it doesn't fully cancel out your deficit. Contemporaneous means it's tracking where you are now, not punishing you for diets five years ago. The paper is describing what it costs to stay smaller than your body expects.
Zoë: Kevin Hall revisited the study in 2022, which is one of the reasons he's simply the best. He's doing album reissues with really good liner notes. He suggested the slowdown was the body compensating for how much these people were exercising, because the ones who kept weight off had increased their physical activity by 160%, versus 34% for those who regained. Which, yeah. These people were working out all day with Jillian Michaels on TV. Of course you can't sustain that.
Zoë: That's Hall's interpretation, and it rests on a model of energy expenditure that people are still debating. It's not fully understood. The study was 14 people, and a null result in 14 people is not proof of no effect. It just means this study didn't find one. Other research does find adaptation predicting less weight loss. So this isn't settled, but it is complicated.
Kylee: I see people with their metabolism impacted for sure.
Zoë: What does that look like, particularly for people who've done intense restriction and huge amounts of exercise?
Kylee: A lot of times people don't understand that metabolic adaptation and slowing happen, so they keep thinking they have to eat less and less. And you can't just eat 500 calories a day and survive. Sometimes I'll have someone get a resting metabolic rate test, and we compare their actual RMR to what we'd predict. It's wild how slowed it can be, particularly for people who've been dieting for a very long time.
Zoë: I'm glad you mentioned that, because what's universally agreed upon is that something real happens to your metabolism when you lose weight or under-fuel. The argument is about how much sticks around once you're eating enough and your weight is stable, and whether that's what drives regain. In athletes, a suppressed resting metabolic rate is a recognized clinical sign of under-fueling and relative energy deficiency in sport, or RED-S. If your measured RMR comes in below about 90% of what's predicted for you, that's understood to be a marker of low energy availability.
Kylee: I've seen that for sure. Where it gets complicated is that predicted metabolism still comes from an equation, and genetically some people have faster metabolisms.
Zoë: Long story short, humans aren't math. An active person who's been under-fueling having a slow metabolism is a known, predicted thing. The more contested leap is from there to "your dieting broke something permanent, independent of where you are and how you're eating now." The version of the study that keeps getting repeated is that the show broke these people's metabolisms, and that's why the weight came back. That second half is a causal claim. The paper went looking for that causal relationship and didn't find it.
Hunger hormones, leptin and ghrelin, and why diets expire
Kylee: So if it's not the metabolism, what else could it be?
Zoë: Appetite. In that 2011 New England Journal of Medicine study, a full year after a 10-week low-calorie diet, people's hunger and fullness hormones were still shifted toward driving them to eat more. Ghrelin was up, leptin was down, and a bunch of other biomarkers went right along with them. Ghrelin makes you feel hungry. Leptin tells your brain you have enough energy stored. After weight loss, one goes up, the other comes down, and they tend to stay there. That's not willpower. That's endocrinology.
Kylee: And that persisted a year after dieting. Wow.
Zoë: So the mechanism is less about metabolism and more that your appetite got louder and stayed louder for at least a year. Another thing that gets glossed over is that diets have an end date. They expire. That's the design. You start Monday, you're done in so many days, there's a before photo and an after photo. These are not maintenance plans. Do you see that with athletes, that they have no plan for what comes after?
Kylee: They don't think about it at all. Or it's a perpetual diet: I've got to always be in a huge energy deficit. It's literally a lifetime of restriction. I work with people who have dieted their entire lives, one diet to the next, like the majority of American women.
Zoë: I'm so glad we called the podcast Your Diet Sucks and not A Lifetime of Restriction, which was the runner-up title. That also comes out in the Biggest Loser data. If you take somebody out of their actual life, cook every meal for them, have Jillian Michaels yell at them, and keep them on an elliptical for 11 hours a day, you'll get dramatic results. But once they're back in their life, the results depend on whether any of it was portable. And thank God you cannot scale Jillian Michaels. Big interventions produce big results that require a big ongoing effort no human can sustain.
Zoë: And it goes back to environment. People get stuck in a loop where restriction turns up the volume on their appetite, appetite drives regain, and the industry sells them the next thing. Paleo didn't do it, but maybe keto will. If keto didn't, maybe South Beach or the Zone.
Kylee: Or you're just broken and have no willpower or self-control.
Zoë: When there's an actual endocrine piece here. That's why the leptin and ghrelin research is useful. It tells people that what they've been told about willpower and self-control maybe isn't the thing to focus on.
Rigid vs. flexible dieting
Kylee: There's also research on how people diet, not just whether they diet, right?
Zoë: That's the rigid versus flexible piece, and it's very clear in the literature. Rigid restraint is all or nothing: foods are allowed or off limits, you're on the plan or you've blown it, and any single deviation is pass/fail. That pattern is associated with much worse outcomes, more disinhibited eating, and more of the binge-restrict cycle, and it tends to be what people psychologically prefer. Flexible restraint means you can adjust without the whole structure collapsing. One meal out doesn't mean you're off the bandwagon. That pattern is associated with better long-term outcomes.
Zoë: You and I were both curious whether a certain type of person does better on diets, and the research isn't clear on that. It's more that any type of person does worse on a rigid approach and better on a flexible one, again as defined on terms we're suspicious of. So the thing that predicts failure isn't how hard the diet is. It's how brittle it is. Rigidity is the biggest risk factor, and people are drawn to it like moths to a flame. That's the part diet culture is structurally incapable of selling, because rigidity is the product. That's how you keep people on the hamster wheel, buying things and blaming themselves when it doesn't work. Nobody's building a brand on "it's fine, be flexible, be cool, do the Mediterranean diet." There's no 30-day chill-the-fuck-out challenge.
Why is long-term weight loss so hard to study?
Kylee: Why can't anybody just settle this? Why is it so hard to study?
Zoë: To get kinky with our methodology, the study you'd need is basically impossible. You'd have to randomize thousands of people, follow them for a decade, keep them from dropping out, and weigh them instead of asking them. Pick any two and the whole thing falls apart. And attrition cuts in an unexpected direction: the people who regain are typically the people who stop showing up.
Kylee: They probably feel shame. You're not excited to go to your doctor and show them how you're doing on these arbitrary terms.
Zoë: Terms decided for you by culture, and by science in the 1950s. So whoever you can still measure at year five is disproportionately the group it worked for. That biases the whole literature toward optimism. The published numbers are likely too good, not too harsh, and most researchers acknowledge that in the fine print. It doesn't mean we throw the research out, but it's a wrench worth throwing in. Some older studies handled dropouts by carrying the last weight forward, which assumes the person who disappeared froze in place at their lowest weight. That's not how it works. Self-report is also extremely unreliable.
Kylee: People underreport.
Zoë: They underreport their weight and way underreport what they ate. They tend to be aspirational. It's extremely consistent. It's a blunt tool, but it's one of the only ones we have, unless you're Kevin Hall and you lock everyone in the NIH for two weeks.
Zoë: To your earlier point, there's not even agreement on what a diet is. Whole30, a medically supervised very-low-calorie protocol, and "I stopped drinking soda" all get averaged into the same bucket, even though they're wildly different interventions. The Zone diet and "I'm trying to avoid being prediabetic" aren't the same thing. No one agrees on what failure means. And there's no money here. No one wants to fund a 15-year trial with nothing to sell at the end. Most of the long-term data we have exists because it got attached to a diabetes question someone wanted answered, and there are challenges in taking research from a specific, metabolically unhealthy population and asking how it applies to Zoë and Kylee. It's not a choice between a good number and a bad number. It's a deeply flawed number or no number at all, and we tend to take the flawed number.
The National Weight Control Registry
Kylee: What about the National Weight Control Registry? That gets cited on the other side.
Zoë: It was started in 1994 by Rena Wing and James Hill, and it contains thousands of people who lost 30 pounds or more and kept it off. The average loss was around 70 pounds, and average maintenance was about five and a half years from enrollment. It's useful for one question: what maintenance looks like when it does work. Most people in the registry said they changed how they ate and increased activity. Most weigh themselves regularly, and most eat breakfast.
Kylee: We're going to have to do a whole episode on breakfast. And the weighing themselves regularly thing. Okay.
Zoë: But the registry can't tell you how common maintenance is, because it only recruited people it worked for. If you only interview lottery winners, the lottery looks like a surefire financial plan. It's the dark twin of the 1959 study. One only looks at people who failed with no support. The other only looks at people who succeeded. Both start with the outcome, which shapes the conclusion. Neither gives you anything I'd base life advice on, but both get quoted as if they do.
What about GLP-1s like Ozempic?
Kylee: One thing I keep thinking about is the rise of GLP-1s. How does that intertwine with this?
Zoë: In a big way, though it's hard to say exactly how much because these drugs are newer and the science isn't settled. GLP-1s are the class of drugs like semaglutide and tirzepatide that mimic a gut hormone involved in appetite and blood sugar. The brand names are Ozempic, Wegovy, Mounjaro, and Zepbound. They have their own version of the same problem: when you stop taking them, the weight tends to come back. If you stay on them, the weight loss tends to be stable. We just don't have great long-term data yet. It does complicate the picture.
Kylee: These drugs were developed for blood sugar control initially. If somebody goes off them, their blood sugar is probably going to go back up.
Zoë: Which is how all medication works. If I stop taking an antidepressant, my brain goes back to the way it was. If you stop taking blood pressure medication, your blood pressure goes back up. That's how medication for chronic conditions works. I don't think it's a point for or against.
Kylee: So maintenance after GLP-1s still hasn't been solved. And it brings up questions about access and affordability, because these drugs aren't cheap and you or your insurance, if you're lucky, have to keep paying in perpetuity. When I ask people on them what they'll do when they stop, they say, I don't know, I haven't thought about stopping.
Zoë: And I don't know that we have a good answer either. So far, the data says that as long as people stay on them, body composition tends to remain pretty stable. It's a tension that's unresolved.
What to measure instead of weight
Zoë: The flaw this whole question assumes is that body weight is the holy grail of health. Everyone has a number and a jean size where they're sure they'd be better, happier, faster, healthier. You're not a fan of the scale as the be-all and end-all. If we acknowledge that logic is flawed, what outcomes do you want people to focus on?
Kylee: It depends on your definition of health. There's mental health, there are internal blood biomarkers, and people in the weight loss conversation will point to studies showing that when people lose weight, chronic disease risk goes down. With athletes, understanding why someone wants to reach a certain weight is really important. Someone who says, I just need to lose these last five pounds to run faster, is a different conversation. So first I ask: who are we talking to, what are we talking about, and what are your reasons for wanting to reach a certain number? Then, what if we focused on internal health markers? What could this do to your mental health? Have you thought about those things before we get into weight loss?
Zoë: I'm always interested in what decisions people would make if weight loss were taken off the table as an outcome. How much would you still want to exercise? How would you still want to eat? It might be totally valid: I have concerns about my blood sugar, or my energy levels. Those are real outcomes. But time and again, the research measures one thing that's an imperfect proxy for health at best, and it's susceptible to people grafting their own assumptions onto it. A 40-pound bucket with no conversation about where you started makes zero sense. For women, or people with smaller frames, that might be an impossible amount of weight.
Kylee: And understanding the reasoning matters, because a lot of people in fitness and athletics are doing it for "five pounds will make me faster," or vanity. Is that the right way for you to be thinking about this?
Zoë: On the other hand, the way I often hear this zombie statistic weaponized is "don't diet," and I'm not super comfortable with that either. If we disown the statistic, what are other reasons not to diet, the way "diet" is defined here?
Kylee: Like I was saying: living life, enjoying food, and having good mental health.
Zoë: I wanted to hold the statistic up to the light because so many assumptions on both sides get pinned on bad statistics. If we wiggle loose some of the baseline assumptions in this cultural Jenga tower, what remains?
Bottom line: do 95% of diets fail?
Kylee: So what should people walk away with?
Zoë: Throw the 95% out the window. It is not good science. So many things are nuanced and "it depends." This doesn't depend. I'm good on disowning 95%. It goes back to a 1959 case series of 100 patients who got handed a sheet of paper, a quarter of whom never came back. 95 isn't even in that paper, and the guy who wrote it told the New York Times in 1999 that people were stretching it way past his data. We just didn't move on.
Zoë: Using a serious definition of success, somewhere between 10% and 33% of people who undergo a specific diet keep off a meaningful amount of weight long term. The failure rate is somewhere between two-thirds and 90%, even with a lot of support. That's still real bad. It's not 95%, but it's not a hopeful narrative. And much of what looks like total failure is actually partial regain that gets scored as zero. 80% of the weight coming back is not 80% of people ending up where they started or worse, which was the prejudice I brought to this conversation.
Zoë: The mechanism also isn't a broken metabolism. It's a more complicated story. And what predicts whether weight loss holds up isn't whether it's Mediterranean, keto, or the Zone. It's the brittleness of the plan. Rigidity is the failure point, and it's the point people love.
Zoë: None of that means you should diet, and none of it means you should tell everyone not to diet. This is not a prescriptive podcast. I just want to talk about Kevin Hall and get paid to do it. What it does mean is that the number people use to launder their opinions is something we should all walk away from. If a program opens its pitch with "95% of diets fail," ask the question. They likely don't know. Whenever someone throws a statistic at you, pause and say: define failure, define diet, define percent. Go to the methodology section. End the conversation by asking your friend to show you the methodology. Because a statistic you can't source and can't explain isn't evidence. It's a bumper sticker.
Zoë: Thanks for hanging out with us. Your Diet Sucks is hosted, researched, and produced by me, Zoë Rom. Kylee Van Horn is my amazing co-host and researcher, and the reason we get the science right. Cover art is by the wonderful Cleo Rom. Our partnerships are handled by Kelly Newlon at RAD Boulder, who finds us sponsors who actually share our values. And our story editor, Emma Zimmerman, makes every episode sharper, smarter, and more fun. If you want more of us, come find us on Patreon at patreon.com/YourDietSucks. New episodes every other week. In the meantime, go eat something. Bye.
Where does the “95% of diets fail” statistic come from?
We trace it from Albert Stunkard’s 1959 paper through decades of bad-science telephone, NIH panels, and a New York Times debunk that changed nothing. Then we ask the annoying questions. Define fail. Define diet. Define over what period. Because “80% of the weight comes back” and “80% of people regain everything” are very different claims.
What long-term weight loss studies actually show
Kylee and I dig into the Look AHEAD trial, a review of 80 weight loss trials, and Kevin Hall’s famous Biggest Loser study. That study is far more complicated than the “broken metabolism” headlines suggest, which we started unpacking in our calorie counting episode. We also cover why hunger hormones stay shifted a year after dieting and why rigid diets fail more often than flexible ones. Then there’s the National Weight Control Registry, and where GLP-1s fit into all of it.
The bottom line: do 95% of diets fail?
No. Using the standard research definition of success, keeping off 10% of your starting weight for at least a year, somewhere between 10% and 33% of people succeed, even with intensive support. A failure rate of two-thirds to 90% is not great. It’s also not 95. A statistic you can’t source isn’t evidence. It’s a bumper sticker. Find this and the rest of the archive, including our episode on diet cults, at yourdietsuckspodcast.com/your-diet-sucks-podcast.
If you want to keep going, come hang out with us on Patreon. That’s where Kylee’s fueling masterclasses and our bonus episodes live, along with the most scientifically literate comment threads on the internet.
References
Anderson, J. W., Konz, E. C., Frederich, R. C., & Wood, C. L. (2001). Long-term weight-loss maintenance: A meta-analysis of US studies. American Journal of Clinical Nutrition, 74(5), 579–584.
Dansinger, M. L., Gleason, J. A., Griffith, J. L., Selker, H. P., & Schaefer, E. J. (2005). Comparison of the Atkins, Ornish, Weight Watchers, and Zone diets for weight loss and heart disease risk reduction: A randomized trial. JAMA, 293(1), 43–53.
Fothergill, E., Guo, J., Howard, L., Kerns, J. C., Knuth, N. D., Brychta, R., Chen, K. Y., Skarulis, M. C., Walter, M., Walter, P. J., & Hall, K. D. (2016). Persistent metabolic adaptation 6 years after “The Biggest Loser” competition. Obesity, 24(8), 1612–1619.
Franz, M. J., VanWormer, J. J., Crain, A. L., Boucher, J. L., Histon, T., Caplan, W., Bowman, J. D., & Pronk, N. P. (2007). Weight-loss outcomes: A systematic review and meta-analysis of weight-loss clinical trials with a minimum 1-year follow-up. Journal of the American Dietetic Association, 107(10), 1755–1767.
Fritsch, J. (1999, May 25). 95% regain lost weight. Or do they? The New York Times.
Hall, K. D. (2022). Energy compensation and metabolic adaptation: “The Biggest Loser” study reinterpreted. Obesity, 30(1), 11–13. https://doi.org/10.1002/oby.23308
Hall, K. D., & Kahan, S. (2018). Maintenance of lost weight and long-term management of obesity. Medical Clinics of North America, 102(1), 183–197.
Jakicic, J. M., Tate, D. F., Lang, W., Davis, K. K., Polzien, K., Rickman, A. D., Erickson, K., Neiberg, R. H., & Finkelstein, E. A. (2012). Effect of a stepped-care intervention approach on weight loss in adults: A randomized clinical trial. JAMA, 307(24), 2617–2626. https://doi.org/10.1001/jama.2012.6866
Look AHEAD Research Group. (2014). Eight-year weight losses with an intensive lifestyle intervention: The Look AHEAD study. Obesity, 22(1), 5–13.
Stunkard, A., & McLaren-Hume, M. (1959). The results of treatment for obesity: A review of the literature and report of a series. AMA Archives of Internal Medicine, 103(1), 79–85.
Sumithran, P., Prendergast, L. A., Delbridge, E., Purcell, K., Shulkes, A., Kriketos, A., & Proietto, J. (2011). Long-term persistence of hormonal adaptations to weight loss. New England Journal of Medicine, 365(17), 1597–1604.
Westenhoefer, J., Stunkard, A. J., & Pudel, V. (1999). Validation of the flexible and rigid control dimensions of dietary restraint. International Journal of Eating Disorders, 26(1), 53–64.
Wing, R. R., & Phelan, S. (2005). Long-term weight loss maintenance. American Journal of Clinical Nutrition, 82(1 Suppl), 222S–225S.
