Post-GLP-1 Motivation Design: The 2026 Reset

A friend told me a story recently that I have not been able to stop thinking about. She had ridden her Peloton four times a week for three years. In January 2025 her doctor put her on Wegovy. By July she had lost 32 pounds. By October she had stopped opening the app. Not because it broke. Not because she got bored. Because the whole reason she had opened it for three years, which was the low hum of anxiety that she might be gaining weight back, had been chemically switched off. When the anxiety left, so did the ritual it had been feeding.

She is not an outlier. She is the market.

Weight Watchers ended 2024 with 3.34 million subscribers and closed 2025 with 2.76 million, a 17 percent drop in a single year, and filed a “pre-packaged” Chapter 11 restructuring on its way to relaunching as a clinical-plus-behavioral GLP-1 hybrid (SEC 8-K, FY2026; WW Q1 2026 update). Peloton has publicly declared the pivot from cardio to strength, with 2 million members already engaging in resistance training every quarter, because GLP-1 users need to preserve lean mass and the bike alone will not do it (Athletech News, 2026; Peloton 2026 Roadmap, The Clip Out). MyFitnessPal shipped GLP-1 medication tracking as a native feature in April 2026 (MyFitnessPal press release). Noom now sells its own microdosed compounded semaglutide starting at $491 a month (Noom Med).

The demand curve for the entire “help me eat less and move more” industry got rewritten in about 30 months, and most designers I talk to are still working from the old assumptions. This piece is my working map of post-GLP-1 motivation design: what the drug did to the eight core drives, which product categories need to redesign, and where the durable design surfaces still live.

This piece is my attempt to say out loud what the new assumptions are, using the framework I know best. If you have followed my work on the 8 Core Drives, the Octalysis Score, and the White Hat versus Black Hat split, you already have the tools. What you may not have is a map of which of those tools still cut and which ones now bounce off a chemically different user. That map is what I want to draw here.

§1. What GLP-1 actually did to motivation

The popular framing is that GLP-1 medications “make willpower easier.” That framing is wrong in a way that matters, because it treats motivation as a single dial. Motivation was never a single dial. It was the shape of an octagon. And what GLP-1 did was reach into that octagon and turn some drives all the way down while leaving others untouched or actively amplified.

Let me go drive by drive, the way I would score any experience (Octalysis scoring method).

Core Drive 6, Scarcity and Impatience. For a person losing weight the old way, food scarcity was the whole game. Every 3 PM sugar craving, every “just one more chip” moment, every late-night raid on the fridge. That was CD6 operating on the body itself. Semaglutide and tirzepatide bind to receptors in the hypothalamus and satiety centers of the brain and functionally silence that signal. The person is not resisting more. There is less to resist. If you designed a product whose engagement loop depended on the user’s daily hunger anxiety (Noom’s calorie budget, MFP’s food logging as vigilance, a bariatric coach’s “let’s talk about your urges”), that loop is now firing into empty air for 15 percent of American adults (Forbes, 2026; Medscape: 1 in 9 Americans on GLP-1s, 2026).

Core Drive 8, Loss and Avoidance. This drive gets louder. Before the drugs, “I might gain it all back” was one anxiety among many. On the drugs, it becomes the central anxiety, because everyone on GLP-1s has seen the data. Discontinuation studies show semaglutide users regain roughly 0.8 kg per month after stopping, with projected return to baseline in about 18 months (BMJ 2026 meta-analysis via TCTMD; Oxford, 2026). The user knows this. The user has read the same articles you have. CD8 in a GLP-1 user is no longer “fear of getting fat.” It is “fear of losing this new self.” That is a different emotional shape entirely, and it opens a different design surface.

Core Drive 2, Development and Accomplishment. The traditional weight-loss product ran on CD2 through the pride of mastery over cravings. “I resisted the donut.” “I stayed under my calorie budget for 30 days.” “I did the hard thing.” On a GLP-1, that entire achievement graph flattens. There is no craving to resist. Beating hunger stops feeling like a victory when hunger has already been benched. CD2 in this population needs a new object of mastery, and it is not going to be food discipline.

Core Drive 4, Ownership and Possession. Weight-loss apps have always leaned on CD4 through avatars, streaks, and progress bars that the user comes to feel are theirs (From Wanting to Protecting: The Ownership Lifecycle). On GLP-1 the deepest CD4 target shifts from the app to the body itself. The person becomes acutely aware they now own a body that behaves differently, and they want to protect the version of themselves the drug produced. That is a much richer CD4 hook than a green streak dot, if a designer knows how to reach for it.

Core Drive 5, Social Influence and Relatedness. Weight Watchers meetings were CD5 masquerading as a nutrition service. The accountability was the product. GLP-1 users often feel unable to fully belong in that room, because they suspect their weight loss will be dismissed as “not real.” They also generate a different social hunger: for other GLP-1 users, for people who understand nausea onset days, for people who lost 60 pounds and want to talk about identity. Existing communities do not serve them well. New ones will.

Core Drive 7, Unpredictability and Curiosity. For anyone whose relationship to food was compulsive, GLP-1 flattens the unpredictable dopamine spike of the binge. There is more emerging evidence that GLP-1 signaling modulates dopamine circuitry in the ventral tegmental area, dampening reward salience across the board (Psychiatric effects of GLP-1 receptor agonists, 2026; Psychology Today, 2026). Some users report parallel drops in alcohol interest, shopping compulsion, and other CD7-driven habits. This is a genuine gift for people whose CD7 was pathological. It is a design problem for products whose engagement depended on it.

Core Drive 1, Epic Meaning and Calling; Core Drive 3, Empowerment of Creativity and Feedback. These two are largely unaffected. The pharmacology does not touch a person’s sense of purpose or their appetite for agency. Which means they become disproportionately important in the new mix, because the two intrinsic Right Brain drives are now the most reliable engines you have.

The one-line summary: GLP-1 did not lower motivation. It rebalanced the octagon. CD6 shrank, CD7 dimmed, CD8 got sharper and shifted objects, CD2 and CD5 need new targets, CD4 moved from the app to the body, and CD1 and CD3 became more valuable by contrast. Any product still designed for the pre-GLP-1 octagon shape is now missing on multiple drives at once.

§2. The behavior-design assumptions GLP-1 broke

For 40 years the field of behavior change ran on a shared premise: the user’s body was working against them, and design existed to help them win that fight. The MyFitnessPal calorie log, the Noom red/yellow/green traffic light, the Weight Watchers point, the fitness tracker’s nagging step count. All of these assumed a hungry, resistant, self-sabotaging organism that needed to be tricked, gamified, or shamed into compliance. Design was the crutch that made discipline possible.

GLP-1 changed the chemistry. The organism is no longer fighting.

That is the reset. And it forces a set of questions that most behavior-design shops have not yet answered.

Question one: if the daily anxiety about food is gone, what is the product’s daily open hook? Weight-loss apps have historically opened at meal times because that is when the fight was happening. Take the fight away and the meal-time open goes with it.

Question two: if compliance with a plan is no longer the primary Desired Action (Mastering Desired Actions), what is? “Take your Wednesday shot” is a Desired Action a robot could do. Does the product exist to protect that shot, to layer strength training on top of it, to prepare the user for eventual off-ramp, or to help them build a new identity? Products that do not pick will drift.

Question three: what happens when the user off-ramps? A design that only functions while the user is on the drug is a design that ends. If you want the relationship to survive discontinuation, you have to be building something durable inside the window, and most apps are not.

Question four, which is the deepest one: if half of the “unhealthy defaults” your product was built to counteract are now pharmacologically absent, does the product still deserve to exist in its current form? For a lot of legacy weight-loss brands, the answer is no. That is a hard sentence, and I do not write it lightly. I write it because the market is already saying it out loud. WW’s subscriber base fell 17 percent in a year. That is what “the product no longer deserves to exist in its current form” looks like from the P&L side.

There is a corollary here that behavioral designers keep missing. When you remove one drive from an experience, you do not simply subtract engagement. You destabilize the whole octagon, because the drives were compensating for each other. In a CMP-style strategy dashboard, you would say the Business Metric shifted, the Player shifted, and the Desired Actions shifted, so the Feedback Mechanics and Rewards have to be rebuilt from scratch. There is no version of this where the correct move is to keep the same product and add a “GLP-1 mode” toggle. That is decoration, and users can tell.

§3. Categories that need to redesign

Let me walk through the categories most exposed to this shift, in decreasing order of how urgently their teams should be re-reading their own strategy docs.

Traditional weight-loss subscription apps. Noom, WW’s classic app, and the middle tier of coach-plus-plan services. This is the bullseye. Noom’s response has been the most instructive: rather than defend the behavioral-only model, they built Noom Med, acquired the prescribing capacity, and now sell microdose compounded semaglutide with their coaching layered on top (Noom microdosing overview). They also released observational data claiming Noom-supported users regained 21.2 percent of lost weight at 12 months post-discontinuation, versus 61.4 percent in published benchmarks, roughly a threefold reduction in modeled regain (Noom in the news, 2026). Whether those numbers hold up under RCT scrutiny is another question. As a positioning move, it is exactly the right play: sell the pharmacology, own the transition, be indispensable at the off-ramp.

Fitness apps and connected hardware. Peloton, Nike Training Club, Apple Fitness+, Strava. The reflex was that GLP-1 would kill them, because users no longer needed exercise to lose weight. What the data actually shows is subtler. A 2026 Endocrine Society study found that GLP-1 users’ daily steps dropped from 5,047 to 4,487 and moderate-to-vigorous activity fell from 28 to 22 minutes per day after initiation (Endocrine Society, 2026; Healio, 2026). The people on the drug are exercising less than before. That is a signal fitness brands can go either direction on. Peloton picked the smart one, which is to reframe their value proposition around resistance training and lean-mass preservation rather than calorie burn. If you strip the “burn fat” logic out of a bike class, you are selling cardiovascular fitness and enjoyment of movement, which are Right Brain, White Hat, and much more durable.

Nutrition tracking. MyFitnessPal and Cronometer. These apps have a legitimate second life, because a GLP-1 user eating roughly 30 percent less food faces a genuine micronutrient risk. The question flips from “am I eating too much” to “am I getting enough protein, iron, B12, calcium.” Cronometer, which was already the nutrient-density specialist, is positioned unusually well here. MyFitnessPal is playing catch-up with its GLP-1 Support feature, which adds medication logging, side-effect tracking, and injection-site rotation to the food diary. The intent shift matters: this stops being a compliance app and starts being a defensive-nutrition app.

Diabetes management. Different animal, and this is worth naming. Type 2 diabetics have used GLP-1s for years, and the design surface there is not about weight loss identity. It is about glycemic control, medication adherence, and complication avoidance. The category has continuity. What is new is the crossover: the Type 2 patient who is now on the same molecule as a colleague using it for weight loss and does not know whether to feel included or otherwise about it. Diabetes apps that acknowledge the cultural shift will do better than ones that pretend it is not happening.

Metabolic health and longevity. Levels, Function, Prenuvo, the Peter Attia audience. This was a niche category two years ago and is about to become much larger, because a lot of the newly-thinner users start asking deeper questions once the acute weight problem is solved. “What is my HbA1c.” “What is my VO2 max.” “How do I add years, not just subtract pounds.” A product that can carry a user from weight loss into healthspan is a product that survives long past the injection.

If your product falls in any of these categories and the last strategy offsite did not include the words “our user is now chemically different,” you are late.

§4. What still needs designing (the durable stuff)

Here is the part the doomsayers miss. GLP-1 solved exactly one problem, which is the acute biological pull toward overeating. It did nothing for the following list, and in many cases it made the following list matter more.

Muscle building and lean-mass preservation. Meta-analyses of GLP-1 trials show 20 to 30 percent of total weight lost is lean mass, with the risk concentrated in older adults and those with baseline sarcopenia (Prokopidis et al., BJP 2026; Medscape, 2026). Clinical guidance now routinely includes 1.2 to 1.6 grams of protein per kg of body weight and progressive resistance training. Any product that helps a GLP-1 user hit that spec, without treating them like a bodybuilder, has a real job to do. Design surface: CD2 gets a new object of mastery (strength progression), CD3 gets a real feedback loop (visible muscle, load progression), CD4 gets a physical asset the user is now protecting.

Cardiovascular fitness. VO2 max and resting heart rate do not improve on their own when a user loses weight through a receptor agonist. They improve through cardio, period. GLP-1 users need this more than the general population, because they are simultaneously less active on average (Healio). A cardio app that positions itself around lifespan gain, not weight loss, has a bigger addressable audience today than it did two years ago.

Sleep. GLP-1 does not fix your sleep. Your sleep still fixes you or does not. Products in this category have been under-invested because they do not have the acute urgency of weight loss, and that changes now, because a person who has fixed the acute weight problem starts looking for the next lever. The Happiness Equation has a lot to say about which levers actually move the needle. Sleep is one of the biggest.

Mental health. Emerging evidence on GLP-1 and mood is mixed and important. A Lancet Psychiatry cohort study found no signal of worsened depression or self-harm in patients with pre-existing psychiatric conditions (Lancet Psychiatry, 2026), and one 2026 analysis found a 42 percent lower risk of worsening mental illness among users (Healthline, 2026). But the reward-dampening effect that helps with binge eating also flattens some of the joy people got from other sources. Clinicians report patients months into treatment developing depression-adjacent symptoms because food had been managing emotions the medication now suppresses without addressing (APA Monitor, 2025). Products that help users notice and re-populate the emotional real estate that food used to occupy have a real new opening.

Community and meaning. Core Drive 5 and Core Drive 1 never went away, and they were always the two drives most under-served by legacy weight-loss products. A weight-loss app that made a user feel purposeful and connected as a byproduct was, by definition, better than the ones that made them feel disciplined. In a post-GLP-1 world, community and meaning are no longer nice-to-have flavors on top of a compliance loop. They are the loop.

Sustained joy of movement. This is the one I care about most, because it maps directly to the play-based-weight-loss thesis I wrote about last year. Long before GLP-1, my complaint with the fitness category was that it treated exercise as punishment for eating. Anyone raised on games knows that a well-designed activity has its own gravity. You do not need to be bullied into playing chess or into VR boxing. You need to be shown the door and left alone. Now that the acute weight problem has a pharmacological answer, the fitness category can stop being punitive and start being playful. That is a much better business, and it is a much better product.

Each of these behaviors sits high on the White Hat side of the octagon: intrinsic, self-controlled, and durable. Which is exactly the shape of the design surface that pre-GLP-1 weight-loss products systematically avoided, because White Hat is worse at converting today’s user and better at keeping them for a decade.

§5. New product categories that make sense

A whole shelf of products has room to exist that did not have room three years ago. Some are already being built. Some are still open.

GLP-1 adherence and side-effect coaching. More than half of patients who discontinue GLP-1s restart within a year (Endocrine Society, 2026), and the primary reasons for discontinuation are adverse reactions, cost, and non-adherence (JMCP, 2024). This is a CD8 design surface. The user has lost 40 pounds, wants to keep them, and is afraid the nausea or the price will force them off. A product that helps them time doses, manage side effects, find cheaper suppliers, and stay on until they choose to leave is doing real work. The framing here has to be careful, because CD8 done crudely creates the loss-aversion loops I have been critical of for a decade. Done well, it is a White Hat rescue inside a CD8 frame: “protect the version of yourself you built.”

Post-weight-loss identity products. This is the category I think is most under-built. People who lose 60 pounds do not just have new bodies. They have new social interactions, new clothes, new photos, new confidence, new grief, new relationship dynamics (Psychology Today: Becoming Someone New, 2026). The body changes faster than the self-concept. There is a real opening for products that help someone integrate a new physical identity: journaling, therapy, wardrobe help, community, story-tracking. Design surface: CD4 (ownership of the new body), CD5 (belonging in the new social layer), CD1 (making sense of the transformation), CD2 (mastery of the new state).

Muscle-composition-first fitness apps. Peloton is moving toward this. The clearer play is a product built from the ground up for GLP-1 users who need to hit protein targets, do progressive resistance training three days a week, and see body composition change even when scale weight stays flat. DEXA integration, grip strength as a headline metric, protein intake in the same app as workout logs. The scale becomes a supporting character. Muscle mass becomes the hero.

Longevity-oriented gamification. The healthspan community was small and expensive. GLP-1 gave it a much larger inflow, because users who solved the acute weight problem naturally look for what is next. This is CD1 territory if you design it right. “Add ten good years to your life” reads as epic if it is credibly built into the mechanic, and hollow if it is a mission-statement banner. Waze earned CD1 by having every user actually contribute to real-time traffic data. A longevity app earns CD1 by having every logged workout, meal, and sleep window feed into a real, visible healthspan projection.

Off-ramp products. No one is really selling this well yet. The reality is that most GLP-1 users will eventually come off, either because of cost, side effects, pregnancy, or plateau. What does the app look like that says “let’s build the six months that come after your last shot”? Not a maintenance program that assumes you failed. A structured off-ramp that treats the transition as a designed phase in its own right, with its own Onboarding-Scaffolding-Endgame arc. This is the product I would build if I were starting a company in this space in 2026.

Each of these categories has a distinct octagon. If you are building any of them, the strategy dashboard is a good place to start, because the Business Metric, the Player, and the Desired Actions are all new.

§6. The behavior-design ethics of GLP-1

I want to sit with this one carefully, because it is where a lot of the field will make expensive mistakes.

The instinct in the fitness and diet industry is to treat GLP-1 as the enemy. That instinct is wrong, and it is also going to lose. The drugs work, the obesity rate has already fallen from 39.9 percent in 2022 to 36.4 percent in 2026 (Forbes 2026), and the population of people whose lives have measurably improved on these medications is now in the tens of millions. Any product that positions itself against GLP-1 is positioning itself against its own future customers.

The opposite instinct, which is to build maximum-friction lock-in around the pharmacology, is worse. This is where the ethics get real. A GLP-1 user is already in a CD8-heavy state. They are afraid of regain. They are anxious about the drug supply chain, the price, the potential need to stay on indefinitely. A designer with even average skill can construct engagement loops that exploit that anxiety and produce very high retention numbers. The A/B tests will look great. The user will feel worse.

This is the classic Black Hat trap. I have written before that Black Hat sits in a design category defined by the user feeling out of control, and it is ethical when it is transparent, infrequent, and serves the user’s own stated goals. A short-term Black Hat push toward remembering the Wednesday injection is fine. A chronic loss-avoidance loop that keeps a user opening the app three times a day out of fear is not.

The frequency test I use for e-commerce (Booking.com vs Amazon) applies here too. GLP-1 adherence is a weekly action, not a daily one. The engagement loops around it should respect that cadence. A product that hijacks daily attention to keep the fear alive is doing exactly what critics of the wellness industry have accused us of for years, and doing it inside a population that is more vulnerable to the exploitation than the general one.

There is a specific pattern I want to name and warn against. Call it the shame-then-rescue loop: remind the user daily of what they might lose, then offer the product as the way to protect it. This works. It also breeds the exact dependency the user is trying to escape. If your product needs the user to be afraid every day in order to function, your product is a subscription to fear, and eventually your user will figure that out and leave.

The White Hat alternative is real. It looks like: help the user build genuine strength (CD2 through resistance training), give them meaningful data about their body (CD3 through composition tracking, CD4 through psychological ownership), connect them to other people in the same phase (CD5 through community that respects the medication rather than mocks it), and give the whole enterprise a purpose bigger than the number on the scale (CD1 through healthspan or grandkids or the mountain hike at 60). A user who is engaging with your product because they want to, not because they are afraid to stop, is a user who compounds.

If you are building in this space, pick your side of the White Hat / Black Hat line deliberately. And if you are shipping a product that a user would be embarrassed to describe honestly to their doctor, you are on the wrong side.

§7. The AI × GLP-1 double punch

Here is where the two big shifts of the decade collide.

AI companions and AI-personalized coaching are moving from novelty to substrate exactly as the GLP-1 population reaches critical mass. This convergence is more consequential than either shift alone. The reason is that GLP-1 users need coaching that is highly personalized (protein targets that account for their exact medication dose, side-effect timing, muscle-mass baseline, activity level), delivered in a way that respects that they now have far more mental bandwidth than they used to. AI is uniquely positioned to serve that user because it can afford to remember every detail without the labor cost of a human coach, and it can deliver at any hour with a patient, non-shaming affect.

I have written elsewhere about AI as a layer, not a pillar, meaning that AI does not replace motivation design, it strengthens the three pillars of immersion, ownership, and motivation. Post-GLP-1, that principle sharpens. The user’s baseline motivation on food discipline is not the problem anymore. What they need from AI is not “help me want to do this,” which is what pre-GLP-1 coaching was mostly about. What they need is “help me do it well now that I am willing.” Those are different products.

The single biggest opportunity I see is what I would call the composition coach. An always-on AI that has read every study on GLP-1 muscle preservation, that knows the user’s protein intake pattern from a photo of their plate, that adjusts their resistance program based on grip strength and DEXA scans, that flags the sleep window that torpedoed last week’s recovery, that connects them with a real trainer when the friction warrants it. This is not sci-fi. Every component exists. What is missing is a product-level integration and, more importantly, an octagon-designed motivation layer that keeps the user engaged for the two years it takes to change body composition.

A word of warning here that applies to any AI health product but especially this one. Human QA is non-negotiable. A GLP-1 patient asking an AI about a rare side effect deserves the answer they would get from a well-trained clinician, not the answer the model happens to hallucinate. I have said before that AI does not replace humans, it replaces humans who do not use AI, and that principle is doubly true when the stakes are medical.

The second convergence surface is longevity. An AI coach with rich longitudinal data on a GLP-1 user is generating a personalized dataset that would have been impossible five years ago. If the coach can translate that dataset into a credible healthspan projection (“this week’s choices added 12 days”), CD1 becomes a live wire, and CD2 gets a scoring system that is not vanity but actuarial. The tools to do this responsibly exist now. The market for it is being handed to us by the pharmacology.

§8. What I have come to believe

Three years into watching this play out, here is where I have landed.

I believe GLP-1 is a genuine medical advance and that treating it as cheating is a mistake made by people who have not sat with someone whose life was being wrecked by their appetite. When my friend told me she stopped opening Peloton because she stopped being anxious, she was not describing a loss. She was describing recovery from a low-grade siege that had been running in the background of her life for a decade. Designers who lecture her about the value of struggle are, in my view, in the wrong.

I also believe the drugs will not save behavior design. The population of people on GLP-1s is going to grow. What follows is a set of second-order problems that the drugs do not touch and probably create: how do we help people preserve muscle, sleep well, build community, find meaning, and stay off the sofa now that food anxiety is gone. Every one of those is a motivation-design problem. The framework I have spent 15 years developing, the 8 Core Drives and the White Hat / Black Hat map, is more useful in this environment than it was in the last one, because the drives worth designing for have shifted toward the intrinsic, White Hat quadrant where good design has always lived and cheap design has always failed.

I believe the biggest strategic mistake a behavior-design shop can make in 2026 is to treat this as “another wellness cycle” and wait it out. This is not a diet trend. It is a permanent chemical shift in what half of the market wants. Products built for the old shape are already dying, and the ones being built for the new shape are being built by teams who understand pharmacology and motivation design at the same time. The consulting demand curve for the traditional “help us make our diet app stickier” work is going to keep flattening. The demand curve for “help us design a durable identity product for post-GLP-1 users” is going to steepen fast.

My prediction for 2026 through 2028: the winners in this category will be products that combine a defensive-nutrition or medication-adherence layer (the CD8 rescue), a mastery layer built around body composition or healthspan (CD2 and CD3), a real community of same-phase users (CD5), and a meaning frame that is credible rather than decorative (CD1). The losers will be the products that either ignored the shift or tried to weaponize the fear. There is a middle path here that respects the user, works with the medicine, and builds something the user is proud to be part of. That is the design brief.

Closing

The Peloton user I opened with called me a few weeks later. She told me she had started strength training three times a week, joined a small Discord for women on GLP-1s who were focused on lean mass, and hired a nutritionist to help her stay above her protein floor. She said the interesting part was not that she was doing less than before, but that what she was doing now felt chosen instead of coerced. When the anxiety went away, so did the loop that was masking what she actually wanted. What she actually wanted was to be strong and to be around people who understood the phase she was in.

That is the design brief for the next five years, and it does not require Ozempic to be understood. It only becomes clearer once the acute biological pull is quiet enough for the person underneath to be heard.

If you build products in health, fitness, nutrition, or wellness, the assignment for Monday morning is small and specific. Open your product’s octagon and re-score it for a user who is chemically different from the one you designed for. Ask which drives just went dark, which just got louder, and which you have been under-serving because they were less commercially convenient in the old world. Then rebuild the loop that used to run on the drive that is now missing. That is the whole reset. It will take most teams a year to do it well. It is worth every hour.

WOULD YOU LIKE YU-KAI CHOU TO WORK WITH YOUR ORGANIZATION?