Most people build a workout plan by starting with the exercises, which is the last decision rather than the first. Choose the movements before you know how many days you can train and you get a session that will not fit a Tuesday, a leg day you keep postponing, and no rule at all for what happens when the weight finally moves.
Build it in the order the decisions depend on each other and the whole thing takes about twenty minutes: what the plan is for, the days you will really train, how those days divide the body, which exercises fill them, how many hard sets each muscle gets, and the rule for making it harder next week. ChatRPE is an AI coach built on a Universal Logbook, so you can make those decisions in conversation, run the sessions from the same place, and read back later whether the plan you wrote is the plan you actually ran.
The short version
- Decide what the plan is for first. It barely changes the structure and it changes almost everything about exercise selection and rep ranges.
- Count the days a normal week can hold, not a good week. That number picks your split for you.
- Fill each session with movement patterns, then choose a specific exercise for each pattern. Patterns are the plan; exercise names are swappable.
- Two to four exercises per session covers most people. The first one carries the session, so it goes first while you are fresh.
- Write the progression rule before week one. A plan with no rule for adding load is a list of exercises.
- Put a review date about a month out, and expect to change one variable rather than the whole plan.
Make the decisions in this order
Each of these depends on the one above it, which is why skipping ahead is what turns a promising plan into a plan you quietly stop running.
| Order | The decision | What it settles |
|---|---|---|
| 1 | What the plan is for | Rep ranges, and which lifts get priority when time runs out |
| 2 | Days per week | Which splits are even available to you |
| 3 | The split | What each session is responsible for |
| 4 | The exercises | What you do on the gym floor |
| 5 | Sets and reps | How much work each muscle group gets per week |
| 6 | The progression rule | What changes next week |
| 7 | The review date | When you are allowed to rewrite any of the above |
The dependency runs one way. Session length limits how many exercises fit, exercise count limits how many hard sets you can put on a muscle group in one visit, and weekly sets depend on both. Pick five exercises for a 40-minute session and something gets rushed, and the thing that gets rushed is almost always the rest between the heavy sets, which is the part that was doing the work.
It also runs backwards once the plan exists, which is worth checking before you commit to it. Multiply the hard sets in a session by the length of a set plus the rest you intend to take, add the warm-up, and you have the session your plan actually implies. How long a workout should be walks through that arithmetic and through measuring what your sessions really run, which is usually the fastest way to find out that a plan needs one fewer exercise.
Step 1: name what the plan is for
One goal per plan. "Muscle and strength and conditioning and flexibility" is four plans fighting for the same hour.
Muscle growth wants moderate loads, a fair number of hard sets, and enough food to recover from them, which the muscle building guide covers. Getting stronger at specific lifts wants heavier work on fewer movements, practised often, covered in how to get stronger. Fat loss barely changes the training at all: the deficit does that work, and lifting is there to keep the muscle you have, which is the argument in the best workout to lose weight.
If you cannot say what the plan is for, say what you want to be able to check in twelve weeks instead. That is usually the same answer in more useful clothes, and how to set fitness goals covers turning it into something with a readout.
Step 2: count the days you will really train
Not the days you could train in a clear week. The number that survives a late meeting, a sick child, and a Friday you do not feel like it.
Three real days beat five aspirational ones, because a plan built for five days and run at three is not a three-day plan. It is a five-day plan with two holes in it, and the holes are never in the same place twice. If your week is genuinely unpredictable, the busy-schedule guide covers writing a full, short, and minimum version of each session in advance so the decision is already made when Thursday collapses.
Step 3: divide those days across the body
Two or three days points to full body, four points to upper/lower, five or six points to push pull legs. Aim to train each muscle group about twice a week and the structure argument is mostly settled. For the comparison in full, including how to arrange the days across the week, read the workout split guide.
The split is a container. It says what each session is responsible for and nothing about what you do in it, which is the next decision and the one most plans get wrong.
How to choose exercises for your workout plan
This is the step where people start browsing, and browsing is how you end up with nine exercises for the chest and none for the hamstrings. Fill the session by movement pattern first, then choose a specific exercise for each pattern.
The patterns a plan has to cover
| Movement pattern | Trains | Common choices |
|---|---|---|
| Squat | Quads, glutes | Back squat, front squat, goblet squat, leg press |
| Hinge | Hamstrings, glutes, lower back | Deadlift, Romanian deadlift, hip thrust, back extension |
| Horizontal push | Chest, front delts, triceps | Bench press, dumbbell press, push-up, machine press |
| Horizontal pull | Mid back, lats, biceps | Barbell row, chest-supported row, cable row |
| Vertical push | Shoulders, triceps | Overhead press, dumbbell shoulder press |
| Vertical pull | Lats, biceps | Pull-up, lat pulldown, assisted pull-up |
| Single leg and trunk | Hips, balance, midsection | Split squat, step-up, lunge, carry, hanging leg raise |
A full-body day takes one squat, one hinge, one push, one pull, and something for the trunk. An upper day takes both pushes, both pulls, and arms if there is time. A lower day takes a squat pattern, a hinge pattern, and one single-leg movement. Written that way, a session you have never done before still takes about ten seconds to fill in.
Three questions that pick the specific exercise
- Can you load it in small steps and repeat it? A main lift needs somewhere to go. Barbells and machines with a decent weight stack give you that; an exercise where the only progression is "more reps forever" belongs later in the session.
- Does it fit the equipment you will actually have on a normal day? The squat rack at 6 p.m. is a different piece of equipment from the squat rack at 6 a.m. If your gym has one rack and a queue, plan a second option for the same pattern.
- Does it feel fine for the joints involved? Mild discomfort in a new movement is ordinary. Sharp or persistent joint pain belongs with a clinician before it belongs in a program.
For the four lifts that most plans are built on, the technique guides go deeper than a plan needs to: how to squat, how to deadlift, how to bench press, and how to do a pull-up. Training at home changes the answer to question one, not the pattern list, and the no-equipment guide covers making bodyweight movements harder when you cannot add a plate.
How many exercises a session needs
Two to four for most sessions, and four is already generous for an hour once you rest properly between the heavy sets. The first exercise is the one the session is really about, so it goes first, at full rest, while you are fresh. Everything after it is accessory work that fills in what the first movement missed.
A common shape for an hour: one heavy compound in the 5 to 8 rep range, one second compound from a different pattern, then one or two accessories in the 8 to 15 range. Treat those numbers as ordinary coaching practice rather than exact prescriptions, and read the RPE guide if you want a way to record how hard each of them actually was.
Substitutions are part of the plan, not a break from it
Because the plan is written as patterns, a swap is obvious rather than improvised. The rack is taken, so the goblet squat covers the squat pattern today. Your shoulder dislikes flat pressing this month, so an incline dumbbell press covers the horizontal push. The session still did its job, and the log still says what you did.
Write the alternates down when you build the plan. Deciding on a substitution while standing in a busy gym is how sessions turn into whatever machine was free.
Step 4: set how many hard sets each muscle gets
Roughly 10 to 20 hard sets per muscle group per week is the range most lifters and coaches work in, with beginners doing well at the bottom of it. Divide that across the sessions the split gives you: four upper/lower days means each muscle group has two visits to spread its sets over. How many sets per week covers the counting, what actually qualifies as a hard set, and what to do when adding more stops helping.
Start lower than you think. A plan you can add to is more useful than one you have to rescue.
Step 5: write the progression rule before week one
This is the decision that separates a program from a workout list, and it is the one most self-built plans skip entirely.
Double progression is the usual default and it fits almost any exercise. Pick a rep range, add reps inside it week to week, and when you hit the top of the range on every set, add load and drop back to the bottom. Three sets of 6 to 8 at 135 becomes three sets of 8, which becomes three sets of 6 at 145. Written down, that rule answers the "what do I do today" question for months without you thinking about it.
Decide the failure case too. Two sessions in a row where the reps go backwards usually means repeating the week rather than pushing, and a third means cutting the load by about 10% and building back. The progressive overload guide covers reading which of those you are in from your own logged sets, and the beginner plan covers the simpler version of the rule for a first program.
Step 6: give the plan a review date
A deadline says when you get judged. A review date says when you are allowed to change things, which stops you rewriting the plan in week two because one session went badly.
Four to six weeks is a sensible first review. At it, change one variable for a stated reason: more sets on a lagging muscle group, a different exercise for a pattern that keeps aggravating something, or a lighter week if the loads have been climbing for a while. The deload guide covers timing that lighter week off your own logs instead of a fixed calendar.
Build the plan in one conversation, then run it from the same place
| The usual workflow | The ChatRPE upgrade |
|---|---|
| Stitch a plan together from five articles and a video | Describe your days, equipment, and goal and work through the decisions in order |
| The plan lives in a notes app you stop opening | The plan, the sessions, and the results share one Universal Logbook |
| Remember mid-set what the progression rule was | The rule is written into the plan and applied to what you just logged |
| Guess whether you ran the plan as written | Ask for a heatmap of your training days and read the answer off your own log |
| Rebuild from scratch when something stops working | Change one variable at the review, with the previous block as the comparison |
ChatRPE handles workout programming as a conversation rather than a form, so the brief is where the quality comes from: how many days, how long a session, what equipment, what you are training for, and anything that hurts. Give it those and the plan comes back with the split, the exercises for each pattern, the set counts, and the progression rule already attached.
Then it runs. Ask for a live session and you get the current exercise, its prescription, what is coming next, and a rest countdown on screen and on the iPhone Lock Screen, covered in the guided workout guide. Sets go in by voice through the microphone in the chat composer, on the web and on Android as well as through Siri on an iPhone, which the hands-free logging guide covers in detail.
The reading-back is what makes the review honest. Because each set is stored with its weight, reps, and RPE, you can ask for the heaviest set of one named lift charted session by session, or an estimated 1RM computed from the same sets, without ever having saved a derived number. Training days chart as a calendar heatmap that counts sessions logged rather than sessions intended. New personal bests get called out as they happen rather than waiting for you to notice.
Tell ChatRPE to remember the plan and the review date and it holds them as durable coaching context outside any single chat, so the check-in four weeks from now arrives framed around the block you are actually running. The goal-setting guide covers how that memory works and what it is good for.
Try this: "I can train four days a week for about an hour, full gym, goal is muscle. Build me an upper/lower plan: pick one exercise per movement pattern, give each muscle group 10 to 14 hard sets a week, and write the progression rule. Then remember it and check in with me in four weeks."
A worked example, decision by decision
One reader, four days, an hour a session, a commercial gym, and a goal of gaining muscle.
| Decision | The answer | Why |
|---|---|---|
| What it is for | Muscle | One goal, so the rep ranges stop arguing with each other |
| Days | 4 | The number that survives a bad week, not the ideal one |
| Split | Upper/lower, twice each | Every muscle group gets two visits |
| Exercises | Upper: bench, row, overhead press, lat pulldown, arms. Lower: squat, Romanian deadlift, split squat, calves | One exercise per pattern, heaviest first |
| Sets | 12 to 14 per muscle group per week | Middle of the usual range, room to add later |
| Progression | Double progression, 6 to 8 on compounds, 10 to 15 on accessories | Answers "what do I do today" without thinking |
| Review | Six weeks | Long enough for the top sets to say something |
Nothing in that table is unusual, and that is the point. A good self-built plan is a boring plan with a rule attached, run for long enough that the logbook can tell you something.
Frequently asked questions
How do I make my own workout plan?
Work through the decisions in the order they depend on each other. Name what the plan is for, count the days you will genuinely train, pick the split that fits that number, fill each session with one exercise per movement pattern, decide how many hard sets each muscle group gets per week, and write the rule for adding load before you start. Then set a review date four to six weeks out. Skipping straight to the exercise list is the single most common reason a self-made plan falls apart by week three.
How many exercises should be in a workout plan?
Two to four per session for most people, which usually works out to somewhere between eight and sixteen distinct exercises across a whole week depending on the split. More than four in an hour means either the rest periods or the last exercise gets sacrificed. Put the heaviest compound first while you are fresh, and treat everything after it as accessory work that fills in what the first movement did not cover.
How do I choose which exercises to do?
Choose the movement pattern first, then the specific exercise. Every plan needs a squat, a hinge, a horizontal push, a horizontal pull, a vertical push, a vertical pull, and something for the trunk, spread across whatever sessions your split gives you. Within a pattern, pick the option you can load in small steps, that will be free when you actually train, and that feels fine on the joints involved. That last one is a real filter rather than a formality, and sharp or persistent joint pain is worth a clinician's opinion rather than a workaround.
What is the easiest way to create a workout routine for a beginner?
Two full-body days a week with one exercise per pattern, three sets each, and a rule to add a small amount of weight whenever the last set moves cleanly. That is a complete plan, and it produces progress for months. The beginner workout plan has the version written out, and how to start going to the gym covers the part that is usually harder than the programming.
Can an app build a workout plan for me?
Yes, and the useful question is what happens after it generates one. ChatRPE builds the plan in conversation from your real days, equipment, and goal, runs the session live with the prescription and rest timer in front of you, records what you actually lifted, and compares the block against the one before it. The AI workout planner guide covers the constraints worth handing over and the progression rules worth demanding.
How long should I stay on the same workout plan?
Until it stops producing progress, which is usually longer than it feels. Six to twelve weeks on the same exercises is normal, and changing the plan while the loads are still climbing throws away the only evidence you had. Review at four to six weeks, change one variable with a reason, and let the next block answer whether the change helped. Running a plan in blocks like that is what periodization means in practice, and that guide covers how long a block should be and how to judge the one you just finished.
Do I need a different workout plan to build muscle and to lose fat?
Mostly no, and this surprises people. The training looks close to identical: the same patterns, the same progression rule, roughly the same set counts. Fat loss is decided by the calorie deficit, and lifting during one exists to keep the muscle you already have, so the honest change is to expect the loads to stall rather than climb. Body recomposition covers running both at once, and the calorie deficit calculator covers the number that does the actual work.