Writing a program for someone standing in front of you is easy — you watch them warm up, you see what they can lift, and you fix it in real time. Writing one for a client you have met once on a video call, who trains at 5:40am in a gym you have never seen, is a different job, and most coaches do it by opening the builder and starting to type.
That is where the time goes and where the first block goes wrong. The program is not the hard part. The decisions that come before it are, and there are about six of them.
Collect the six answers before you open the builder
You cannot write a program from a goal. "Lose 8 kg and get stronger" tells you nothing about what goes on Tuesday. Before you write a line, you need six specific answers, and every one of them should come from the intake form rather than from a follow-up message three days later.
| What you need | What it decides |
|---|---|
| Days per week they can train, and which days | Split, session count, how much you can fit per session |
| Time per session | Exercise count, superset use, rest periods |
| Where they train and what is in that gym | Every exercise selection you make |
| Training history and current lifts | Starting loads, complexity, rate of progression |
| Injuries, surgeries, painful ranges | What comes out and what replaces it |
| What they have already tried and hated | Adherence — the most underrated input in programming |
The last one gets skipped and it should not. A client who has quit two previous programs at week three because of 90-minute sessions is telling you the constraint that matters most. If your intake does not ask these, fix the intake — a client intake form built for programming saves you a week of back-and-forth on every signup.
If an answer is missing, ask before you write. Guessing a client's gym equipment and getting it wrong costs you a rewrite and costs them their first session.
Program the week they have, not the week they want
Ask a client how often they can train and most say four or five. Ask what they actually did in the last month and it is two or three. Both numbers are real — one is intention, one is behaviour — and you program the second one.
The practical rule: take the number they give you and subtract one. If they say five, write four. If they say four, write three. A three-day program completed every week beats a five-day program half-done, because the four-of-five client spends every week failing at their own plan, and clients who feel like they are failing cancel.
Then get specific about which days. "Three times a week" is not a schedule; "Tuesday, Thursday and Saturday morning" is. Named days let you sequence sessions properly — you know Saturday follows a rest day and can carry the heaviest work, and you know Tuesday follows a weekend that may have been neither restful nor sober.
Ask about the shape of their week too. A client with Wednesday night kids' sport and a Friday commute is not the same as one with a flat, open week, even if both say three sessions.
Pick the split from the days, not from a philosophy
Once you know the days, the split mostly picks itself. Two or three days a week means full-body or upper/lower — you cannot run a body-part split on three sessions without leaving muscle groups untrained for ten days at a time. Four days is where upper/lower and push-pull-legs both work. Five or six days is where you can split by body part if there is a reason to.
The reason coaches get this wrong is that they run the split they enjoy. If you train six days a week on a bro split, a three-day full-body program feels thin to write, so you compensate by cramming eleven exercises into 45 minutes and the client's first session takes 80.
Session length is the second constraint. A useful working figure is one exercise per five to seven minutes of session time once warm-up is accounted for — so a 45-minute session is four to six exercises, not nine. Count it before you save.
There is more to it than days and minutes — training age, goal and recovery all shift the answer — and choosing a workout split for a client goes through the trade-offs properly.
Select exercises for the gym they are actually standing in
Ask for equipment specifically or you will get "full gym", which means nothing. Ask whether there is a squat rack, whether there are barbells or only fixed machines, what the heaviest dumbbell is, whether there is a cable stack, and how busy it is at their training time.
That last question is the one nobody asks and it changes the program. A client training at 5:30pm in a commercial gym cannot run a program built on the two squat racks. Program a machine or dumbbell alternative for anything likely to be occupied, or accept that they will skip it.
Build each session in this order:
- One primary compound, loaded and progressed deliberately. This is the lift the block is about.
- One or two secondary compounds covering the patterns the primary missed.
- Two or three accessories, chosen for the client's weak point, goal or joint history.
- A warm-up that is specific to the session, not a generic five minutes of cardio.
Then check coverage across the week rather than within a session: squat, hinge, push, pull, carry or core. If the week has no hinge in it, no amount of good session design fixes that.
For a beginner, keep the total number of distinct exercises low — repeating the same movements gives them enough exposure to get technically competent and gives you clean data to progress from. A program with 22 different exercises across three sessions teaches nobody anything.
Set starting loads for someone you have never seen lift
This is the part coaches find genuinely hard, and the answer is that you do not set the loads. You set the effort and let week one tell you the loads.
Three ways to get a starting number, in order of how much you should trust them:
Option three is not a cop-out — it is the correct answer for most new online clients, and it is why RPE exists. "3 sets of 8 at RPE 7, leaving 3 reps in reserve" is a complete instruction that works without you knowing their squat. It also protects you: a client who cannot judge effort yet will land somewhere sane, whereas a prescribed 80 kg can be either a warm-up or an injury.
You do have to teach the scale. A client who has never been asked to rate effort will report everything as an 8. Explain it in the notes, ask for it in week one only on the primary lifts, and expect the numbers to be useless for a fortnight — using RPE with online clients covers the teaching part, and RPE versus RIR is worth reading before you decide which scale to put in front of them.
Whichever you choose, week one is a calibration week. Say so to the client, and then hold your nerve when the loads look light on Tuesday.
Write the notes so the session runs without you
The notes are where an online program is won. In person you cue; online, the note is the cue, and it is the only thing standing between your intent and a client doing tempo squats at half speed for the wrong reason.
Write a note when it changes what the client does. Skip it when it does not.
- What "good" looks like on this set. "Stop when bar speed slows noticeably" rather than "go hard".
- The one technical cue that matters most for this client on this lift. One, not four. Four cues is a paragraph they will not read on their phone between sets.
- What to do when it goes wrong. "If the left knee aches, drop to the goblet variation and tell me." Every prescription should have a named fallback.
- Why the odd-looking thing is there. A client who does not know why they are doing 3-second eccentrics will quietly stop doing them.
Keep each note to a sentence or two and put it on the set it applies to, not in a wall of text at the top of the session. A note the client reads while resting is coaching. A note at the top of the workout is a document.
This is also where the platform stops being neutral. A note attached to the set, an alternate exercise the client can swap to mid-session when the rack is taken, and a video on the exercise itself all mean fewer messages on Tuesday night — Fitsly's training tools put those on the set the client is actually looking at, and once a program is on a client's calendar it is their own copy, so tidying up your library later never rewrites a session they are mid-way through.
Make the first block four weeks, not twelve
Twelve-week programs get sold because they sound like a commitment. They are the wrong length for a first block, because everything you assumed at week zero is a guess and you will not know which guesses were wrong until week two.
Four weeks is the working default: long enough for the client to repeat sessions and for loads to mean something, short enough that a bad assumption costs you a month rather than a quarter.
| Block length | Suits | Risk |
|---|---|---|
| 2 weeks | Complete unknowns, returning-from-injury clients | Too short for progression to show |
| 4 weeks | Almost every new online client | Very little — this is the default |
| 6–8 weeks | Established clients with a known training response | Wrong assumptions run for two months |
| 12 weeks | Nobody, as a first block | The client outgrows or abandons it by week five |
Within those four weeks, progress one variable at a time — usually load on the primaries and reps on the accessories — so that when something stalls you know what stalled. Progressing a client's exercises covers the week-to-week decisions.
At the end of the block you rewrite, but not from scratch: you keep the exercises that worked, swap the ones they hated, and adjust the volume based on what they actually completed.
Worked example: one brief, one week
The brief, straight off an intake form:
Priya, 34. Two kids, office job. Trains at a 24-hour gym before work. Says four mornings, has managed two or three most weeks this year. Forty-five minutes, hard stop. Lifted at university, nothing structured for six years. Wants to be stronger and lose about 6 kg. Right shoulder aches overhead, no diagnosis. Full commercial gym, quiet at 5:40am. Quit her last program because sessions ran over an hour.
The decisions that follow from it: three days, not four. Full-body, because three days. Tuesday, Thursday, Saturday. Five exercises per session at 45 minutes. No overhead pressing in block one — horizontal and incline pressing instead, and a note asking her to report the shoulder. Loads by RPE, because there is no recent data worth trusting. Four-week block.
| Tuesday | Thursday | Saturday | |
|---|---|---|---|
| Primary | Goblet squat 3×8 @ RPE 7 | Trap-bar deadlift 3×5 @ RPE 7 | Goblet squat 3×8, +2.5 kg on Tuesday |
| Secondary | Dumbbell bench press 3×10 | Chest-supported row 3×10 | Incline dumbbell press 3×10 |
| Secondary | Half-kneeling cable row 3×10 | Dumbbell RDL 3×10 | Lat pulldown 3×10 |
| Accessory | Split squat 2×10 each side | Cable face pull 2×15 | Hamstring curl 2×12 |
| Accessory | Dead bug 2×8 each side | Suitcase carry 3×30 m | Side plank 2×30 s |
Five movements, three of them repeated across the week, no barbell back squat in week one because nobody has watched her squat yet. It fits in 45 minutes with three minutes on the primary and 90 seconds elsewhere, and it produces a usable picture of her by Saturday.
Hand it over so week one is quiet
You have written a good program. It can still generate twenty questions on Monday night, and every one of those questions is a five-minute reply you did not budget for.
Send three things before the first session, not after it:
- A short walkthrough of the week. Which days, roughly how long, what the point of the block is. Two minutes of video beats a page of text.
- The rules for week one. It is a calibration week, the weights will feel light, log everything anyway, and tell me if a session runs over 45 minutes.
- One named escape hatch. What to do if the rack is taken, if a session gets missed, or if something hurts. Clients freeze on these and then skip.
Then check in mid-week rather than waiting for Sunday. One message after the first session — did it fit the time, did anything hurt, how did the effort feel — catches the wrong assumption while it is still one session old.
The mistake that costs the most
The beginner's mistake is writing the program you would want to do. It shows up as too many exercises, too many training days, exercise variety chosen because writing the same session twice feels lazy, and a session length that assumes nobody has to be at work at seven.
The tell is easy to check: if the program you just wrote would be fun for you to run, and your client is a 34-year-old with two kids who has not trained seriously in six years, you have written it for the wrong person. The program that works is usually less interesting than the one you wanted to write, repeats itself more than feels comfortable, and is short enough to fit in a real morning.
Write for adherence first. You can always add complexity at block two, once the client has proved they will turn up — and by then you will have four weeks of actual data instead of a guess.
