You prescribe three sets of eight at 60kg. The client logs three sets of eight at 60kg. You add 2.5kg. What you do not know is whether the last two reps were clean, whether the third set was actually a set of six with a long pause in the middle, or whether the bar ever reached their chest.
In a gym you would have seen all of that. Online you are progressing on a number the client typed, and the number is only as good as their honesty, their attention and how clearly you wrote the prescription. Everything below is about making progression work anyway: which variable to move, how fast, what to do when the numbers stop moving, and how to tell that apart from a client who had a bad week.
If you need the variables themselves and what each one costs in recovery, that is what progressive overload is. This is the applied version.
Write the progression rule into the program
The most common failure is not choosing the wrong scheme. It is having no scheme written down, so progression happens whenever you remember to look at that client.
With five clients you can carry it in your head. With thirty-five you cannot, and what happens instead is that the loud clients get progressed every fortnight and the quiet ones sit on the same weight until they leave.
The fix is to put the rule in the prescription, where the client reads it:
Dumbbell bench press — 3 sets of 8 to 10 at 27.5kg. When you get 10 reps on all three sets, move to 30kg next session and start again at 8.
That is a rule the client can run for six weeks without asking you anything. Your job shifts from deciding every increment to checking that the rule is being followed and stepping in when it breaks. It is also the difference between a program that survives your week off and one that stalls the moment you get busy — the rest of that build process is in how to write an online coaching program.
Double progression should be your default
Double progression means the client progresses reps first and load second. You fix a rep range, they add reps inside it at a fixed weight, and when they hit the top of the range on every set they add the smallest available load increase and drop back to the bottom of the range.
Three reasons it suits online coaching specifically:
- It removes the judgement call from the client. "Go heavier when it feels easy" asks an untrained person to make a decision they are not equipped to make. "Ten on all three sets, then add 2.5kg" does not.
- It self-corrects for a bad week. A client who slept badly repeats the same weight and moves on. Nothing needs rescuing, and nothing needs a message from you.
- It gives you a binary signal in the log. Either the reps went up or they did not. You are not interpreting how a set looked, which you cannot see.
The alternative schemes both have an online cost. Straight load progression — add 2.5kg every week regardless — works for a genuine beginner and then produces a client grinding a weight they cannot handle in the rep range you asked for. Effort-based progression, where the client trains to a target RPE and picks the weight themselves, is more accurate in principle but only as good as the client's rating, which takes months to become reliable — see how to use RPE with online clients for what that process actually involves.
Start every client on double progression. Move them to effort-based work once their ratings agree with their numbers.
Choose one variable per exercise, and choose it for what the log shows
Some variables leave evidence in a log you can read. Others leave nothing, and progressing on them online is faith rather than coaching.
| Variable | What the log tells you | When to use it |
|---|---|---|
| Reps | Exact, set by set. The highest-trust number you get | Default, inside a fixed rep range |
| Load | Exact, once the top of the rep range is reached on every set | The second half of double progression |
| Sets | Exact, and it moves weekly volume with it | Recovery has room and the client has the time |
| Range of motion | Nothing at all | Only when you have a recent video of the lift |
| Tempo | Nothing, unless the set is timed | Late, on isolation work, with a cue the client can hear |
| Rest periods | Session duration, which is too noisy to read | Conditioning blocks, not strength work |
The practical rule is one variable per exercise per block. If load, reps and sets all move in the same four weeks and the client stalls, you have no idea which change caused it and no clean way back.
Frequency is the exception worth planning rather than progressing. Adding a session is not a small change — it moves weekly volume for every muscle that session touches, and it usually means a different structure rather than a fourth day bolted onto a three-day week. That is a split decision, covered in how to choose a workout split for a client.
How fast to add load
The honest answer is that the log decides and these are planning defaults. As typical ranges rather than researched figures, this is roughly where most clients land:
| Training age | Lower-body compounds | Upper-body compounds | Isolation work |
|---|---|---|---|
| Under 6 months | 2.5–5kg per session | 2.5kg per session | Next dumbbell up when reps top out |
| 6–24 months | 2.5kg every 1–2 weeks | 1.25–2.5kg every 2–3 weeks | Next dumbbell up every 3–4 weeks |
| Over 2 years | 2.5kg every 4–8 weeks | 1.25kg every 6–12 weeks | A few increases a year |
Two things fall out of that table.
The first is that a beginner's progression is fast, boring and almost automatic, and the main way to ruin it is to add weight faster than their technique can absorb. A client who added 20kg to their squat and lost 15cm of depth got weaker, and you will not see it in the log.
The second is that intermediate progression is measured in reps, not kilos. A client who goes from 8, 8, 7 to 10, 10, 9 at the same weight across a block has added about 15% more work on that exercise without touching the load. If you only count kilos, that client looks stalled for two months and you will start changing things that were working.
Review per exercise, not per session
Session-level review hides the problem you are looking for. A client can complete every workout for four months, rate them all as challenging, and be lifting the same weights they lifted in March. Nothing in a completion percentage tells you that.
What you need is the last few entries for a given exercise, side by side, and you need to be able to get them in seconds or you will not do it for thirty clients. On Fitsly, each client's Training tab has an Exercises view listing every exercise they have logged with their history and last-logged numbers per exercise, and the client sees their previous result for that exercise while they are logging the next one — so the person doing the reps knows what they have to beat. Fitsly's training tools hold that history against the exercise rather than against the session, which is the difference between answering "is she progressing on the row?" in ten seconds and reading back through eight workouts.
Whatever you use, the review question is per exercise: what were the last three entries, and did anything move?
A worked example: twelve weeks on one lift
Client is fourteen months into training, three sessions a week, dumbbell bench press at 3 sets of 8 to 10, starting at 27.5kg. The rule is ten reps on all three sets, then up to 30kg.
| Week | Load | Logged reps | What you do |
|---|---|---|---|
| 1 | 27.5kg | 10, 9, 8 | Nothing. The rule has not been met |
| 2 | 27.5kg | 10, 10, 9 | Nothing |
| 3 | 27.5kg | 10, 10, 10 | Rule met. Move to 30kg |
| 4 | 30kg | 8, 8, 7 | Nothing. This is the expected drop |
| 5 | 30kg | 9, 8, 8 | Nothing |
| 6 | 30kg | 9, 9, 8 | Nothing |
| 7 | 30kg | 8, 8, 7 | Flag it. One backwards week is noise |
| 8 | 30kg | 9, 9, 8 | Back on track. Do nothing |
| 9 | 30kg | 10, 9, 9 | Nothing |
| 10 | 30kg | 10, 10, 10 | Rule met. Move to 32.5kg |
| 11 | 32.5kg | 8, 7, 7 | Nothing |
| 12 | 32.5kg | 8, 8, 7 | Nothing |
Twelve weeks, one exercise, two load increases, and you made a decision on three of them. That is what progression looks like for an intermediate client, and it is far less dramatic than most coaches expect. The client added 5kg and roughly a rep per set. Told that in advance, they would have been disappointed. Shown the log at week twelve, they can see it.
Note week 7. If you were reviewing session by session you might have moved the exercise, changed the rep range or written a paragraph about intensity. It was one week, and the following week resolved it.
Telling a real stall from a bad week
A stall is three consecutive sessions with no movement in load or reps on the same lift, in a phase where progress was expected, with attendance intact. Anything less than that is noise, and sleep, work, illness and a bad night out all produce it.
"In a phase where progress was expected" is doing real work in that sentence. A client six weeks into a fat-loss phase is not meant to be adding weight — holding their loads is the win, and treating it as a stall leads you to add volume to someone who cannot recover from it. That is why progression is easier to read against a planned structure than against a rolling program; periodization for online coaches covers setting those expectations before the block starts.
When something looks like a stall, work through it in this order before you change any programming:
- Did they train? Count completed sessions over the last three weeks, not the last one.
- Did they do all the sets? Trimming the last set of everything is the most common silent change a client makes, and it will not be in a message.
- Is it one lift or everything? One lift is usually a technique or exercise-selection problem. Everything at once is recovery, sleep or calories, and no programming change fixes it.
- Where are they in the phase? Weeks four to six of a hard block are supposed to feel heavy.
Most of what looks like a programming problem is answered in those four steps.
What to change when the stall is real
Go in order of cost. Do one thing, then wait two sessions before doing another, or you will not know which one worked.
- Repeat the week. Do nothing at all and let them run the same weights again. This clears a surprising share of stalls, because the client was fatigued rather than at their ceiling.
- Cut back for a week. Drop to two-thirds of the working sets at the same loads. Fatigue clears, and the following week usually produces the reps you were waiting for.
- Reset and rebuild. Take 10% off the load, run the same rep range up, and arrive back at the stalled weight in three or four weeks with more reps in reserve.
- Change the variable. If load has stopped, hold it and chase reps. If both have stopped, change the exercise to a close variation and progress that.
Changing the exercise feels like the obvious first move and should be near-last. A new exercise resets the numbers, so the client looks like they are progressing for three weeks whether or not anything actually improved.
Usually, nothing is progressing because nothing is happening
Before any of the above, check the attendance. A client training two of four prescribed sessions is not stalled. They are training half a program, and no progression scheme rescues that.
This is the most common cause of a flat log and the one coaches spend the least time on, because programming is the interesting problem and adherence is not. But a client who trains three of four sessions for a month has done three-quarters of the work, and the numbers will read exactly like someone who has plateaued.
So put attendance ahead of programming in your review. If a client has missed sessions, the conversation is about what is getting in the way — the session is too long, the gym is too far, the 6am start was aspirational — not about set and rep schemes. Fixing the schedule usually restores the progression on its own, and keeping online clients accountable is the harder half of that job.
Progressing a client who trains at home
Fixed dumbbells break load progression. Going from 12.5kg to 17.5kg is a 40% jump, which no client absorbs inside a rep range of 8 to 10. If the only home kit is a pair of dumbbells and a mat, load is the one variable you mostly cannot use.
What you do instead, roughly in order:
- Widen the rep range. Prescribe 8 to 15 rather than 8 to 10, so there is enough room for the next dumbbell to be survivable when they get there.
- Add sets. Three to four working sets is a real increase in stimulus and costs nothing but time.
- Add tempo or pauses. A three-second lowering phase, or a two-second pause at the bottom, makes the same dumbbell harder. Write the count in the prescription so it is not left to feel.
- Change to a harder variation. Split squat to Bulgarian split squat, push-up to feet-elevated push-up, two-arm row to single-arm. This is where most home progression comes from once the range and the sets have been used.
- Add external load crudely. A loaded backpack or a resistance band adds resistance in small enough steps to keep double progression alive.
Two rules make this work. Progress one of them at a time, the same as anywhere else. And prescribe the variation yourself — a client who is told to "make it harder" will change the exercise, the tempo and the rest all in one session, and the log after that is unreadable.
