Ask why someone uses a tracking app and the answer is usually "to see what I lifted last time." Fair — but that's half of what a log is for.
For that purpose one week of history would do. The other eleven weeks sit there doing nothing. They can answer a different question entirely.
Deciding that by feel is usually wrong. After a good session you want to keep everything; after a bad one you want to scrap it all. Three months of data frequently says neither.
1. Per-Lift Trends, Not Total Volume
The first number people reach for is weekly total volume (weight × reps × sets), and it's nearly useless for making decisions.
A total lets unrelated causes cancel each other out. Legs climb, upper body slides, and the sum holds steady — actively signaling "all is well" when it isn't.
What to read instead: individual trends for four to six main lifts.
What matters is the slope, not any single entry. One off week is just a bad day. Four or more weeks at the same number is the signal.
Don't Leave "Up, Flat, Down" to Your Eyeballs
Eyeballing a slope produces a different verdict every time. Fix the thresholds in advance and every quarter gets measured the same way. The input is percent change over 12 weeks.
One caveat is mandatory here. These bands shift wholesale with training age. Inside your first six months, +15-30% over 12 weeks is common — and +5% in that window is itself a slowdown signal. At three years in, +3% over 12 weeks is an excellent result.
Not knowing your own band means reading normal progress as a plateau and tearing up a program that was working. It's the single most common mistake in a review.
2. Not Which Days You Missed — Which Lifts
Most people already know their attendance. What they don't know is which exercises go missing.
Read the log column-wise and a bias shows up almost every time.
- Three sessions a week, but legs happened nine times in twelve weeks — the same body part gets dropped on busy days
- Whatever sits last in the session consistently gets fewer sets than planned
- Within one body part, only the lifts that are going well survive; the ones that aren't quietly vanish
The third is the common one. Lifts that go well are enjoyable and the numbers climb, so they stay — and the weak movements disappear from the routine without a decision ever being made. The weak areas then stay weak.
You remember what you did. What you skipped leaves no trace in memory — only in the log.
Catch the Bias With a Completion Rate, Not a Hunch
"I think legs slip a bit" is almost always more generous than reality. Replace it with one line of arithmetic.
Two lines are enough to judge against.
- 85% and above — normal. That's illness and schedule slippage.
- 70-85% — bias. Something is dropping it repeatedly, and the cause usually narrows to one thing (day of week, position in the session, how long it takes).
- Below 70% — that body part isn't in your program. It's in your plan; it isn't in your training.
In the example, legs at 64% isn't "I skip it sometimes" — it's thirteen sessions in twelve weeks that never happened. Changing the program because legs aren't growing accomplishes nothing here. What needs changing is the condition under which it gets dropped — move legs earlier in the week, put it first in the session, or cut the exercise count so it takes less time. Reworking the split itself is worth considering at exactly this point.
3. When the Weight Went Up but Nothing Improved
One reliable illusion: watching load while ignoring reps.
The number on the bar went up while the actual work went down. That isn't automatically bad — a deliberate shift toward lower reps looks exactly like this. The problem is arriving here without deciding to.
The usual cause is impatience. Load is the most visible number, so it gets pushed weekly and the reps never catch up. This is why progressive overload isn't a story about weight alone.
When reviewing, always read load, reps, and sets together. Look at one and you have no idea where the other two went.
Collapsing Weight and Reps Into One Number — e1RM
Which is stronger, 70kg × 10 or 75kg × 6? You can't see it. This is what estimated 1RM (e1RM) is for: back-calculating your one-rep max from a set you actually did, without ever testing it.
Same conclusion as the volume math above, reached much faster. The bar got heavier while the strength estimate went down. Pull one of these per lift out of three months of logs and the trend reads as a single line even when weight and reps move in opposite directions.
Two limits worth knowing before you rely on it.
- Accuracy degrades as reps climb. It holds up well from 1-8 reps and overestimates badly past 15. Sets above 12 reps are better left out of e1RM comparisons.
- Only compare a lift against itself. Bench e1RM and incline e1RM are separate axes. This is a tool for tracking one lift's change over time, not for absolute numbers.
If you logged whether the set was taken to failure, accuracy improves another notch. Eight reps at RIR 2 and eight reps to failure aren't the same eight reps, and only the latter belongs in the calculation. The working-weight math used to pick starting loads rests on the same formula.
4. Real Plateaus and Fake Ones
Four weeks at the same number starts to look like a plateau, and a good share of those aren't plateaus at all. The log tells them apart.
The distinction matters because the fixes are unrelated. A recovery problem doesn't care what program you switch to — start with sleep and rest days. An ordering problem is solved by moving one exercise.
Logging sleep or a note about how you felt makes this call much faster. It doesn't need to be daily — one line on the days that went unusually badly is enough.
5. The Minimum a Log Needs to Be Readable
To be readable three months later, a log needs at least this much.
The most commonly missing field is whether the set was near failure. 70kg × 8 alone can't tell you whether that was comfortable or barely there — and that difference decides whether you add weight next time.
Conversely, naming an exercise differently each time destroys the trend. Mix "incline bench" with "incline barbell press" and one lift splits into two lines. Pick a name once and keep it fixed.
6. A Thirty-Minute Review
Once a quarter, in this order:
Steps 1–4 are observation; step 5 is judgment. The order matters — decide what to change first and you'll only shop the log for supporting evidence.
Capping step 5 at one or two changes is also deliberate. Change several things at once and the next review can't tell you which one worked. Part of the point of reading a log is making the next read more informative.
With a Number Attached to Each Step
A procedure alone still leaves you judging differently every time. Pin the thresholds from earlier onto each step and it finishes inside thirty minutes and stays comparable next quarter.
The trick in step 2 is reading only the first and last week. Plotting all twelve won't fit in thirty minutes, and the wobble in between doesn't enter the decision anyway. You go back for the middle weeks only on lifts that came out "plateau" or "regression" in step 3.
Step 6 must be written as a sentence. "Do more legs" can't be verified next quarter. "Move legs to first thing Monday and cut squats from four sets to three" can. And the first item of the next review is checking whether that sentence actually happened.
7. What the Numbers Can't Tell You
Some things should never be judged from a log.
- Pain. A log records outcomes, not causes. Recurring pain is not a numbers problem.
- Form. If load climbs while technique degrades, the log reads as improvement. That needs video or another set of eyes.
- Body weight and composition. A separate axis. Worth tracking alongside, but neither substitutes for the other.
A log is strong on trends and omissions and weak on quality and causes. Know that boundary and it's genuinely useful; ignore it and good-looking numbers will talk you past a real problem.
8. Summary
- Read per-lift trends, not total volume. Totals let opposite causes cancel.
- Find the lifts that went missing. Weak movements disappear quietly.
- Read load, reps, and sets together. More weight with less volume may not be progress.
- Check whether conditions held before calling it a plateau. Most "plateaus" are changed conditions.
- Note failure and anything unusual. That's what makes the review possible later.
- Change one or two things at a time so the next review is still readable.
- When weight and reps diverge, collapse them into e1RM. Compare a lift only against its own history.
- The thresholds depend on training age. At three years in, +3% over 12 weeks is a good result, not a plateau.
Keeping a log is a hard habit, and three months of it already puts you ahead of most people. But a log you only write and never read is a diary. Thirty minutes once a quarter turns it into the basis for what you do in the next three months.
Frequently Asked Questions
I don't have three months yet. Can I review now?
You can, but the shorter the window, the wider your thresholds have to be. Over four weeks, day-to-day variation is larger than the trend, so nearly everything reads as "plateau." Wait for at least six to eight weeks. Before that, the reason to open your log isn't judgment — it's checking that the log is in a readable shape at all: are lift names consistent, is the failure/RIR field actually being filled in?
I switched lifts and the 12-week trend is broken. How do I compare?
You don't. Forcing a broken line back together manufactures progress or regression that never happened. Handle it this way instead: drop the interrupted lift from this quarter's verdict and judge on the remaining lifts. That's why the guidance is 4-6 main lifts — the extras are slack so losing one doesn't blind the review. From next quarter, that lift restarts as a fresh axis.
My app only shows a total volume graph.
Total volume graphs are useful for catching large events like an injury or a long absence. They won't surface the judgments in this article — which lift stalled, which body part is slipping. If the app can filter by exercise, viewing your 4-6 main lifts one at a time substitutes fine. If it can't, copying the first-week and last-week numbers by hand once a quarter is enough. Six lifts is twelve numbers.
Do I have to log every set, or is one representative set enough?
Log every set. The rate at which reps fall off across sets is the fastest available signal of recovery status, and 8-7-6 turning into 8-4-3 is invisible if you only record one set. If that's too much friction, a workable compromise is every set on main lifts, a representative set on accessories. Accessories don't enter the quarterly verdict anyway.
My body weight went up along with my lifts. Is that progress?
Absolute strength went up; relative strength is a separate question. Compute weight lifted ÷ body weight and it separates. Squatting 100kg at 70kg body weight (1.43x) becoming 108kg at 78kg (1.38x) is +8kg in absolute terms and a decline in relative terms. Which one matters depends on your goal. That said, body weight is a different axis from your training log — treat this as context, not as something that overrides the per-lift trends.
What if the review says nothing needs changing?
Then nothing needs changing. And that's the most frequent correct answer and the most frequently ignored one. If every main lift is climbing inside its training-age band and completion rates are above 85%, there is no reason to touch anything. Rebuilding the program here just resets a stimulus that was accumulating properly. Half the value of a review is producing the evidence that you don't have to change anything. If you're early on, reading it alongside first-month expectations makes the call easier.