How to Measure Progress When Learning a New Skill
TL;DR: Measure progress in a new skill by setting a concrete end goal, then tracking performance, not just exposure. Use the same task, score accuracy, speed, errors, and independence, and retest after a delay and after sleep to see whether the skill actually stuck.
Start with a concrete end goal
Measure progress by deciding what “done” looks like before you start. The site’s house rule is blunt in a good way: “A small, concrete goal on day one, so it is obvious at the end whether it worked.” That is the right vibe.
So instead of “get better at spreadsheets,” pick something observable, like “clean up a messy 20-row spreadsheet so the dates, names, and totals are consistent and there are no duplicate entries.” Now you can tell whether the skill improved, because the result can be checked, not just admired from a distance like a suspiciously neat notebook.
Measure what you can actually do, not what you have read
Progress shows up in performance. That usually means four things:
- Accuracy, how many errors you make
- Speed, how long the task takes
- Consistency, whether you can do it the same way more than once
- Independence, whether you can do it without prompts, notes, or constant peeking over the fence
Reading about a skill can help, but it is not the same as doing the skill. The APS review of learning techniques notes that practice tests and distributed practice are high-utility methods, while rereading is much less useful for actual learning (APS review: https://www.psychologicalscience.org/publications/journals/pspi/learning-techniques.html). In plain English, familiarity is not the same as ability.
Use three kinds of checkpoints
A decent measurement setup uses more than one checkpoint, because practice can fool you.
- Immediate performance after practice. Can you do the task right after a session, with fewer errors than before?
- Delayed recall or re-performance. Can you do it later, after some time has passed, without re-reading your notes?
- Transfer to a slightly different task or context. Can you handle a similar but not identical version, which is where real understanding usually stops pretending and shows its cards?
This lines up with the IES practice guide, which recommends quizzing, spacing learning over time, and reviewing after a delay (IES guide: https://ies.ed.gov/ncee/wwc/PracticeGuide/1). If you only test immediately, you are mostly measuring short-term performance. Useful, yes. Complete? Not really.
Build a simple scorecard for the skill
Keep the scorecard boring. Boring is good. Boring is measurable.
A weekly log can include the same few measures every time:
- Time to complete the task
- Error count
- Prompt needed, none, light hint, or full rescue mission
- Whether the result met the target standard
Use the same task or a very close version each week. That makes comparison possible. If the task changes every time, the numbers become decorative, which is a shame because decorative numbers feel productive and are terrible at telling the truth.
Worked example: measuring progress in a practical skill
Say someone is learning basic spreadsheet cleanup. The goal is to take a messy list and make it usable, with consistent dates, tidy names, and duplicates removed.
Day-one baseline
They try to clean a 20-row sheet.
- Time: 18 minutes
- Errors: 6, including one duplicate left in and two dates formatted inconsistently
- Prompt needed: frequent
- Target standard met: no
Midweek practice
Same kind of sheet, similar mess.
- Time: 12 minutes
- Errors: 3
- Prompt needed: light hint twice
- Target standard met: partly
End-of-week performance
Same task again, but this time they wait until the next day and do it without notes.
- Time: 9 minutes
- Errors: 1
- Prompt needed: none
- Target standard met: yes
That is real progress. Not because they saw the steps three times, but because they did the task better, faster, and more independently. Repetition alone would not prove that. Improvement does.
Use self-testing and spaced review to reveal real learning
Self-testing is among the most reliable ways to see whether knowledge has stuck.
For a practical skill, self-testing does not have to mean formal quizzes. It can mean:
- closing the notes and doing the task from memory
- rebuilding the process after a delay
- explaining the steps aloud, then checking whether the explanation matches the actual work
If you can only do the skill while staring at the instructions like they owe you money, the learning is still shaky.
Watch for false signs of progress
A few things feel like progress but can overstate it:
- Feeling fluent during practice
- Recognizing the right steps when you see them
- Doing well only when notes are open
- Getting faster on the same problem just because it is familiar
Those are useful, but they are not the final test. Judge progress by unaided performance after a delay. That is where the honest score lives.
The PLOS ONE replication of Ebbinghaus’ forgetting curve is a useful reminder that memory changes over time, not just during the practice session itself (PLOS ONE: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0120644). If you wait a day and everything evaporates, the skill was not fully learned yet.
Let sleep be part of the measurement
Sleep matters because it affects learning and retention. The NIH/NHLBI notes that sleep deficiency can interfere with learning, focus, and reaction, and that getting enough quality sleep is vital for health and well-being (NHLBI: https://www.nhlbi.nih.gov/health/sleep-deprivation).
That means a better test of progress is not just how you perform at the end of practice, but how the skill holds up after a normal sleep cycle. If you can do it tonight but not tomorrow, the scorecard should say so. Sleep is not a bonus feature. It is part of the system.
Adjust your plan when the numbers stall
Plateaus, mistakes, and forgetting are not proof that you are bad at the skill. They can be signals that the practice setup needs work.
If the numbers stall, try one of these:
- Improve practice quality, not just quantity
- Increase spacing between sessions
- Simplify the task and rebuild from a narrower version
- Get feedback on the specific step that keeps breaking
- Add more self-testing instead of more rereading
This also fits the deliberate practice review in Frontiers, which treats training as necessary for improving performance while also showing that the evidence is more complicated than the simple “just practice a lot” story (Frontiers review: https://www.frontiersin.org/articles/10.3389/fpsyg.2020.01134/full). If progress slows, that usually means the next practice session should be more targeted, not that the skill has personally insulted you.
Keep expectations realistic for a two-week window
In two weeks, the goal is usually not mastery. It is reliable improvement in a narrow, concrete version of the skill.
So set expectations like this:
- You should expect measurable improvement in a defined task
- You should not expect full, flexible mastery of everything related to the skill
- You should expect some forgetting, then less forgetting if you space practice and self-test
That is a perfectly respectable outcome. Small wins, clearly measured, beat vague confidence every time.
Related guides that help you measure and plan
If you are still picking the skill itself, how to choose which skill to learn next helps you choose something compact enough to measure quickly, which matters because fuzzy goals make fuzzy progress.
If you want a simple way to structure the learning mix, what is the 70-20-10 rule for learning is useful for thinking about practice, feedback, and instruction without turning the whole thing into a spreadsheet about spreadsheets.
Bottom line
Measure progress by performance on a concrete task, over time, with the notes closed when it counts. Track accuracy, speed, consistency, and independence. Test again later, test after sleep, and compare the same task or a close variant. If the numbers improve, that can be a sign of learning. If they stall, change the practice, not your self-worth.