Continuous feedback, done right
Vaidy Bala
January 25, 2026
Last month, I wrote about feedback mirrors - the idea that we need constant reflection to know whether we're growing or slipping.
Today, I want to show what that looks like in practice.
The problem with traditional feedback
Remember report cards? A single number that supposedly captured months of learning.
But if Maria scored 72%, what should a teacher actually do next?
Is she improving? Was she at 60% last month and climbing - or 85% and slipping? Which topics need work?
You can't tell from a single number.
Traditional feedback is like looking at a single frame of a movie and trying to understand the plot. You're missing the story.
What a better mirror looks like
This is why we built the Student Performance dashboard in Spark.
It shows the complete learning journey - not as a snapshot, but as a timeline.
You see all your students at once.
All students at a glance
From there, you can drill down layer by layer:
Student-level performance trends over time
Test-level results within any time window
Question-level detail within a specific test
Individual student performance graph
Tooltip showing multiple tests
Quiz report with question details
This multi-level drilldown goes from subjects → specific topics → individual questions.
Teachers and students can explore this together, identifying exactly where help is needed.
And once a weak topic is identified, the student can dive straight into it with Spark's AI tutor.
The mirror doesn't just show the problem - it helps solve it.
Why this actually works long-term
There's a reason most feedback systems don't survive in the real world.
They add work.
Teachers are asked to:
Design quizzes
Schedule tests
Grade submissions
Analyze results
Prepare reports
It works for a while.
Then reality kicks in.
What makes this different is that none of that work exists.
In Spark:
Quizzes are generated automatically from teaching content
Tests are deployed without manual setup
Submissions are graded automatically
Performance is tracked continuously
Reports update in real time
The only thing the teacher does is teach - exactly as they already do.
In this case, the only setup required was connecting a Zoom account.
That's it.
The feedback mirror keeps working quietly in the background - class after class, week after week - without adding a single new task to the teacher's plate.
That's the only way feedback mirrors can work at scale.
What's next
Right now, the dashboard shows the data.
Next, we're teaching the system how to interpret it.
We're building automatic insights and personalized recommendations on top of this foundation.
The system will:
Detect learning gaps
Identify patterns over time
Suggest exactly what each student should focus on next
This is our first real step toward building a set of virtual assistants that work individually with each student - offering guidance based on their actual performance, not generic rules.
If you're already using Spark, you can explore the Performance dashboard today.
And if you're not using Spark yet but this kind of feedback feels overdue, just reply to this email - I'll set you up with early access.