If You Can’t Explain the Score, Don’t Expect People to Trust It

I opened my Cosmo teaching dashboard recently and saw that I’d received a Silver badge for August.

My first thought wasn’t, “Why didn’t I get Gold?”

It was:

What am I supposed to do with this information?

That question feels familiar.

• A performance review says you’re “meeting expectations.”
• A sales dashboard says you’re below target.
• A customer metric turns red.
• An employee survey produces a score.
• A project health indicator lands at 82%.

Now an algorithm tells you that you’re Silver.

Knowledge workers spend a lot of time being measured. The problem isn’t measurement. Useful measurement can help us see patterns, make better decisions, and improve.

The problem starts when the score makes more sense to the people who created it than to the people expected to act on it.

A Number Is Not Automatically Feedback

Cosmo says my August badge reflects my “overall teaching contribution last month.”

That sounds reasonable.

The dashboard also shows attendance, quality, trials, and renewals. But the current figures are labeled as this month’s activity, so they don’t explain the August result.

I know the outcome: Silver.

I don’t know enough about why I received it to decide whether I should change anything.

Maybe Silver accurately represents my contribution. Maybe I’m missing something important. Maybe Gold depends on something I can influence. Maybe it doesn’t.

I simply don’t know.

That makes the badge an evaluation.

It doesn’t necessarily make it useful feedback.

Useful feedback should help someone make a better decision.

We’ve All Been On Both Sides Of This

This isn’t really about Cosmo.

If you’ve worked in technology, consulting, professional services, education, operations, sales, or management, you’ve probably seen the same pattern.

A real problem appears:

• We need more consistent performance reviews.
• We need better accountability.
• We need to reduce favoritism.
• We need to know whether customers are happy.
• We need to recognize stronger performance.

So someone builds a system.

I’ve been one of those people.

At Axelerant, we automated parts of our appraisal and salary revision processes because the manual approach was cumbersome and inconsistent. We wanted clearer criteria, more consistency, less administrative work, and better conversations around the exceptions.

Those were good goals.

But sitting on the receiving end of a Silver badge reminded me of something easy to forget:

A system almost always makes more sense to the people who built it.

• They sat through the meetings.
• They know why each metric exists.
• They know the assumptions, exceptions, and tradeoffs.
• They’ve seen the spreadsheet, workflow, or code.

Everyone else gets the dashboard.

Precision Is Not The Same As Understanding

Knowledge work has no shortage of measurement.

• Velocity.
• Utilization.
• Conversions.
• NPS.
• Performance ratings.
• Completion rates.
• Customer health scores.
• AI-generated summaries and rankings.

We are very good at turning complicated work into numbers.

Sometimes that helps.

Sometimes we take a complicated reality, produce a precise-looking score, and mistake precision for understanding.

Automation can make a process faster and more consistent. It can reduce repetitive work and some forms of bias.

But automation doesn’t automatically improve the underlying judgment.

And it certainly doesn’t make it easier to understand.

Sometimes it just makes an unclear decision happen faster.

The Useful Question Is: What Can I Do Differently?

I don’t need to know whether some hidden variable contributes 6.7 percent to my result.

I need enough context to understand:

• What mattered?
• What evidence affected the result?
• What could I reasonably influence?
• What would stronger performance look like?
• What should I try differently next time?

If student participation is too low, tell me.

If renewals matter significantly, I should know that.

If teaching quality depends partly on a behavior I’m overlooking, show me.

I don’t need the source code.

I need a clear enough connection between what I did, what happened, and what I might do next.

That’s when measurement becomes feedback.

More Transparency Is Not Always Better

There is an obvious counterargument.

If organizations publish every weighting and threshold, people will optimize for them.

We’ve all seen that too.

If closing tickets is rewarded, people close tickets.

If utilization drives evaluation, people maximize billable hours.

If renewals determine teacher quality, teachers may start optimizing for renewals instead of learning.

The metric becomes the work.

So I don’t think the answer is radical transparency.

The goal is simpler:

Enough transparency for informed judgment.

Tell people what matters.

Explain how the major pieces fit together.

Show the evidence that affected the outcome.

Give examples of stronger performance.

Provide a reasonable way to ask questions.

Then trust people to use their judgment.

This Is Really An Ownership Problem

I’ve spent a lot of time thinking about Sustainable Ownership.

One lesson keeps coming back: leaders can’t simply demand ownership.

People need the conditions to exercise it.

That includes clarity, context, knowledge, autonomy, accountability, trust, and reasonable capacity.

Take too much of that away, and people stop exercising judgment.

They start trying to figure out what the system wants.

• “What do I need to put in the form?”
• “What number does management want?”
• “What boxes do I need to check?”
• “What do I need to do to get the good rating?”

That’s compliance.

Ownership sounds different:

• “What outcome are we trying to create?”
• “What does the evidence tell me?”
• “What could I improve?”
• “What should I try next?”

A useful performance system should encourage those questions.

The Harder Question Is About The Systems I’ve Built

It’s easy for me to look at a Silver badge and say, “This needs more transparency.”

The more useful question is whether I’ve created the same problem for other people.

When we automated performance processes, did people understand them as well as we thought?

When I built dashboards, did they help people make better decisions, or mostly help management monitor them?

When I asked someone to take ownership, had I actually given them enough context to do it?

I’ve certainly gotten some of this wrong.

That’s probably why this badge has stuck with me.

Opaque systems are easy to notice when you’re being scored.

They’re much harder to recognize when you’re holding the spreadsheet.

A Better Test

I’ve landed on a simple test for performance systems:

After someone receives the score, can they make a better decision because they saw it?

• If yes, you’ve probably created feedback.
• If no, you’ve probably created a scoreboard.

And if people have to reverse-engineer the scoreboard to figure out what the organization wants from them, don’t be surprised when they start optimizing for the score instead of the work.

I don’t need Cosmo to explain every detail behind my Silver badge.

But if the badge is meant to help me become a better teacher, I need enough information to understand what it’s trying to tell me.

The same standard applies to the systems I’ve built.

And probably to more of the dashboards we look at every week than we’d like to admit.

If you can’t explain the score well enough for someone to learn from it, don’t expect them to trust it enough to own it.