Knowing If It's Working

Making a change is not the same thing as improving a system. Some changes feel productive because they create emotional relief, novelty, or a brief sense of control. The real question is whether the change made the practice more workable.

In Adaptable Discipline, evaluation is not about asking whether the intervention felt exciting or correct. It is about asking whether it changed the underlying conditions in a useful way. That is also why evaluation belongs with running small experiments. A change is often a test of a working explanation, not a final answer.

What A Good Change Should Improve

A good change usually improves one or more of these:

• return: it becomes easier to come back after drift
• friction: the cost of beginning or re-entering goes down
• capacity fit: the practice becomes more realistic under current conditions
• direction: the practice becomes more clearly tied to what matters
• visibility: it becomes easier to tell what is happening in the system

If none of those changed, then the intervention may have been emotionally satisfying without being structurally useful. It may also mean the hypothesis behind the intervention was incomplete or pointed at the wrong constraint.

Evaluate The Right Thing

Many people evaluate a change too early or by the wrong signal. They ask whether it felt good, whether it looked disciplined, or whether they performed perfectly for a few days. Those signals can mislead.

A better evaluation asks:

• is return cheaper now?
• is the practice easier to begin?
• does the system survive low capacity better than before?
• is there less confusion about what to do next?
• is comeback speed improving?

Those questions keep the evaluation tied to the framework instead of to mood. They also help you tell whether the experiment is confirming the hypothesis, weakening it, or revealing a different problem than the one you thought you were solving.

Look For Friction Migration

Sometimes an intervention removes friction in one place but creates it somewhere else. A new tool may preserve state but add maintenance burden. A reduced version may make return easier but weaken direction if it becomes the only version that ever gets used. A metric may improve visibility but increase self-surveillance.

That does not automatically make the intervention bad. It means you need to evaluate the whole effect, not just the first benefit. In experimental terms, you are looking not only for the intended effect, but also for side effects and friction migration.

What A Result Usually Means

Evaluation gets easier when you have a small set of interpretations available.

• better result: the change made return cheaper, clearer, or steadier
• partial result: the change helped, but only in narrow conditions
• moved problem: the original bottleneck eased, but another one now limits the system
• false relief: the change felt good but did not improve return, clarity, or stability
• new burden: the change solved one problem by creating too much maintenance, pressure, or confusion somewhere else

This is often the difference between useful iteration and random churn. You are not only asking whether the change worked. You are asking what the result is telling you about the system.

Give The Change Enough Time To Show Itself

Not every intervention reveals its value immediately. Some changes help at once. Others only show their value when the next hard day arrives. If a change is meant to help with re-entry, you may not really know whether it works until the next time the system wobbles.

That is why evaluation should include both immediate feel and stress-test value. A practice that feels elegant under ideal conditions but fails under variance still needs work. A useful experiment often needs enough time and enough pressure to reveal what is actually true.

What To Record

Evaluation gets better when it is concrete. You do not need an elaborate dashboard, but it helps to note:

• what change you made
• what hypothesis the change was testing
• what it was meant to improve
• what actually became easier
• what stayed hard
• what new friction appeared

This keeps the system from becoming a blur of vague impressions.

For example, a simple note might say: "Left the next step visible after each writing session. Re-entry was easier for two days. After the first miss, shame still delayed the return. Friction improved, but mindset is still part of the bottleneck."

The Real Test

The real test of a change is simple: does it make the practice more buildable?

If it makes return more available, reduces unnecessary cost, improves alignment, or helps the system hold under real conditions, it is probably worth keeping. If it mainly adds complexity, pressure, or noise, it probably needs adjustment or removal. Either way, the evaluation should leave you with a better hypothesis for the next pass through the system.

Try it: Evaluate a change you made recently

Pick a change you made to a practice in the last week or two.

1. Name what you changed. One sentence.
2. Name what it was supposed to improve. Return cheaper? Entry clearer? Less shame? Better capacity fit? If you didn't have a target, that's information too.
3. Look at the actual signal. Is comeback speed different? Does the practice hold better under harder days? Is re-entry cheaper, or does it still cost as much as before?
4. Classify the result. Better (the target improved), partial (helped only under good conditions), moved problem (one bottleneck eased, another appeared), false relief (felt good, changed nothing structural), or new burden (solved one problem, created another).

You're done when you have a classification and a sentence about what it suggests for the next change.

Where this leads: Adjusting as You Go takes the result and turns it into the next iteration.