Measure the friction.Act on it. Track the result.

One difficult week is noise. Repetition is a signal.

Built around your institution.

Your problem. Your teams. Your words.

What would you want to know?

• Decisions

  How often are decisions made unsure, or redone?

• Clarity

  How often is the owner, or the rule, unclear?

• Workload

  Do high-load weeks keep coinciding with rework?

• Pressure

  How often do people escalate because time ran out?

• Working together

  Where is work sitting still, waiting on another team?

• Sales

  What holds deals up?

• Production

  What stops the line for the same reason twice?

• Yours

  The question only your institution would ask.

Examples only. Every engagement starts from what you want to understand.

Tailored before the first check-in. Consistent after it, so each reading can be compared with the last.

The method

1. 1Count

   Each week, count the share of check-ins that report the condition.

   Formula

   $$p = \frac{k}{n}$$

   Worked example23 of 60 check-ins = 38.3%.

   WhyEach check-in describes the most significant decision of the week. The method shows where friction concentrates and recurs. It does not estimate how often it occurs across all work.

2. 2Baseline

   Compare the team with its own recent history, using measures that one unusual week cannot distort.

   Formula

   $$\begin{gathered}m = \text{median of the team's history} \\ s = \text{robust spread of the team's history}\end{gathered}$$

   Worked examplemedian 22.0%, robust spread 5.9 points.

   WhyA mean and a standard deviation move when one crisis week enters the history. The median and the robust spread do not.

   • The robust spread is built on the median, not the mean.
   • The reference window holds up to 12 completed weeks, with a minimum of four. Readings are provisional until it matures.
   • The week being tested is never part of its own reference.
   • The spread never falls below what chance alone would produce for a share measured on that many responses.
   Add a crisis week (55%) to the history

   With a mean, one crisis week would hide the next problem. With a median, it does not.

3. 3Distance

   Measure how far this week sits from the team's own normal, in units of its own spread.

   Formula

   $$z = \frac{p - m}{s}$$

   Worked examplez = (38.3 − 22.0) / 5.9 = 2.75

   WhyThe same number of points means different things for a steady team and a variable one. Dividing by the team's own spread makes readings comparable over time for that team.

4. 4Strength

   Convert the distance into a score from 0 to 100, then reduce it when participation is thin.

   Formula

   $$\text{strength} = 100 \times \sigma(z - 1.5) \times a,\quad \sigma(x) = \frac{1}{1 + e^{-x}}$$

   Worked example100 × σ(1.25) × 1.00 = 78

   WhyThe curve is bounded, so one extreme week cannot dominate. It puts “elevated” at 50, which makes scores easy to read. The adequacy factor a lowers a score smoothly when few people answer: a thin week can mute a signal, never create one.

   Distance z2.75Sample adequacy a1.00

5. 5Persistence

   Wait for the pattern before recommending anything.

   Rule

   Elevated 2 of the last 4 weeks = WATCH. Elevated 3 of the last 4 weeks = RECOMMENDED. Otherwise IN RANGE.

   Worked exampleelevated in 3 of the last 4 weeks = RECOMMENDED.

   WhyOne difficult week is noise. A recommendation requires repetition.

   W1W2W3W4

   RECOMMENDEDelevated 3 of 4 weeks

   Press a week to toggle it. The status follows the rule.

• How change is checked

  When a reading moves from one period to the next, the change is tested before it is reported. Shares are compared with a two-proportion test. Ordered scales are compared with the Mann–Whitney U test, with the rank-biserial correlation as the size of the effect. Because the same people answer repeatedly, the checks resample by respondent.

• Fixed in advance

  The thresholds, 50 for elevated and 75 for strong, are set before reporting begins. Any recalibration applies going forward only, is versioned and documented. Past scores are never restated.

• What the method does not claim

  It describes; it does not establish cause. It shows where friction concentrates, not how often it occurs overall. It never compares one team with another. Groups below 10 people are not shown.

Everyone in the room has a reason to say yes.

• The executive

  • Built around your priorities.
  • One page to forward.
  • A report that reads in 90 seconds.

• HR

  • Working conditions, not people.
  • No individual scores.
  • Voluntary throughout.

• Risk and compliance

  • Conditions, not verdicts.
  • No finding about misconduct or control adequacy.
  • Thresholds fixed in advance.

• The teams

  • About a minute at a time.
  • No app and no login.
  • They hear back: you said, here is what changed.