Back to the on-screen lesson ·
A people review runs from the role's evidence and asks whether a gap is one person or everyone — a support need or a process need — without blame, ending in one agreed next step.
Paper packet. Every task here also exists on screen, where it is checked automatically; answers written on paper are not assessed by Nydus. When you are back at a device, enter your answers there.
You will sort review findings into support needs, process needs and no action, read whether evidence points to one person or the whole team, keep impressions out of review notes, and order a people review that ends in one next step.
An operations review improves a process one change at a time. A people review does the same for how each role is working — from the same kind of evidence, and with the same refusal to blame.
| Term | What it means |
|---|---|
| People review | A regular conversation about one role, run from the role's evidence. |
| Evidence | The role's measures, the training record and dated feedback notes. |
| Support need | Something the person lacks: training, a tool, time. |
| Process need | Something missing from how the work is set up. |
| Next step | One agreed action, with a date, that ends the review. |
| Impression | A feeling about a person; not evidence. |
A review that starts 'how do you think you're doing?' or 'I've not been happy lately' runs on feelings. A review that starts with the evidence runs on facts both people can see: the driver brings the quarter's on-time delivery rate; the assistant brings the unbroken-order rate and the training record.
Then ask what the evidence points to. The most useful single question is: is this one person, or everyone?
Another way: table
What the evidence points to.
| Evidence | Points to |
|---|---|
| Everyone misses the same step | A process need |
| One person below the rest | A support need |
| On target all quarter | Say so; no action |
Another way: steps
Blame is tempting because it is quick: 'the cleaners are careless', 'the new one is letting the side down'. But blame teaches people to hide problems, and a business that cannot see its problems cannot fix them. A cleaner who expects to be blamed for a failed inspection stops reporting near-misses; a mechanic who expects a telling-off does not mention the torque wrench he never had.
So the review asks two open questions — what helped, and what got in the way? — before deciding anything, and the notes it relies on are dated facts, not impressions like 'seems less enthusiastic'. Most gaps turn out to be support or process needs that are cheap to meet. Where a real performance problem remains after support has been given, the earlier lessons apply: specific feedback, a record, and — only if it comes to that — a process that follows the local rules.
Gather the evidence. Before the review, collect the role's measures for the period, the training record and the dated feedback notes. Ask the person to bring them too, so both start from the same page.
Read against the target. Turn counts into rates the same way for everyone: passed inspections over all inspections, times a hundred. Find the gap to the target in percentage points.
Compare with the team. Put the person's rate beside the others'. If the whole team misses the same thing, look at the process. If one person stands apart, look for a support need.
Ask before deciding. What helped? What got in the way? The answer often names the need directly: a step never shown, a tool missing, a schedule too tight.
Agree one next step. Training booked, a tool bought, a checklist line added, with a date. Name specifically what went well.
Check the work. Is every note a measure or a dated record? Were the rates worked out the same way for everyone? Does the next step answer the need the evidence pointed to?
Starting from evidence is needed because memory favors the recent and the dramatic. One bad week in the last month can outweigh eleven good ones in the owner's mind; the measure for the quarter does not.
Comparing with the team is allowed because the team works the same process. If everyone misses the oven, the difference cannot lie in any one person, so it must lie in what they share: the standard work and the checklist.
Asking before deciding is needed because the person knows things the records do not: the bridge works on the delivery route, the wrench that went missing, the step nobody showed them.
Ending with one next step is needed for the same reason as in an operations review: a single change can be checked at the next review; a list of complaints cannot.
Comparing people by raw counts misleads when they did different amounts of work. A cleaner with 4 failed inspections out of 40 has done better than one with 3 out of 15: 10 percent against 20. Turn every count into a rate over the work done before comparing, and use the same period for everyone.
Small numbers need care too. One failed inspection out of 5 is 20 percent, but it is one clean. Read a short period as a question to ask, not a conclusion.
A review is also the place to say, specifically, what is working. 'Your move-outs passed first time 19 times in 20 this quarter, and the oven checklist you suggested is now used by everyone' tells the person exactly what to keep doing. It also spreads good practice: the team learns what the standard looks like when someone meets it.
Vague praise — 'good job' — is pleasant but teaches nothing. Specific praise is as useful as specific feedback, and costs less. Write it into the review note as well, so the record shows the good quarters as clearly as the hard ones.
A quarterly review of thirty minutes suits most roles in a small business: often enough that the evidence is fresh, rare enough that each review has a quarter's measures to read. New team members benefit from a short review after their first month, when support needs are most likely and cheapest to meet. Keep the reviews on the calendar even when things are going well; a review that only happens when something is wrong soon feels like a summons, and people arrive defensive instead of ready to talk.
Sometimes the person reads the same evidence and sees it differently. That is useful, not a threat. Ask what they see: perhaps the measure counts something outside their control, such as a late parts delivery counted against a mechanic, or perhaps the target was never explained. If the measure is unfair, fix the measure for everyone. If the target was unclear, explain it and agree a date to read the measure again. Write down both views in the note. A review where the person can disagree safely is one they will speak honestly in, and honest answers are what find the real need.
Most support needs are cheaper than owners fear. A second training session costs an hour or two. A missing tool might cost fifty dollars. A checklist line costs nothing. Moving a task to a quieter time of day, or pairing a new team member with an experienced one for a week, changes the result without changing the pay bill. Compare that with the cost of replacing someone, which earlier lessons showed can run to weeks of lost work and the cost of hiring. Meeting a support need is usually the cheapest improvement a small business can make.
Write a short note after each review: the measures read, the need found, the next step and its date, and what went well. Give the person a copy. At the next review, start by reading whether the next step happened and whether the measure moved. Over a year, the notes show how the role grew and what the business did to help, which is the record a fair business would want to have.
A dental office in Minneapolis with four hygienists began quarterly people reviews after patients complained about long waits. Before the reviews, the office manager pulled each hygienist's measures from the scheduling system: appointments that ran more than 10 minutes over, out of all appointments in the quarter.
Three hygienists ran over on about 8 percent of their appointments. The fourth ran over on 22 percent. Read as raw counts it looked worse still, but the manager compared rates, and asked the fourth hygienist what got in the way. The answer was specific: she was the only one booked for new patients, whose first cleanings always take longer, and the schedule gave them the same 60 minutes as a routine visit.
That was a process need, not a performance problem. The office lengthened new-patient slots to 80 minutes and spread new patients across all four hygienists. At the next quarter's review her over-run rate was 9 percent, in line with the others, and the office's wait complaints fell by more than half. The review note recorded the change and why it was made.
The manager noted one more thing. Had the review started from the complaints, or from a feeling that one hygienist was slow, it would have ended in blame and a resignation, and the schedule would have made the next hygienist look just as slow.
US hospitals and airlines use blame-free reporting for the same reason as this lesson: people report problems they expect to be fixed, and hide problems they expect to be punished for.
A people review is about whether the owner is happy. It is about what the role's evidence shows.
A problem means someone is at fault. Most are support or process needs.
Everyone failing proves everyone is careless. It points to the process.
Good results need no mention. Name them; it is how good practice spreads.
The most failures means the worst performer. Compare rates over the work done, not raw counts.
Read the measure.
$\text{on time: } 85 \text{ percent; target } 90$
The quarter's evidence.
Find the gap.
$90 - 85 = 5 \text{ points}$
Below target.
Ask what got in the way.
$\text{the bridge works on the route}$
The driver knows.
Decide the kind of need.
$\text{the schedule, not the driver: process}$
Anyone driving that route would be late, so it is not the driver.
Agree one next step.
$\text{bridge deliveries first, from Tuesday}$
Read at the next review.
Read the pass rates.
$85, \ 90, \ 60 \text{ percent}$
One for each cleaner.
Find the shared failure.
$\text{the oven, for all three}$
Everyone shows it.
Fix the process.
$\text{oven onto the standard work and checklist}$
A process need.
Find who stands apart.
$\text{Cleaner C at } 60$
Even aside from ovens.
Ask what got in the way.
$\text{bathroom standard never shown}$
A support need.
Agree one next step.
$\text{training booked for Friday}$
With a date.
Read the comeback counts.
$\text{Dev: } 4 \text{ of } 80; \text{ Sam: } 3 \text{ of } 30$
Different amounts of work.
Find Dev's rate.
$4 \div 80 \times 100 = 5 \text{ percent}$
Over the work done.
Find Sam's rate.
$3 \div 30 \times 100 = 10 \text{ percent}$
The same way.
Compare the rates.
$10 - 5 = 5 \text{ points}$
Sam, not Dev, stands apart once the work done is counted.
Ask what got in the way.
$\text{no torque wrench of his own}$
Borrowing Dev's.
Decide the kind of need.
$\text{a tool: support}$
Cheap to meet.
Agree one next step.
$\text{buy the wrench this week; read next quarter}$
And name Dev's good record.
Read the measure.
$\text{unbroken: } 88 \text{ percent; target } 96$
Eight points short.
Ask what got in the way.
$\text{the new mugs don't fit the dividers}$
Anyone would struggle.
Name the need and the next step.
Findings from Northside Repairs' quarterly people reviews. Sort each by what it points to.
| A support need | A process need | No action needed | |
|---|---|---|---|
| Never trained on hydraulic brakes, and gets them wrong | |||
| Every mechanic misses the same check, which is not on the job tag | |||
| Has no torque wrench of their own | |||
| On target all quarter, no complaints |
Complete the worked solution: at Bright Home Cleaning, Cleaner C passed $10$ of $20$ move-out inspections first time last quarter, while the other cleaners averaged $85$ percent. Find C's pass rate, the gap to the others in percentage points, and how many more inspections C would have needed to pass to match them.
Find C's pass rate.
$(\text{passed}) \div \text{twenty} \times \text{a hundred} =$ p
Percent passed first time.
Find the gap to the others.
$(\text{others' rate}) - (\text{C's rate}) =$ g
Percentage points.
Find the inspections short.
$(\text{gap}) \div \text{five} =$ n
Each of twenty is five points.
Say what it points to.
$\text{one person below the rest: support}$
Ask what C was never shown.
At Bright Home Cleaning's people reviews, all $5$ cleaners' move-out cleans failed inspection on the oven more often than anything else. What does that point to?
Notes prepared for a people review of Maya's assistant. Mark every note that comes from the role's evidence.
This task has no paper form; do it on a device.
Put Neighborhood Kitchen's quarterly people review with its driver in order.
Number the steps in order (write the number in the box):
Maya's assistant packed $50$ orders last month and $41$ arrived unbroken. The target is $96$ percent unbroken. How many percentage points short of the target was the assistant's rate? Fill in each figure on the sheet.
| Amount | |
|---|---|
| Unbroken rate, percent | |
| Points short of the target |
A restaurant in Chicago tracks order errors at its people reviews. One server had $14$ errors in $200$ orders last month; the rest of the team averaged an error rate of $3$ percent. How many percentage points above the team's rate was the server's error rate?
Answer:
A hotel housekeeping team had $20$ room inspections fail last month, across five housekeepers. $12$ of them failed on the same item: the coffee maker not refilled. Complete the sentence.
The coffee maker accounted for p percent of failed inspections.
Lesson test: one question per skill, one attempt each, no hints. Your answers are checked when you submit.
Bright Home Cleaning's three cleaners each had $20$ move-out inspections last quarter. Passed first time: $15$, $20$ and $10$. Fill in each cleaner's first-time pass rate, as a percentage.
| First-time pass rate, percent | |
|---|---|
| Cleaner A | |
| Cleaner B | |
| Cleaner C |
You can run a people review that finds what someone needs rather than who to blame. Tell someone why 'is it one person or everyone?' is the key question. Next: naming the evidence an operations decision still needs.
20. Your turn: Maya's assistant, step 3
$\text{process: new dividers, ordered Monday}$
Read at the next review.