Go back

In October we asked our Lumos parents how much their children were learning. Then we asked the Lumos students the same question. The answers did not match. The parents were more confident than their children.
We could not have seen that gap from inside the classroom, and that is why feedback loops matter to us. A school cannot see itself. So we ask: every parent and every student, every quarter. We sort what we hear into three piles, which are start, stop and continue. Then we act on it, and we tell everyone what we did.
We do not ask for the sake of asking. Within two weeks of that first survey, we had changed the lessons our students liked least.
What a feedback loop is
A feedback loop is what you get when the result of what you do comes back to you and changes what you do next. A thermostat is the usual example. It measures the room, compares that with the temperature you asked for, and turns the heating up or down. Take the measurement away and the heating just runs.
Anything that gets no feedback repeats itself. For a school that means teaching this year the way you taught last year, whether or not it worked.
Schools have the slowest feedback loop there is
A restaurant knows by Friday whether people liked the food. A software company knows within hours whether a change helped. A school finds out years later, when a child sits an exam or starts a job. By then the child has left.
It is worse than slow. In most businesses the person who pays is the person who uses the thing. In a school the parent chooses and the child sits in the classroom. The parent cannot see the lessons. The child has little to compare them with. Each holds half the picture, which is why we ask both, and why the two can disagree as they did in October.
Slow feedback makes institutions comfortable. Nothing tells you that you are getting worse, so you assume you are fine. We assume the opposite. There is always something to improve, and what worked in September may not work in March, because children change and so does the world they are growing up into. An organisation that is happy with itself has stopped looking.
How fast you can improve depends on how fast you find out, and by that measure schools are built to improve slowly. We cannot speed up the slow loop. No survey will tell us what a child can do at twenty. What we can do is run a fast loop beside it, one that tells us how school is going this term.
Schools get feedback whether they ask for it or not. The kind you do not ask for arrives as a family leaving. We would rather hear it a year earlier, while we can still do something about it.
There is a simpler reason too. We ask children to take feedback on their work every week. It would be strange to exempt ourselves.
What we ask
Every quarter we send one short survey to all parents and another to all students. Parents can answer in German or English and need not give their name. Students answer anonymously. We ran the first round in October, in the fifth week of the school year. Every student answered.
We ask everyone because the feedback a school gets without asking comes from the people who speak up. A survey also reaches the ones who do not.
We ask every quarter because timing matters as much as asking. Feedback that comes once a year arrives too late to use. Feedback that comes every week makes you lurch, like a learner driver stamping on the brake. A quarter is short enough to change course within the same school year and long enough to tell whether the last change worked. We would rather make four small corrections a year than one large reform every five.
We ask whether the child is happy, whether they feel they belong, and how much they are learning. Then come the two questions that matter most: what do you especially like, and what would you most like us to change?
One number: Net promoter score
The first question never changes: how likely are you to recommend Lumos School to another parent?
This is the question behind the Net Promoter Score. People answer on a scale that goes up to 10. A 9 or a 10 makes you a promoter. A 6 or below makes you a detractor. The score is the percentage of promoters minus the percentage of detractors, so it runs from minus 100 to plus 100. A 7 or an 8 counts for nothing, and that harshness is the point. Being liked is easy. The score asks whether people will put their own name behind you.
We tracked this number at Prisma, the school we founded in the United States before Lumos, and we track it here. Our first score at Lumos was 83: 83 per cent of the answers were a 9 or a 10 and 17 per cent were an 8. Nobody gave us 6 or below.

Illustration: Our first Net Promoter Score is 83, which the scale calls world class and we call a good start for a young school in Zurich.
We are glad about that, and we do not lean on it. We are a small school, and one answer can move the score a long way. So we treat the number as a smoke alarm rather than a report card. It tells us whether to worry. It does not tell us what to fix.
Start, stop, continue
What to fix is in the written answers. We sort them into three piles.
Start
French was our students’ least favourite subject, and parents told us the lessons were not meeting their children’s needs. Within two weeks we had changed the lessons, and the children are much happier with them. Parents asked for more photos of the school day, so they now get an album every week.
The gap this post began with gave us the biggest start. We have two explanations for it. One is that the first weeks went into getting to know each child and introducing the subjects, which left less room for new material. The other is that children who learn through projects and discussion do not always notice how much they are learning. So we started on both: more new material, the goal of each piece of work stated plainly, and regular time for students to look back at what they have learned. We do not know yet which explanation is right. The next survey will show whether the gap has narrowed. That is why we ask more than once.
Stop
Parents of our younger children asked how much German their children were hearing. In the first weeks we had leaned towards English on purpose, because some children arrived with no German. We stopped. That class now has days that run mostly in German and days that run mostly in English.
Stop is also where we sometimes say no. Some families asked us to drop sport, because their children already train outside school. We cannot. The canton requires it, and club training does not count. So we said that. But the students had told us something we could act on. Sport was their second least favourite subject, because they did not enjoy the activities we had chosen. We changed our approach straight away, and they are much happier with it.
Continue
This is the pile people forget. Parents told us what they value most: a warm school where children can be themselves, teachers who know each child, and learning through projects and real challenges. Students said much the same, and added the freedom to choose some of their own work. None of that needs fixing. It needs protecting. A school is one system, and you cannot change one part of it without touching the others, so the easiest way to break what works is to be busy fixing everything else.
That is the point of the three piles. Feedback is only worth asking for if you act on it, and when you have acted there is always a next thing to improve. Good enough is never good enough for us.
The half that gets skipped
Asking is the easy half. If nothing comes back to the people who answered, it is not a loop. It is a suggestion box. Three weeks after our survey went out, every family had a written summary of what we heard and what we were doing about it, including the parts we did not enjoy reading.
That reply matters more than the survey, because its effect compounds. When people see their answers turn into changes, they answer more candidly next time. Then we learn more, and the changes get better. It runs the other way just as easily. A school that asks and then goes quiet teaches people that answering is pointless. A school that bristles at criticism teaches them to be polite, and polite answers are worth very little. Either way it knows less every year.
Answering does not mean agreeing. Not every piece of feedback is right, and part of the job is telling which is which. A feedback loop is not a promise to do whatever is asked most loudly, and a decision has to be right for the whole school. But everyone who answers should be able to see where their answer went.
Every school should do this
Plenty of schools survey their parents once a year. We do not know another that asks every parent and every student every quarter and says the score out loud, except Prisma, and we started that one too.
We find that odd, though we think we understand it. The answers can sting, and it is easy not to ask. But parents trust a school with their child for years. Asking them regularly whether they would recommend it seems like the minimum. It costs a short form and the nerve to read what comes back. It is the responsible way to find out whether you are a good school instead of assuming you are.
We will ask again next quarter. We expect to hear things we would rather not hear. That is the point.