Why Parent Coach works this way

Not to make you a compliant parent. To make you a parent who notices their own patterns and decides for themselves.

There is a paper that describes this better than we could. Sanders and Mazzucchelli, writing in 2013 about what parenting programmes are really for, argue that the target is not a set of techniques a parent applies correctly. It is parental self-regulation — a parent's capacity to watch their own behaviour, set their own goals, judge their own results, and adjust. Techniques are the means. The durable outcome is a parent who can run that loop without anyone standing over them, and who has, in their words, "a greater sense of personal control and mastery".

That is the whole product, so it is worth being precise about how it maps:

Why this matters more than it sounds. It decides what counts as success. If the goal were compliance, the obvious thing to measure would be whether you followed the advice — and the obvious thing to optimise would be advice that people say they like. Because the goal is your own judgement, we deliberately do the opposite: advice is never ranked by whether it was pleasant to read.

You are often reading this at the end of a bad day. Anything that arrives uninvited and tells you what to do lands as blame, however carefully it is worded.

So information and intensity only ever increase when you tap. Never because you logged more, never because you logged worse, never because you ignored the last one. There is a ladder, and you choose the rung:

It also changes how everything is written. Not "try this" but "something you could try". Not "what to avoid" but "what tends to backfire". You decide; the app supplies what is known.

One exception, and it is absolute. Being gentle never means being vague about a risk. Where something could hurt a child — separate them if someone is being hurt, and this never applies to a child who is distressed rather than aggressive — it is said plainly, first, and never hidden behind a tap.

Your response is the half of the exchange you can actually change.

There is a mechanism behind this, and it is more interesting than "be a better parent". Across a large body of work on how parents think, what a parent believes a behaviour means shapes how they respond to it — and reading a behaviour as deliberate defiance is associated with responding more harshly than reading the same behaviour as a tired four-year-old (Bugental & Johnston, 2000).

Which makes the act of logging more than record-keeping. Writing down what happened separately from what it meant is a small intervention on the interpretation. The pause is doing work before any advice arrives.

It also means the most valuable thing this app can say to you is about your own move — and that is the easiest thing in the world to hear as an accusation. So anything about your behaviour sits behind an explicit invitation: want to look at your side of this? It is never the first thing you see.

The app would rather you disagreed with it thoughtfully than followed it blindly.

When the same situation keeps happening — the same trigger, the same reaction from your child, the same response from you — the app will occasionally show you your own record and ask one question: looking at those, is this working? Seems to be helping · doesn't seem to help · too early to tell.

That is not a test, and there is no right answer. If you work out for yourself that holding the limit settled it three times out of five, you have learned something no advice card could have told you, because it is about your child and not about children.

What we deliberately do not do. We do not ask you to rate advice and then use those ratings to decide what to show people. That question is asked before you have tried anything, so it measures whether advice was pleasant to read — and advice that is correct and unwelcome would lose. "Stop pressuring them to eat" is exactly the kind of true, annoying suggestion that would quietly disappear.

Every piece of advice here says where it came from — and what that research does not show.

The first thing you read after logging a moment is written for you, by a model, in the moment — not picked from a list of pre-written replies. It is held to every rule on this page, and it never cites a study, because a model asked to recall a reference will eventually invent one. It offers nothing and suggests nothing: it says what happened and that it makes sense. Everything with a claim in it — the suggestion, the research, the limits — is written by a person, checked against the paper it came from, and sits one tap away.

Every study behind the advice is a real, checkable paper with a link. None are summarised from memory. But a citation is not a guarantee, so each one also carries its limits, in plain words, on the same screen: the ages it was tested on, the size of the sample, whether anyone was randomised. The card's own honest limits sit higher still — they open with the advice, not one tap deeper, because a limit you have to go looking for is not doing its job.

Two examples of how far that goes.

One study is in our library because it argues against us. A review of 41 studies found that praise has mixed effects on whether a child does what you asked in the next thirty seconds. So no advice here promises that noticing good behaviour will stop an interruption quickly, even though that would be a nicer thing to say.

And the assumption underneath this entire product — change what the parent does, and the child's behaviour changes — was tested directly. It held up in under half the studies that properly examined it (Forehand et al., 2014). The programmes still work; what is less certain is why. You will not find a sentence here claiming that link is settled.

Finally, you can switch any of it off. Every study behind the advice can be muted, and doing so stops the advice built on it from appearing. Over time you shape which evidence this app is allowed to use on you.

No grades, no streaks, no comparisons.

A count of what you wrote down is not a count of what happened. A busy week with no logging looks exactly like a calm one, so any chart here says so on its face rather than reading a fall in entries as progress.

Some specific things follow from that:

No streaks. They reliably increase engagement and reliably punish the week your life falls apart. In an app for tired parents, a broken streak is a message about your parenting delivered on the worst day of your month.

Firsts are yours, not your child's. When the app says "first time you've caught this", it means exactly that. It is the first time someone wrote it down, which is a fact about your attention — not a claim about your child's development.

Unknown is never treated as bad. A field you left blank stays blank. Nothing falls through into the negative column because you were too tired to finish.

And a guess is always labelled a guess. When the app suggests what a behaviour might be doing for your child, that is a hypothesis drawn from your own notes. Confirming it would take a qualified clinician watching your child. Every screen that shows one says so.

Worth writing down, so it stays true as the product grows.

Diagnose, screen, or label your child — or you.

Tell you that something you did caused something your child did.

Show you advice you did not open.

Compare your child to other children, or you to other parents.

Cite a study it cannot link you to.

Decide that a piece of research applies to your child after you have said it doesn't.

Suggest a goal that serves only the adults. Harmless self-soothing is not a problem to solve.

These four don't tell you what to do in a hard moment. They are the reason the app is shaped the way it is.

The four clinical frameworks the app borrows — ABC observation, the four functions, PRIDE skills and a slow questionnaire — are explained from scratch, with their sources, in How Parent Coach works.