The dashboard says: 14-day streak, 312 minutes this month, nine badges, "Level 4 Maths." It is confident, colourful, and it does not answer the only question the parent actually walked over to ask, which is whether their eight-year-old can subtract across a zero without counting on his fingers under the table.
Most parent dashboards are built to reassure. A few are built to inform. The difference is which numbers they put in the largest font.
Three numbers that carry real information
First-attempt accuracy on new material. The word "new" is doing all the work. Many dashboards blend fresh items with review items into one accuracy figure that sits comfortably at 92% and tells you nothing, because review is supposed to be easy. What you want is how often the child gets something right the first time they meet it. Roughly 70–85% is the productive zone. Consistently above 90% means the material is too easy and the time is being spent, not used. Below about 60% means guessing, and guessing compounds — the child is practising the wrong answer as often as the right one.
Retention after a gap. This is the only honest test of learning, and it is the metric dashboards most often lack. A child who scored 100% on the seven times table on a Tuesday and 55% on it two weeks later did not learn it; she held it briefly. Look for a check that happens seven to fourteen days after the original lesson, cold, without a warm-up round. If the app does not schedule that, you can do it in four minutes at the kitchen table with five items.
Attempts before success. An average of 1.1 attempts means the work is too easy. Four or more means the lesson is testing rather than teaching. The number itself matters less than its direction: 3.2 dropping to 1.8 on the same kind of item across three weeks is exactly what improvement looks like, even when accuracy has not moved at all.
There is a quieter fourth: hesitation, or time to first response. A child answering multiplication facts at 100% accuracy but eleven seconds a fact is calculating, not recalling. That is fine at first meeting and a problem by the time she is doing two-digit multiplication, where every fact she has to work out steals attention from the procedure she is trying to learn.
What the popular metrics are actually measuring
| Metric | What it really measures | Where it still earns its place |
|---|---|---|
| Minutes | How long the app was open | Spotting change — a child who did 12 minutes and now does 3 is telling you something |
| Streak | The adult's routine, mostly | Establishing a habit in the first month, then it turns into pressure |
| Badges and levels | The publisher's pacing decisions | The child's motivation, not your judgement |
| Lessons completed | Throughput | Only as a denominator for the accuracy figures |
| "Grade level" labels | One company's mapping of its own content | Almost nothing — these do not line up between apps, and they line up even less across national curricula |
That last row matters if your child is not in a US school. An app's "Grade 3" is a marketing label attached to a content sequence, and it may sit a year either side of what your child's school is covering in Year 3, P4, CE2 or Standard 3. It is not a verdict on your child, and it is not comparable with the teacher's assessment.
The underlying problem with minutes and streaks is that they are inputs. Ten focused minutes on material at the right edge of difficulty beats forty minutes of a child replaying a level he mastered a fortnight ago because it feels good and the coins are guaranteed.
How to read a plateau
Flat lines panic parents more than dips do. In practice a plateau is usually one of four things.
- Consolidation. Accuracy is flat but speed is improving, or attempts-before-success is falling. The child is getting fluent at what she already knows. This is progress the accuracy line is structurally incapable of showing.
- A ceiling in the app. The genuinely new content has run out and the child is cycling through review dressed up as lessons. Check whether new items are still appearing at all.
- A missing prerequisite underneath. A nine-year-old stalling on long division almost never has a division problem; he has a subtraction or times-table hole that surfaces four steps in. Fractions stalls usually trace back to equivalence. Reading-comprehension stalls often trace back to decoding multisyllabic words slowly enough that the sentence falls apart before the end.
- Reading load, disguised. Somewhere around ages eight to ten, word problems and content text get denser fast. Read the question aloud to your child. If accuracy jumps, the maths was never the issue.
A plateau shorter than about three weeks is noise. Past four to six weeks with no movement in any of the three real numbers, change something — but change one thing. Difficulty, or time of day, or format. Change all three and you will not know which one worked.
Fatigue looks different from regression
Both show up as a drop. They call for opposite responses, and the shape of the drop tells you which you have.
It is probably fatigue when the last third of a session is worse than the first third; when the bad sessions cluster after 5:30pm, after football training, or during the week of the school trip; when accuracy falls but attempts-before-success stays flat, meaning the child is rushing rather than lost; and when the next session recovers on its own with no reteaching.
It is worth taking seriously when the drop persists across three or more separate sessions on different days; when it is specific to one skill while everything else holds steady; when attempts-before-success climbs at the same time; and — the decisive one — when the child cannot do it when you ask off-screen, away from the app's format and phrasing.
Expect a real dip after any break of two weeks or more. Number facts and spelling go first, and they come back faster than they went. Two or three short sessions in the final week of a long holiday do more to cushion the return than daily practice through the whole break, which mostly produces resentment. And bear in mind that a dashboard knows nothing about a new baby, a house move, a lost friendship, or a fortnight of broken sleep.
What to say to a child about their own numbers
Children read numbers about themselves as verdicts unless taught otherwise, and the teaching is in how you talk.
- Under about seven: do not show percentages at all. Talk about the specific thing. "You got 'ph' every single time today" is information a six-year-old can use. "78%" is not.
- Eight to eleven: they have already started comparing themselves to classmates, whether you invited it or not. Aim comments at the task, never the person. "That set was too easy for you" describes the set.
- Name what changed, then ask why. "First time round you needed three tries; today you got most of them first go. What did you do differently?" Asked genuinely, that question teaches a child to notice their own strategies, which is worth more than the score.
- Keep the dashboard out of arguments. The moment progress data becomes leverage over screen time or chores, children start gaming it — replaying easy levels, satisfying the streak with a single item, handing the hard set to an older sibling. They are good at this.
- Never put two siblings' dashboards side by side. Different ages, different content, different mapping, and a comparison nobody wins.
- If asked "am I behind?", answer narrowly and honestly. "You're still working on borrowing across a zero. That's what we're practising." Not a ranking, not a reassurance you cannot back up.
A ten-minute monthly review
- Skip minutes, streaks and badges entirely.
- Find first-attempt accuracy on new material. Is it in the 70–85% band?
- Pick five items from three or more weeks ago and ask them cold, offline, with no warning.
- Look at attempts-before-success on one recurring skill across the month. Which direction?
- Ask your child which part they would skip if allowed. The answer is nearly always either too hard or too dull, and both are fixable.
- Change at most one variable, then leave it alone for three weeks.
Finally, hold the whole thing loosely. Any dashboard — the one in Kid Genius World included — can only see what happened inside that app. It cannot see the child who reads fluently to a grandparent and freezes at a timed quiz, or the one who clears every level and still cannot lay out a written method on paper in class. Treat it as one of three sources, alongside the teacher's report and what you observe at your own table. When those three disagree, the software is usually the one that is wrong.