25% growth in teaching practice across 250 teachers

Pallavi already observed its teachers. What came back was the problem every school knows: a rating without a reason. A 3 out of 4 tells a teacher nothing they can use on Monday, and by the next term nothing has carried forward.

Early Years made it harder. In Pre-primary and the early primary years, instruction is play, transitions dominate the period, and much of what matters between a teacher and a five-year-old is not verbal at all. Most observation frameworks are written for a secondary classroom and simply do not see it.

The group needed an instrument that worked in their classrooms, and produced something a teacher would actually act on.

Six parameters and eighteen criteria, purpose-built for Early Years. Every rating is made by selecting the written expectation statement that matches what was observed - so the score follows the description, not the other way round.

Two choices shaped everything that followed. Application of Learning is scored on what students do, not on what the teacher delivers - which stops the framework rewarding performance over effect. And Evaluating the Session requires the teacher's own reading of their lesson, making self-assessment a rated instrument rather than a comment box.

Each of the 250+ teachers was observed four times - once externally, once internally, once by a peer, and once by themselves. At this scale that is near-universal coverage rather than a sample.

The power is in comparing them. Four ratings of the same teacher on the same eighteen criteria show where a self-assessment diverges from every external view, which strengths hold across every perspective, and where a teacher is stronger than they believe. None of that is visible from a single annual visit.

The improvement held at aggregate level across the group rather than appearing only in individual cases - evidence that the instrument was measuring something real, and that the intervention loop around it was working.

Thirteen months inside live classrooms shaped the platform in ways no specification would have. Each of these exists because a teacher or an observer needed it.

• Teacher-first access. Early Years teachers are rarely at a desk. They needed their own entry point, on their own device, with no training required.
• Works on any school network. Shared devices and weak classroom signal drove offline capture with later sync - an observation form that fails at the back of a Pre-primary room is not a form.
• Reports a teacher reads without help. Rewritten until no explanation was needed. This became the plain-language standard across every IMPACT report.
• Live KPIs. Ratings move as the year moves rather than sitting as a snapshot per visit - the origin of the running teacher profile.
• Evidence against the criterion. Board photographs, student work and worksheets attach to the specific criterion they support. Early Years evidence is overwhelmingly visual.
• Practice shared across schools. With ten schools in one group, what worked in one was invisible to the others. A sharing layer changed that.

Pallavi established that the IMPACT architecture - parameters resolving into criteria, criteria rated against written expectations, multiple stakeholders on one shared rubric - works at scale in live schools, and moves the number it is designed to move.

An observation framework that survives Pre-primary generalises upward easily. One built for Class 10 rarely survives the journey down.

Why we built the alpha in Early Years