The Indian edtech sector has a measurement problem, and it is a simple one. Nobody agrees on what “impact” means.
Product A measures learning gains with a pre-post test aligned to its own curriculum. Product B tracks “engagement”, meaning time on the app and sessions per week. Product C reports teacher satisfaction scores. Product D says its impact is “reach”, the number of schools using it.
A funder looking at all four cannot compare them. A school principal choosing between them has no common yardstick. A government education department trying to bring edtech into its system has no framework for deciding which products are worth the time and attention they take up.
The common framework was built to solve this problem.
What the framework does
The framework sets a standard for what counts as an outcome across edtech products, so that a funder, a school or a government department can compare like with like.
It works at three levels.
Level 1 is learning outcomes. It asks whether the product changes what a child knows or can do, measured with curriculum-aligned assessments such as ASER-style tests, NAS-aligned tests or competency tests independent of the product. This is the hardest level to meet, and most products avoid it.
Level 2 is pedagogical process. It asks whether the product changes what happens in the classroom, including teacher behaviour, student participation, time on task and the quality of questions. It is measured through classroom observation protocols.
Level 3 is adoption and usage. It asks whether the product is being used as intended, measured by download rates, active users, frequency of sessions and uptake among teachers. This is the easiest level to claim, and it is the one most often reported as “impact”.
The politics of the framework
Asking edtech companies to adopt a common evaluation framework is like asking restaurants to agree on a shared food safety inspection. Everyone says they support it and nobody wants to go first. Companies with the weakest evidence prefer the current arrangement, in which each defines its own measures of success. Companies with strong evidence prefer the framework, because comparison favours them.
The framework dealt with this by making Level 3 the entry point, since adoption is easy to report and threatens no one, and then raising expectations over successive funding cycles. A funder using the framework could ask for Level 3 in the first year, expect Level 2 by the second, and expect Level 1 by the third or an explanation of why it was not possible.
This gradual approach worked because it gave companies time to build the capacity to evaluate, instead of demanding it all at once.
What we learned
The main lesson concerns the politics of measurement, which is also the subject of the book we are writing. When you standardise what “impact” means, you change who wins. The companies that reported reach as their main measure were the ones with the least evidence of learning gains, and the framework made that visible.
Measurement is power, because the choice of indicator decides the story. When every product tells its own story with its own measures, nobody can tell whether the sector as a whole is helping children learn more. The framework gave the sector a shared language, and the resistance to it showed which companies were confident in their products and which were not.
The same thing happens in development more broadly. When a country defines its own poverty line and measures progress against it, the definition shapes the result. If the line moves, the number of poor people changes overnight, though nobody has eaten a better meal.
Measurement and politics cannot be separated, and that idea runs through all of this work.