For the past several weeks, I have been digging through soil test data from seven Indian states as part of my internship with the Ambuja Foundation. The question driving the work was simple to state and hard to answer: are Ambuja's sustainable soil interventions actually improving soil health, and if so, which ones are worth scaling?
The Foundation runs a wide portfolio of interventions across Maharashtra, Rajasthan, Punjab, Gujarat, Uttar Pradesh, Himachal Pradesh, and Uttarakhand: organic inputs, crop diversification, water management, mechanization, and more, all aimed at improving six core soil parameters. Of those six, organic carbon matters most. It is the parameter that best reflects long term soil health, and it is also the slowest to move, which makes it a fair test of whether these interventions are working or just look good in a slide deck.
To answer the question, I combined three kinds of evidence: multi-year farmer-level datasets (574 farmers in Maharashtra alone, plus matched cohorts in Nagaur and Himachal Pradesh), interviews with regional coordinators across all seven states, and my own field visits to talk with farmers and entrepreneurs in Rajasthan and Uttar Pradesh. One honest caveat before the findings: none of these locations have a comparison group of non-adopters, so what follows is correlational, not causal.
The headline is encouraging. Organic carbon is rising in every region tracked over multiple years. It has not yet reached the recommended range anywhere that started below it (soil chemistry does not move fast), but the direction is consistently right. In Maharashtra, the share of farmers below the target range dropped from 78.4% to 65.7% in a single year. In Nagaur, Rajasthan, one of the toughest growing environments in the country, organic carbon nearly doubled in three years, from 0.21% to 0.37%. In Himachal Pradesh, soils that were already comparatively strong got stronger still.
What is actually driving the results that work is, to me, the most useful finding in the whole project. The practices with the strongest evidence, crop residue incorporation, farmyard manure and composting, and soil testing, are also the cheapest and least equipment dependent. In Maharashtra, farmers who incorporated crop residue were 10.1 percentage points more likely to see their organic carbon improve. Practices that stall tend to be the ones that need a supply chain, a machine, or a subsidy to keep going.
Some of the best evidence for impact was not a soil measurement at all. In Gujarat, a local entrepreneur named Kanubhai Mevada spent 900 rupees on a season of cow based inputs and earned 51,902 rupees, a net profit of 51,002 rupees, despite Cyclone Tauktae disrupting the season. Twenty four entrepreneurs like him have been trained across six categories in that state, and 16,685 hectares of cotton residue have been incorporated into soil instead of burned, avoiding an estimated 184,000 tonnes of CO2 equivalent emissions.
To summarize my recommendations, I would suggest the following. Center the cheap, proven practices instead of treating them as one option among many. Recognize that the real constraint is rarely farmer willingness (people already want better soil), it is input supply, machinery access, and unit economics. And invest in the interventions that outlast a project's active support, the Organic Resource Centers, Learning Demonstration Centres, and entrepreneur led production units that farmers run themselves, because those are the ones still standing after the funding cycle ends.