- Team ThinkAg
'Planting Intention', 'Production Dynamics' & 'Warning Signals' - ushering the 'Data Democracy'
The HummingBird team makes a case for upgrading Indian Agriculture with sharp insights from the farm and enhancing productivity with capabilities to learn and predict from smart data
Bharat ( India) contributed the most essential numerical value, without which no bank account today will be worth its envy. In the debate between basing judgements on scientific outcomes versus judgements based on experience, we as a country chose the later process for many of our decision making. Though, given the deep scientific temper that existed in the early years of the formation of the Land beyond the Hindukush mountains, it gave way to judgemental values more importance, which slowly gave way to another word, which found its way into the Webster dictionary= “JUGAAD”.
However the as the economy grew and many economic agents realised the importance of science in decision making, so did the respect for data grew. As technology enveloped many of the core and peripheral sectors, Agriculture-the largest part of the economy somehow kept itself aloof from the tech revolution wave. The entire agriculture ecosystem continues to adopt a “Jugaad” approach across the ecosystem, largely because it is besieged by its structural design, and the emotional quotient attached to the underlying activity.
The entire data collection program is based on a designed as conceived by Sher Shah Suri, almost about 500 years back, wherein the local village head known differently across the country as “ Talati, Patwari, Kulkarni” who provide what his village is sowing or how the crop is shaping up. Over time, they remain key assets, even though we have been able to launch a satellite to observe Mars.
Given the Giga structure of the entire agri economy, even with so-called infrastructure across the administrative structure, the information evaporates, leading to a decision which remains shrouded in mystery & reactive based on sentiments rather than data- this is across the value chain.
Hummingbird: The Journey
Opportunity always remains a key inspiration for any entrepreneur. Opportunity to create value for the ecosystem, solve a problem which never existed or remained untouchable. Given the extensive experience, we have had in the looking at number be it market research, financial industry or technology – to dig deeper into the numbers to procrastinate came naturally. When we applied our standard yardsticks to look at the agriculture sector, we realised that there is only a Yard but no stick, no benchmark which can help to measure or gauge. It was like starting raw, from zero. Further, when we stepped into the soil, we saw the entire ecosystem was consistently consistent in the Jugaad framework. All the numbers and insights were done using the back of the mind envelope calculation with “little adjustments”. When we further dug deeper into the policymaking process, we didn't find much difference in approach. However, there was a good amount of sophistication that was being adopted by just a handful of the key players.
The classic words from the world of trading kept on haunting us “INFORMATION GAP LEADS TO GREAT ARBITRAGE OPPORTUNITY”. As it became increasingly evident that the Given the Information Asymmetry in the Agricultural sector, the terms of trade were completely skewed towards those who had access to the data.
Thus began the journey towards building the country’s first Agricultural information system (AIS)- HUMMINGBIRD since Sept 2017. One thing we surely realised is that HUMMINGBIRD will be ushering in Data democracy as we kept moving each day.
Based on our experience in the sector and our previous experience in the financial and other related industry, we were able to translate our thought process into a substantial offering.
We realised we are addressing the following key questions at all time:
What is the intent of farmers concerning the sowing of crops & consumption of inputs like tractors, agro chem etc?
How are the agro markets behaving, prices & arrivals?
What will be the farmers' income & its repercussion be on loan repayment?
What do the coming months look like for Rural demand?
What is the Agricultural emerging scenario for the next quarter?
As we listed down these key questions, one thing which emerged was that these 5 questions, contained everything that one would need to have to be able to provide a meaningful response.
Our first Step was to Build a Data Structure:
One thing which was certain that, to be dependable, reliable and credible our data sources and collection mechanism had to be extremely robust. We had to build very high-level hygiene factors at each stage of the collection of data. Given our prior intensive knowledge of collecting multifaceted data sets across the entire economy, which would entail consumer behaviour or others, we knew exactly the kind of challenges we were up against. Further, being agriculture all the more we had to ensure the ground network that we were planning to set up to collect the data was robust enough. As our data source would be only the producers directly. One thing was certain, its impossible to map all the millions of farmers spread across 5 lakh villages. However statistical science has developed robust techniques which if applied judiciously, can surely provide the much-needed impact, as it would be covering each farmer.
We often get asked, if the sample size we have selected is large enough, to measure the phenomenon under consideration. Much of the confusion in people's minds arise from the notion that sample accuracy is linked to the size of the universe we are measuring. This confused mental linkage of sample size and universe size should vanish if we take examples from our normal lives - To see if rice is cooked, we just need a spoonful scooped from the pot. We don't need to take more spoonfuls if the pot is bigger, or less if the pot is smaller. Of course, skilled researchers will take care to select those few scoops from different randomly selected spots to ensure representativeness.
Interestingly the Agrisk panel is unique in the world and there are no comparable data sets available of farmer panels anywhere else in the world.
The Data Collection Mechanism: Bifocal- Feet in the Soil & Eye in the Sky
The Outputs: So we decided to compartmentalise the responses, and put them under various headings, which turned out to 7 different elements. Each component having a different data point and insight.
The Most Unique feature of the Platform, HUMMINGBIRD is its ability to report the crop status based on the stage at which it is, this gives far sharper insight and early warning signal in case any drastic changes are happening on the ground.