"Data Alone Won't Feed the World": A Conversation with Dr. Priya Nambiar on AgriTech, Smallholder Research, and the Future of Agricultural Market Intelligence
Introduction
Dr. Priya Nambiar is a Senior Research Director specialising in agricultural markets and rural consumer behaviour, with over 18 years of experience conducting field research across Sub-Saharan Africa, South Asia, and Southeast Asia. She has led market intelligence programmes for the International Food Policy Research Institute (IFPRI), several leading agribusiness multinationals, and a number of development finance institutions. We sat down with her to discuss the rapidly evolving landscape of agricultural market research — from the explosion of remote sensing data to the enduring importance of qualitative fieldwork in farming communities.
The State of Agricultural Market Research Today
Let's start with the big picture. How has agricultural market research changed over the past decade?
Dr. Nambiar: It has changed profoundly, and in some ways it has become almost unrecognisable from what I was doing when I started out. The most obvious shift is the sheer volume of data now available. Satellite imagery, IoT sensors on farm equipment, mobile phone-based survey platforms, drone-based crop monitoring, weather station networks — we are swimming in agricultural data in a way that simply did not exist ten years ago.
The global precision agriculture market is now valued at over $10 billion, and the broader agricultural technology sector attracted more than $6 billion in private investment in 2022 alone. That investment is generating an enormous amount of research infrastructure that market analysts can leverage.
But — and this is critical — I think there is a real danger that our community becomes seduced by the abundance of remote and automated data, and loses sight of the fact that the most important agricultural markets in terms of food security, development impact, and frankly commercial opportunity are smallholder farming systems. And smallholder systems are precisely where a lot of this shiny new data infrastructure is least applicable and least reliable.
Can you say more about that tension?
Dr. Nambiar: Absolutely. Consider this: approximately 500 million smallholder farms produce roughly 70% of the food consumed in developing countries. These are farms often under five hectares, operated by families with complex livelihood strategies, operating in informal markets, making input and output decisions based on a combination of agronomic knowledge, social norms, liquidity constraints, and risk perception that no satellite can capture.
When an agribusiness client asks me to assess the market opportunity for a new crop protection product in, say, smallholder maize systems in East Africa, I can look at yield gap data from remote sensing, I can look at input market distribution data, I can look at mobile money transaction data. All of that is useful contextual intelligence. But if I do not also go and sit under a tree with a group of women farmers and understand why they make the decisions they make — what they fear, what they aspire to, who they trust — I will get the research fundamentally wrong.
"The farmers I research are not data points. They are extraordinarily rational actors operating in environments of deep uncertainty. Understanding their logic requires presence, not just processing power."
Research Methodologies and Tools
What methodological approaches do you find most effective for agricultural market research, particularly in emerging markets?
Dr. Nambiar: I use a deliberately mixed architecture for almost every major research programme. On the quantitative side, we have become very sophisticated users of Computer-Assisted Personal Interviewing (CAPI) platforms — SurveyCTO, ODK, and Kobo Toolbox are all widely used in agricultural field research — combined with GPS-stamped farm plot data and integration with remote sensing layers. This allows us to build statistically robust, spatially referenced datasets at scale.
For segmentation work, I am a strong advocate of latent class analysis rather than simple cluster segmentation. Farming household segments based on land size and crop type miss the enormous heterogeneity within those groups in terms of technology adoption propensity, risk appetite, and market orientation. Latent class models, properly specified, reveal segment structures that actually predict behaviour.
On the qualitative side, I rely heavily on a combination of focus group discussions with homogeneous groups — separated by gender, age, and farm type — and in-depth interviews using life history and seasonal calendar techniques. The seasonal calendar approach, where you map a farming household's full annual activity and cash flow cycle, is particularly powerful because it reveals the timing constraints and liquidity dynamics that determine when and whether a farmer can invest in new inputs or technologies.
How do you handle the challenge of conducting survey research in low-literacy environments?
Dr. Nambiar: This is one of the most important methodological questions in agricultural research and it does not get nearly enough attention in mainstream market research literature. The default assumption — that surveys can be directly translated and administered — fails in many smallholder contexts for multiple reasons: low literacy, low numeracy, cultural conventions around answering questions from strangers, and different conceptual frameworks for time, quantity, and probability.
Several adaptations matter enormously. First, picture-based survey instruments — where response options are illustrated visually — dramatically improve data quality in low-literacy contexts. The 5-2-1 visual scale, for example, outperforms Likert scales in many agricultural research contexts. Second, working with highly trained local enumerators who are native speakers and ideally from the farming community themselves is non-negotiable — not just for translation but for cultural mediation. Third, cognitive pre-testing of survey instruments with target respondents before fielding is essential, not optional.
Market Sizing and Commercial Intelligence
When you are doing commercial market sizing work for agribusiness clients, what frameworks and data sources do you typically rely on?
Dr. Nambiar: The analytical foundation for agricultural market sizing is usually a combination of crop area and yield data — sourced from FAO STAT, national agricultural census data, and increasingly remote sensing-based crop mapping — combined with input usage rate data and distributor-level sales data where accessible.
For demand-side intelligence, I build what I call a Farmer Decision Architecture — a structured model of the inputs, constraints, and information sources that drive adoption decisions for whatever product or technology the client is commercialising. This draws on the behavioural economics literature — specifically technology adoption models like the TAM and the Rogers Diffusion of Innovations framework
— but adapted for agricultural contexts where social proof, peer networks, and trusted extension agent relationships play a far larger role than in consumer technology adoption.
Industry bodies like CropLife International, national agricultural research systems (NARS), and development organisation research programmes (CGIAR, USAID Feed the Future) also publish invaluable market intelligence that commercial researchers frequently overlook.
The Future of Agricultural Market Research
Where do you see the field heading over the next five years?
Dr. Nambiar: Three developments excite me most. First, the integration of earth observation data with ground-truth survey data is becoming genuinely powerful — we are now at a point where, in certain geographies, you can build reliable farm-level yield prediction models by combining satellite-derived NDVI data with relatively small-n ground survey programmes. This dramatically reduces the cost of market monitoring at scale.
Second, the proliferation of mobile phone ownership in agricultural communities — even feature phones — is enabling longitudinal panel research with farming households in ways that were logistically impossible a decade ago. Being able to re-contact the same farming families across multiple agricultural seasons gives us an understanding of adoption trajectories and farming system dynamics that cross-sectional research simply cannot provide.
Third, I am watching with great interest the development of AI-assisted qualitative analysis tools that can process large volumes of in-depth interview and focus group transcripts in multiple languages. If these tools can be developed in ways that preserve the nuance and context-sensitivity that qualitative agricultural research depends on, they could be genuinely transformative for our ability to scale insights from deep field research.
But through all of this technological evolution, my fundamental conviction remains the same: the most important thing a market researcher in agriculture can do is spend time in farmers' fields, listening carefully and with genuine respect for the expertise and experience that farming families carry. Data alone will not feed the world. Understanding will.
Dr. Priya Nambiar's research has been published in Food Policy, World Development, and the Journal of Rural Studies. She is a member of the Advisory Board of the Agricultural Market Information System (AMIS) and a Fellow of the Market Research Society.