Grounded: Can big data do for agri-business what it's not doing for retail?
Agriculture proving fertile ground for analytical startups
Like fusion energy, big data has consistently promised amazing results – always at some indeterminate point in the future.
We know grocers track every purchase, index them against our loyalty cards (translation: database index keys), and develop rich models of consumer behaviour that have, despite all of the millions of dollars and thousands of hours poured into them, have not moved the needle of consumer behaviour at all.
Which should be making us all wonder: is that really all there is? Is big data really just about gathering enough information about a consumer to target ads to them that they’ll block as unneeded and annoying?
Big data grew up alongside the Web, was fuelled by the Web, and because the Web has been thought of as a consumer’s medium, people connected big data with consumer behaviours.
There’s nothing inherent about that relationship. It’s just the first place it landed, because people were able to quickly connect the dots. Whether big data was really useful in marketing...well, that’s a question marketers have so far refused to ask, because some answers lay waste to the assumptions they’re founded upon.
Instead, the payoff for big data gets pushed further into the never-never, while nervous retailers continue to pour millions into solutions no one has yet proven they need. Because progress and competition and me too.
The whole situation is both tragic and comic, depending on whether you’re taking or being taken in by this 21st-century snake oil. Big data isn’t a bad idea, but it’s the sort of hammer that makes everything else look like a nail.
Big data for bumpkins is another matter
Despite its failure to perform, over the last year it’s become clear that data analytics will assume a foundational role in the economy, just not in any of the ways we might have suspected. We needn’t be looking to supermarkets for the raison d’etre of big data, when the answers we seek can be found beneath our feet.
Australia is perhaps the most unlikely country in the world to have a significant agricultural sector. For thousands of years most of the continent has received so little rainfall the soils have dried into dead sands with few nutrients. Yet around the margins – and nearly all of Australia’s farmlands lie within a few hundred kilometres of its coasts – the soils are fertile enough, most of the time, to raise a variety of crops and livestock.
It’s a close-run thing. Droughts of increasing frequency and severity threaten farmers everywhere across the nation. River systems, depleted by years of over zealous irrigation, must now be managed down to the last litre, to ensure there’s enough to go around.
At the same time, agriculture has become professionalised as never before. The ‘Green Revolution’ of the 60s and 70s gave rise to business modelling in farming, allowing farmers to calculate the costs of their inputs – seed, water, and nutrients – against the price of their harvests, running the numbers on feasibility just like any other business would.
Of course, outside conditions – such as weather or competition – could make a muddle of any farmer’s plans. How does a farmer hedge against the variability of weather or markets?
This is where big data can provide a meaningful answer.
Tech startups working on innovations in the agricultural sector (‘agritech’) - have already started to show us a bit of what big data looks like when it comes down to Earth. Perth startup Agworld and Melbourne startup The Yield both offers farmers an analysis of their fields, generating a plan to managing those fields on a metre-by-metre basis. Farmers can know exactly how to manage each area to achieve maximum output from minimised inputs - and presumably will soon be able to program their farm robots to tend those fields according to those results.
Equal parts lean startup methodology, growth hacking, and big data (buzzword bingo anyone?) rolled into an ‘agriculture-as-a-service’ offering, agritech transforms a ten thousand year-old guessing game into an agile process that constantly measures, tweaks, and improves its performance, driving productivity per square metre up as it drives costs down.
Australian farmers have done this before; over the last 30 years they’ve become the most efficient irrigators in the world, because water had become so scarce and expensive. Technological solutions, such as drip-line irrigation, became the norm, and productivity per litre of water soared.
With Green Revolution 2.0, mesh networks of simple sensors, cheap farm robots - another Australian startup, SwarmFarm, is working on that - and data analytics will make Australian farmers the most efficient in the world - at just the moment that’s going to become vitally important.
Two things we know can’t be changed: the planet will warm up by at least another degree over the next thirty years, and there’ll be around to nine billion mouths to feed.
We know that farming will only get harder as the planet warms and the soils dry, so we need every bit of help we can get if we’re going to add another thirty to forty percent to our agricultural output. And we need that revolution everywhere, so that a farmer growing potatoes on the side of a volcano in Rwanda can use his smartphone to photograph his small fields, upload them to a cloud-based AI, which will then tell the farmer how to manage those soils, week-by-week, and crop by crop.
Green Revolution 1.0 allowed us to double the human population in a generation without a corresponding famine - the first time in history that had happened. Green Revolution 2.0 will arrive just in time to rescue us from endless conflicts over diminishing food resources.
It sounds like science fiction, but all of the elements are already in place. Agriculture, the foundation of civilisation, has suddenly become its cutting edge. ®