How precision agriculture eats data and CPU cycles
Given their rural isolation, toiling miles away from the coastal centers of technology and finance, it’s easy to overlook the remarkable growth in productivity on the average farm.
One look at a chart of corn yields, which have increased four-fold since the early 1950s, shows that farmers are marching to their own version of Moore’s Law. While past improvements were the result of better plant hybrids, fertilization and production equipment, information technology will be the key to sustaining and perhaps accelerating agricultural productivity.
Precision agriculture, a collection of data collection, analysis and prediction technologies that looks like something out of Google, not John Deere, describes a group of technologies designed to collect and analyze detailed information about growing and crop conditions that feed complex models designed to provide actionable recommendations to improve yields and reduce costs.
Although precision agriculture is an important tool for feeding a growing planet while minimizing environmental damage, the motivation for farmers is less altruistic. According to Eduardo Barros, Accenture’s global products agri-business lead, data-driven decisions about irrigation, fertilization and harvesting can increase corn farm profitability by $5 to $100 an acre.
Barros adds that a six-month pilot study found precision agriculture improved overall crop productivity by 15%. It seems like a no-brainer for farmers if not for the nasty implementation details: new sensors and equipment for granular data measurement, data collection, integration with third-party data sources such as weather models and satellite imagery, and number-crunching data analysis to produce recommendations.
While not insurmountable hurdles for big corporate farms, the technology requirements and expertise are beyond the reach of smaller farmers, particularly in developing countries. Enter cloud services: the same technology equalizer that allows two-person startups to develop software using hundreds of servers can deliver sophisticated agricultural analytics to the family farm.
By combining aspects of IoT and big data, precision agriculture has a lot in common with burgeoning analytics applications in many other industries. The need for prodigious data collection, from many sources, associated storage and computational horsepower makes it a great fit for cloud services. Not only do shared services broaden the available market for precision agriculture, but the cloud enables agricultural crowdsourcing, by aggregating data from a wide variety of smaller operations to improve prediction models.
The field has already attracted the attention of big companies like IBM, which has researchers working on agricultural weather forecasts, models and simulations to improve farm decisions, and Accenture, along with a host of startups as profiled in this Forbes column. Yet farming is a hands-on activity and many of the measurements that feed precision agriculture models require instruments and implementation expertise that small farmers don’t possess.
That’s why Accenture has segmented its offering into two services: one for large agribusiness with the necessary equipment and sophistication to use a pure SaaS product and another for small operations, particularly in developing countries, that rely on an agricultural version of the channel: agro-service agents that work directly with individual farmers. In this case, Accenture’s software provides decision support for companies that already sell a range of agricultural products like seeds, fertilizer and pesticides. Barros says Accenture’s software can even integrate with ERP and HR systems to automate orders and schedule field workers.
An important similarity between precision agriculture and broader trends in business software is the use of location services. Of course, farms are inherently tied to location, making agriculture a natural early adopter of GPS services, such as fertilizer spreaders that can apply different amounts according to location, and autonomous vehicles.
Drones represent the next frontier for data collection (field imagery) accuracy and frequency and perhaps product application (fertilizer, herbicides). Just like an array of equipment sensors in a power plant or aircraft, all of this fine-grained location data can feed analytic models, however as with industrial IoT, the amount of data can be overwhelming, reinforcing the case for cloud deployment.