John Deere: Betting the Farm on Machine Learning

Deere’s acquisition of Blue River Technology underscores its focus on using machine learning to improve agricultural efficiency

In September 2017 Deere & Company agreed to acquire Blue River Technology (“Blue River”), a Sunnyvale, California-based provider of vision-based weed targeting systems.[1] While some may question the $305 million headline price, others have compared this investment to the Company’s 1999 acquisition of NavCom, which helped to springboard Deere into a leadership position in GPS-based agricultural solutions.[2] Will Blue River have the same effect for machine learning-based solutions?

The Need for Machine Learning in Agriculture

With the world’s population expected to continue to balloon and the amount of available agricultural land expected to shrink, there is an increasing need for farmers to become more efficient with their resources.[3] Without additional land dedicated to farming, much of this gap will need to be addressed by technological improvements. Deere and other agricultural equipment manufacturers have responded to this challenge by emphasizing “precision agriculture” solutions geared toward increasing farmers’ yields while lowering their costs.[4] As sensor technology becomes more and more affordable, Deere and other companies are arming farmers with the data and technology that they need to meet these goals.[5]

One specific area where there is opportunity for increased precision is pesticide usage. Pesticides are an essential component of the agriculture industry, but they also have some clear downsides. Pesticides are toxic for the environment and humans, and they are also expensive for farmers.[3] These downsides illustrate the need to use pesticides as efficiently and accurately as possible, and Deere, through its acquisition of Blue River, hopes to find a solution to that problem with machine learning technology.

Deere Investing in Machine Learning

In acquiring Blue River, Deere is directly responding to its customers’ desires for increased efficiency and technology. “As a leader in precision agriculture, John Deere recognizes the importance of technology to our customers. Machine learning is an important capability for Deere’s future,” said John May, President of Agricultural Solutions and CIO at Deere.[1]

Blue River’s “See and Spray” system is able to distinguish weeds from crops and apply targeted sprays of herbicide (or other substances) to the weeds with very high precision.[5] In the past, farmers have made decisions on what plants to spray with herbicide on a field-by-field basis.[3] By targeting only weeds, this spraying technology allows farmers to reduce herbicide usage per acre by ~95% while also improving yield.[4] In addition to reducing farmers’ costs, the technology may broaden the range of herbicide options available to them since it is able to precisely target and spray only the weeds.[5] While the technology was Initially designed with lettuce and cotton fields in mind, the Company is optimistic that See and Spray can be applied to a range of crops, including soybeans, peanuts and corn.[5]

In addition to the See and Spray product, Deere’s interest in the acquisition was largely driven by the Blue River’s machine learning expertise, which management hopes to apply to other products within the Deere portfolio.[6] The See and Spray system’s machine learning capabilities enable the system to constantly check its work and autocorrect so that its precision increases with additional usage.[6] It is believed that this AI engine can be applied to a range of other farming tasks, including soil tilling, seed planting, fertilizer and nutrient spraying and harvesting. [4]

A Look into the Future

Deere’s acquisition represents a huge first step toward improving herbicide efficiency for farmers, but the problems of agricultural efficiency are far from solved, and there are several steps that Deere will need to take in both the short and medium term to commercialize and spread this new technology. In order to maximize the impact of this technology, Deere will need to work with the Blue River team to develop solutions that can work across a range of crop types. In order to commercialize and roll the solution out across the country Deere will need to conduct thorough product testing to make sure that the solution is robust enough to be effective in any geography and for any crop.

Concurrently, Deere will need to invest significantly in developing the market and educating customers. Farmers can be averse to change, especially when technology is involved, so the Company will need to educate farmers on the product and will need to effectively communicate the savings that this solution can bring.[3]

Deere’s bet on Blue River, and on machine learning in agriculture, has the potential to create massive efficiency savings within the industry. Several questions remain however that could determine the ultimate success of this initiative. Will farmers be receptive to this new technology, or will the market development required to gain traction be insurmountable? How should Deere integrate the Blue River team – can they preserve Blue River’s entrepreneurial culture while also enabling the collaboration with Deere that will be needed to commercialize the technology?

 

(792 Words)

 

Sources

[1] Deere & Company. “Deere to Advance Machine Learning Capabilities in Acquisition of Blue River Technology”. https://www.deere.com/en/our-company/news-and-announcements/news-releases/2017/corporate/2017sep06-blue-river-technology/. 6 September 2017. Accessed November 2018.

[2] James Bourne. IoT Tech News. “The AI and Machine Learning Innovations Taking John Deere to the Next Level of Precision Agriculture”. https://www.iottechnews.com/news/2018/mar/22/ai-and-machine-learning-innovations-taking-john-deere-next-level-precision-agriculture/. 22 March 2018. Accessed November 2018.

[3] Bernard Marr. Forbes. “The Incredible Ways John Deere is Using Artificial Intelligence to Transform Farming”. https://www.forbes.com/sites/bernardmarr/2018/03/09/the-incredible-ways-john-deere-is-using-artificial-intelligence-to-transform-farming/#2000fb7b330d. 9 March 2018. Accessed November 2018.

[4] Adele Peters. Fast Company. “How John Deere’s New AI Lab is Designing Farm Equipment for a More Sustainable Future”. https://www.fastcompany.com/40464024/how-john-deeres-new-ai-lab-is-designing-farm-equipment-for-more-sustainable-future. 11 September 2018. Accessed November 2018.

[5] Willie Vogt. Farm Industry News. “Super-Targeted Sprayer Getting Closer to Market”.  https://www.farmindustrynews.com/technology/super-targeted-sprayer-getting-closer-market. 23 August 2017. Accessed November 2018.

[6] Willie Vogt. Farm Industry News. “John Deere Makes a Machine Learning Buy”. https://www.farmindustrynews.com/technology/john-deere-makes-machine-learning-buy. 6 September 2017. Accessed November 2018.

 

 

Previous:

Capturing Value Through Machine Learning, Shell Adapts The Era of Low Oil Prices

Next:

Lily Health – Scaling access to reproductive health advice with machine learning

1 thought on “John Deere: Betting the Farm on Machine Learning

  1. The social and environmental benefit of adopting this technology are striking. To reduce herbicide use by over 90% not only will make food safer, but it will make our soil healthier and reduce waste in the Ag value chain. However I’m curious how this technology has improved yields? I would assume the see & spray can not be 100% accurate, so there may be some “misses” in spraying the pesticide that result in crop death, whereas blanket spraying the entire fields with pesticide would presumably protect all plants and maximize yield? In any case, if the technology is affordable and improves yield, Deere could drive adoption by letting farmers test the technology and see the results.

Leave a comment