Five years building the technology for the future of agriculture: The three problems we solved in ag
By Jason Tatge, Farmobile co-founder and CEO
. . .
At Farmobile, we often say we’re building the foundational technology for the future of agriculture. But what does that really mean? What is the future of agriculture? And how is Farmobile contributing to it?
The truth is, I can boil it down to three problems. Three problems that have held back the agricultural industry while other industries have eclipsed us in terms of digitization. Three problems, that without being solved, would prevent agriculture from being able to make use of the exciting technologies that we see popping up all around us — Machine Learning, AI, robotics and more.
This month, Farmobile turns five. And to celebrate how far we’ve come, I’m going to dig into the three biggest problems we faced — not just as a company, but as an industry. It’s these problems, and the challenges faced in overcoming them, that our team has been focusing on and committed to solving these past five years.
1. Data from a mixed fleet
Precision ag equipment isn’t new but precision data is very new. Combines and tractors have had the ability to record machine activity for years on the in-cab computers or terminals. For instance, since 2011, most John Deere equipment shipped with a factory-grade modem and the vision of remote connectivity. Savvy farmers, who invested the time required to learn how to use this technology, have used thumb drives to download data from terminals to get it out of the cab.
Ag data today is more like an incomplete report card used to issue a final grade to determine how the multiple combination of decisions made on a field performed. Farmers and their trusted advisors (most commonly an agronomist and/or ag retailer) use this data to analyze strategies from the previous growing season and make recommendations for the upcoming year.
For the entirety of this time, though, there were limits to the type of data farmers and their agronomists could patch together for analysis. You could look at your Deere or Case IH thumb drive data, but it required some real technology chops to bring those two sets of data together for analysis. Each equipment brand has unique designs, signature colors and proprietary data systems making it very difficult to compare and/or merge this data — think of trying to open an Apple Pages document in your Microsoft Word program (here is a link to what that requires).
When we started Farmobile in 2013, one of the things that was most important to us was to enable farmers to collect and compare all of their data, in one place, regardless of make or model of machine.
With Farmobile, subscribers are able to collect data across the largest selection of mixed fleet farm equipment on the market and seamlessly stitch and layer this data together by field to better understand the impact of decisions made. You can check out a full list of compatible OEMs and aftermarket systems here.
2. Real-time, accessible data in one place
A big reason why so many farmers, whom I’ve met across the country, haven’t implemented a digital strategy on their farm is that there is no clear way to make use of it — other than passing it on to an agronomist or seed dealer for Monday morning quarterbacking. Data had no real value to farmers in their day-to-day operations.
When Farmobile began, this was one of the key problems we set out to solve. How could we bring consistent, real-time data collection and visibility to farmers operating a mixed fleet of equipment in a sketchy cellular environment? Furthermore, could we provide live fleet monitoring to farm managers, who weren’t in the fields, so that they could make the real-time decisions needed to impact their bottom line?
I like to describe the way we fixed this problem with an analogy. If you’ve ever purchased a movie or song on iTunes, you know you have to wait to download the entire song or movie before enjoying it. This is how most data collection systems in ag were designed to operate — by sending the full file(s) up to the cloud before they could be used.
The problem here, of course, is that the terminals create bloated files that are very large, and when combined with poor internet or cellular connection often lead to 1) files that fail to upload or get corrupted before they make it to the cloud or 2) the farmer turning off the machine during file transmission due to the amount of time this transfer requires.
This results in incomplete data sets being collected and recorded and requires a lot more speculation than certainty. Farmers have long-expressed a strong interest in seeing a real-time remote view of the machines, so they could make use of their data while in the fields.
Rather than the iTunes model, you can think of what we did as the Netflix option — streaming farm data to the cloud, in a one-second, time stamped, geo-tagged packet rather than waiting for an entire field level file to “complete”. That means our data collection device, the PUC, is constantly sending data to the cloud — we don’t need to wait and send a “complete” file. We use the combination of the PUC to store and send the data and the power of Amazon Web Services (AWS) to help us reprocess this data and stitch it back together when cellular connections may be intermittent and unreliable.
The result is that we’ve engineered a solution to work exceptionally well even in low-connectivity cellular environments. If there’s no internet access, the PUC device simply time stamps and stores the data packets and waits until internet is available and then bursts the data that wasn’t sent during the times of lost cellular connectivity. The result? Real-time, self-healing data sets, that allow for remote visualizations and “in field” decision making to help farmers track their progress, manage their fleet and react to potential threats as they happen.
To get a look at the in-depth visualizations that Farmobile provides, check out this video.
3. Creating the standard data format for data
Perhaps the biggest challenge when it comes to data in ag has been the lack of a standardized format. How are farmers and agronomists supposed to make sense of what’s happening on their fields when each machine and make essentially has its own data language? How do you compare apples to oranges?
When we set out to create a new standard for ag data, we wanted it to be the best — in terms of data quality, data richness, and utility to the farmer and agronomist. The result is our Electronic Field Record (EFR).
Rather than relying solely on the machine’s OEM data systems to collect and upload the data, the Farmobile PUC collects thousands of precision ag sensor data points on a second-by-second, point-by-point basis from multiple disparate systems. The quality of the data you collect is no longer contingent upon the accuracy of your in-machine system.
As PUC streams data to the cloud, it’s building a layered data file that tells the complete story of the field being worked on — from fertility, to planting, to spray applications all the way through harvest. These extremely rich data files allow farmers to manage what’s going on in their fields and makes it easier for agronomists to make recommendations based on complete high quality information. EFRs can be exported in multiple formats, accessed in the cab and in the office, are shareable, and can even be streamed via API to a farmer’s preferred precision ag software, like EFC Systems FieldAlytics.
The Future of Agriculture
The problems we’ve spent the past five years addressing at Farmobile haven’t been easy.
Building comprehensive universal solutions that address these entrenched problems was downright hard. It took longer than we thought but with only four seasons under our belt we’ve solved the most massive engineering problems by working at startup speed. When I look back on the past five years, I couldn’t be prouder of my team and what they’ve accomplished. We have out-executed every other AgTech data platform in the space regardless of funding or parent company resources, maintained our neutrality and this was all accomplished using good old-fashioned Midwestern work ethics.
Technologies that the Bay Area is so well known for, like Machine Learning and Artificial Intelligence, need reliable and scalable foundational data sources to make a meaningful difference in ag. All of these technologies rely upon data, and that data needs to be the highest quality, most comprehensive and readily available for others to build upon and accelerate the learning in the world of producing our food.
Farmobile is the only place to legally license these proprietary data sets to train big data algorithms with explicit farmer permission. Our only business is helping customers collect the highest quality data set possible with very little human interaction and making these data sets available in their entirety for licensing. We are better at it than anyone else in the world because it is our only business — no hidden agendas.
Of course, we didn’t get into this business just to fix big technology problems. We did it to support farmers and create a system to drive more accountability in products and services that are being sold to farmers.
In 1980, farmers received about 31 cents for every dollar spent on food by the consumer, and ironically today that number has been cut in half to about 16 cents received by the farmer. In the new world of Big Ag consolidation, the farmers are being sold bundles of products (fertility, seed, spaying applications where traditionally these were each independent decisions made by the farmer) to get the best combination pricing.
To enable the farmers to make the best decisions for their operations, measure field by field level performance statistics, generated from high quality ground truthed data, combined with external data sources like weather and imagery, will be the source of truth for product bundle performance and this can only come from a neutral platform.
Watch and hear more stories from the Farmobile team’s Milestone Moments in Innovation.
Farmobile co-founder and CEO. Passionate about advocating for the farmer. #FarmerPower.