Software venture, boosts farm yields and net income in successful field trials of its predictive analytic application
FORT COLLINS, CO, February 15, 2018 - Visual Farms, LLC, is a Fort Collins, Colorado based start-up whose goal is to optimize farm profitability and sustainability through improved crop yields and better decision making. Using artificial intelligence algorithms (AI), and Machine Learning systems, the company has successfully increased crop yields and net farm incomes in field trials designed to demonstrate the Visual Farms System of matching specific seed choices to each farm's unique location and field.
The farmers planted seeds they selected using their traditional methodology alongside hybrids chosen by the company's BestSeedChoiceTM predictive analytics software. The 2016 and 2017 trials, conducted in northern Illinois and east central Indiana, used 700,000 individual crop yield tests, using 13,368 unique corn seed varieties from 174 suppliers and more than 15 million same-field comparisons.
Visual Farms announced the use of the seed recommended by the BestSeedChoiceTM software increased overall crop yields by an average of 15%. Additionally, whole farm yield forecasts from the software were 100% accurate and field level yield forecasts were within 98% accuracy. The system accurately modeled and projected in-season planting, harvest, growth stages and conditions timing.
"We were able to document that when farmers use our BestSeedChoiceTM tool, eight out of nine recommended hybrids out-performed other hybrids planted and net farm income also increased," said Ron Olson, Visual Farms co-founder. "Over the course of the trial, farmers in northern Illinois saw a net farm income increase of 22% ($47 per acre). In east central Indiana, it was 27% ($87 per acre)."
More Revenue Per Acre
Visual Farms has demonstrated that its proprietary AI solution substantially boosts a farm's grain yields by optimizing seed selection. Additionally, the system enables fine tuning field management during the growing season with data-driven advice tailored to each individual field.
Preliminary field tests conducted in 2016 were so encouraging that in July 2017, the company published its advance projections for the October yields for test sites in three states across the US Corn Belt. In the 2016 trial in Indiana, its advance projection, in bushels per acre, was within 2% of the actual yield. The new field trials significantly exceeded the company's expectations which, prior to the trials, were to increase yields by up to 10%, improve net farm income by up to 20%, and to predict crop yields within 95% accuracy.
To use the tool, farmers first select the fields for which they want to optimize crop yields and provide Visual Farms with the specific criteria important to them such as relative yield, moisture, maturity, risk class, performance variability/stability, genetics and other criteria.
BestSeedChoiceTM provides ranked seed recommendations from a minimum of 100,000 head-to-head comparisons within the farm's local agro-ecosystem in a sortable list or an interactive graphic. Farmers interactively choose hybrids to plant from responsive lists and graphics. The farm can apply searches for their favorite hybrids to compare and then filter selections by its own minimum criteria, before choosing from the recommended hybrids.
Unprecedented Data Resource
The company has compiled 10 years' yield test data from across the US from public sources. Ninety percent of that data is no longer available. A given seed variety is tested side-by-side with other varieties, for a given location and equivalent conditions. Visual Farms has amassed data for the last 10 years on approximately 40,000 different corn varieties, 25,000 different soybean varieties and 12,000 varieties of wheat in field trials from pooled results of universities and seed companies. Visual Farms has the data to make about 20 million comparisons among corn seeds, for example - an order of magnitude more data than its closest competitor would have.
"The BestSeedChoiceTM software puts an unprecedented compilation of data at the farmer's fingertips, allowing him to identify the best-performing hybrids of major crops like corn, soybeans and wheat for his region and his farm," Olson said. "Our database is generated from over 700,000 yield tests with over 13,000 unique seed varieties. In Indiana alone, in an area covering a dozen counties, that dataset includes almost 100,000 head-to-head yield trials comparisons. It enabled our trial partner, Fairholme Farms, to narrow its seed choices down to the best-performing hybrids for its specific location and conditions."
"Farmers need to have an edge every year," Olson said. "Our tool allows them to save time and make better seed selections, to improve their yields, reduce costs and increase profits. We aim to keep pushing that edge continually."
Farmers will pay a one-time, annual subscription fee for the use of the software service.
Seeking First-Round Equity Funding
The founders of Visual Farms include technologists formerly from John Deere in field spatial systems and artificial intelligence, and a senior agronomist formerly from Cargill. The company currently is seeking to raise private equity placement to fund commercialization of the BestSeedChoiceTM software.
Visual Farms estimates its total domestic addressable market at 1.7 million farms producing corn and/or soybeans on a total of 180 million acres, or one-half the nation's cropland area. One-third of these farms grow both crops. The company projects that global corn and soybean markets, which its systems also can address, could be 4.4 times larger.
Domestically, the company's primary target is 324,000 farms growing corn and soybeans on fields larger than 100 acres on a total of 132 million acres. Within this group are the 4,400 largest farms (1,000 to 19,000 acres). About 30% of which employ precision farming technologies, on fields averaging 4,000 acres per farm. Those operations, Visual Farms executives believe, represent the "sweet spot" for its planned SaaS products.
Because they recognize that every farm is different, the company believes farms will be best served when every farm has its own independent Machine Learning system.