USDA Precision Technology Adoption Report Highlights and Analysis
Highlights and analysis from the recent USDA report
Recently, the USDA released a report called Precision Agriculture in the Digital Era: Recent Adoption on U.S. Farms. It is really well done and while there are other technologies out there, it does a great job highlighting adoption trends within the precision ag space.
This week I aggregated some thoughts on adoption considerations along with then highlighted and commented on some of the most compelling data points and insights from the report.
Aggregated Thoughts on Technology Adoption
(Note: I use the term technology a lot. I am always referring specifically to the type of technology referred to in the report that is related to agronomic and operational aspects of the farm).
The Innovation Adoption Curve is heavily cited and frequently discussed in agriculture (and all industries).
Often I hear people say the adoption curve is different for agriculture (when talking acres farmed though I’d suggest the argument is valid).
What many do not realize is that the innovation adoption curve was based off hybrid corn adoption in Iowa.
Acceptance and Diffusion of Hybrid Corn Seed in Two Iowa Communities was the basis for the curve from work done in the 1940’s!
That means that even agriculture is not immune to the normal distribution of innovative mindsets (or lack thereof).
The argument in my opinion is misplaced if we are focusing on the distribution variation.
Ultimately what is important is the speed to move through the curve and across the chasm. This is what delivers returns to investors and success for companies.
The majority of digital technology has moved at tortoise like speed in agriculture. However, biotechnology has moved at lightning speed in adoption terms.
Below is a chart from the USDA on Herbicide Tolerant trait acre growth by year and crop. Look specifically at soybeans in the light blue/teal:
From launch to 80% adoption in ~7 years on soybeans.
In comparison, here are timelines for consumer technology penetration rates:
No influential technology listed hit 80% in under 10 years. In fact, the fastest technology after soybean HT technology is smart phone usage and we all know how influential that has been around the world.
Herbicide tolerant soybeans is one of the fastest adopted technologies ever.
Other agricultural technologies that are digital in nature? They are on the other end of spectrum.
This is a chart from The Kansas Farm Management Association and was released a few years ago:
On average, the technologies listed progressed to key levels at the following speeds:
2.5% adoption from date available = 2.6 years on average
16 percent adoption from date available = 12 years on average
From 2.5% to 16% = 9.8 years was the average.
Only one of these technologies is past 50% even as of the 2019 data in the new report (still automated guidance)
There is not a single digital technology that has reached 80% adoption according to the USDA report.
Why is that? The answer is of course nuanced, but one quote stands out from the source of the consumer adoption chart above regarding why technology is adopted more rapidly today:
It seems partly because modern tech needs less infrastructure in contrast with the water pipes, cable lines, electricity grids, and telephone wires that had to be installed throughout the 20th century. It also says something else about today’s consumers – which is that they are connected, fast-acting, and not afraid to adopt the new technologies that can quickly impact their lives for the better.
Enabling infrastructure is key. In Carlota Perez book Technological Revolutions and Financial Capital she talks about the deployment of capital and where that capital gets allocated to:
The instalment phase of technology where capital is being invested in these important pieces of infrastructure has huge bearing on uptake. When the foundations get laid to enable the wide adoption technology, things can pick up. We are just seeing the enabling infrastructure being laid, think of connectivity, data standardization, cloud capabilities, machinery upgrades and more.
Infrastructure is only one hurdle that has to be overcome though. And this combination of hurdles are why as you’ll see in my highlights of the report (or reading the report itself), digital technology is still not well adopted on the whole.
The farmer ultimately is the one that experiences the brunt of the pain from implementing technology and if companies haven’t done a good job of ensuring their product works smoothly with other technologies or systems or even with the institutional support infrastructure a farmer relies on (eg: agronomist, input retailer, equipment dealership etc), then challenges arise and adoption stagnates. The ease of use and simplicity has to be addressed with all technology.
I look at adoption risks as a hierarchy of hurdles that need to be overcome before successful adoption at scale. They can each be worked on simultaneously, but it’s difficult to see deployment at scale unless each hurdle has been meaningfully addressed. To compound the challenge, the standard to overcome these hurdles becomes higher as you progress through the innovation adoption curve from innovator, to early adopter to early majority:
Adoption of technology in absolute terms is about addressing all these different scenario’s clearly.
Infrastructure was noted earlier. Education is about how clearly you can help a farmer understand the value and how to deploy the capability, resources are how quickly and confidently a farmer will allocate time and labour to the technology, economics are whether there is a true return on the investment being made. And the entire process needs to add as little extra effort as possible (aka be simple). No small feat.
If we consider the adoption of herbicide tolerant soybeans, we can easily see how it overcame all these hurdles rapidly:
Farmer already had the planter and sprayer capabilities on their farm or readily available to them.
Monsanto invested heavily into educating the ag retail system and having representatives available to support the product.
Farmers needed to use LESS resources because they could eliminate complex tank mixing or extra passes.
Weed control was a challenge in the naturally uncompetitive soybean crop. The tolerance naturally controlled weeds better and the genetics tied to the trait were still strong so the return was immediate (that year vs. drawn out longer for lots of digital technologies).
Part of farmer psychology is being looked at favourably in the local area. Having better weed control immediately boosts this preference because fields look cleaner. It also was only a one year commitment and no change in buying process alleviating commitment issues that we see rampant in lots of digital technology.
What isn’t talked about enough are the psychological dynamics of adoption. Fear of failing, uncertainty of the perception at the coffee shop, lack of clarity on whether they’ll be able to do the core job of get seed in the ground or spray on a field if there is a hiccup with the technology all contribute to the psychological challenges (plus more).
Psychology will always be the final hurdle, specifically to cross the chasm.
What reports like this released from the USDA illustrate though is that there is more to gain from technology than there is to lose.
Part of the psychology to overcome is the concept of loss aversion in humans. Loss aversion is a cognitive bias that describes why, for individuals, the pain of losing is psychologically twice as powerful as the pleasure of gaining. The loss felt from money, or any other valuable object, can feel worse than gaining that same thing.
What the data in this report illustrates though is that there is more to lose from not adopting technology than there is to gain. Of course, that comment is nuanced (eg: which technology do I adopt first?), but the take away is that there is a need to start and a need to assess the farm for the areas of opportunity whether operational, agronomic, grain marketing or otherwise.
The report clearly illustrates that a lack of technology adoption tends to be correlated with smaller farm sizes, lower crop yields, less use of crop management recommendations, and limited employment of technical or consulting services. There isn’t explicit profitability called out, but one can infer from the rest of the data profitability lags as well.
What we have seen in the greater landscape of business is that companies that adopt and utilize technology have set themselves up for greater success and the same is true in farming and agribusiness.
The one unfortunate extrapolation from the report in my mind is that the surest way to see technology adoption increase is actually novel financing mechanisms to increase the rate of consolidation. The variation between large and small farms is drastic. It’s obvious that there is a different mentality, need and return for farmers in the top quintile of farm size.
As farms consolidate, incentives change, the poorer companies are competed out (or we reset, like Carlota Perez’s theory talks about) we should see adoption increase.
USDA Report Highlights and Commentary
Precision Agriculture in the Digital Era: Recent Adoption on U.S. Farms - USDA
Here are some of my favourite take aways, charts and quotes from the report:
Auto-steer guidance systems were used on only 5.3 percent of planted corn acres in 2001, growing to 58 percent in 2016. Estimates for 2019 suggested 72.9-percent and 64.5-percent adoption rates for sorghum and cotton (planted acreage). In the same year, GPS applications were used on 40 percent of all U.S. farm and ranchland acreage for on-farm production.
It’s actually incredibly surprising to me that this number is still not well over 80%! Auto steer is one of those functionalities that has a high return on financial investment and human wear and tear.
Adoption rates vary by farm size: At least half of relatively large row crop farms (those at or above the third quintile of acreage, i.e., with at least 60 percent of fields on farms with lower acreage) rely on yield maps, soil maps, VRT, and/or guidance systems. Meanwhile (except for cotton), less than 25 percent of smaller farms (those with acreage in the first quintile) use any of these four technologies.
This isn’t surprising. Bigger farms not only tend to have scale benefits, but they also are more generally individuals that approach farming with a business first mindset that have consolidated acres.
We can even see how size literally means a trend towards more tech utilization:
One of the areas of agriculture that I am incredibly passionate about is soil. The deeper a farmer or agronomist understands soil, the better decisions they can make regarding numerous aspects of the farm - seed variety, seeding rate, fertilizer use, seed treatment usage and more. Over a 20 year period the adoption of soil mapping stayed virtually flat.
The thing about soil mapping is that it is generally expensive to invest in up front with the likes of EC mapping, or grid mapping or other variations of soil testing and mapping. But it is also one of the highest leverage pieces of information a farmer can have about their farm and is the foundation to other valuable farm practices, like VRT, or informing profit driving decisions like fertilization:
Variable rate technology (VRT) is getting adopted, but in only two areas; corn seed VR and corn fertilization VR is it executed on over a quarter of the acres:
The growth in adoption is higher than many other technologies though:
One of the most interesting stats was just how big of a time benefit technology can be:
Using a sample of corn fields from the 2010 ARMS, Schimmelpfennig (2016) found that total labor hours per bushel of corn for adopters of yield and georeferenced soil maps were 35 percent lower relative to those of nonadopters. Although the difference in labor hours per bushel between adopters and nonadopters of guidance systems was insignificant, labor hours per bushel were 28 percent lower for VRT adopters, as compared to nonadopters. These findings are corroborated in a sample of ARMS cornfields from 2010 and 2016 across 10 Midwestern States, as analyzed by McFadden et al. (2022c). In that sample, labor hours per acre (totaled across the farm’s entire corn enterprise) were 0.15 for adopters of yield and soil maps—half as large as the labor input for nonadopters of 0.30 hours/per acre. Moreover, adopters of guidance systems worked 49-percent fewer hours per acre than nonadopters; similarly, VRT adopters worked 41 percent less per acre than VRT nonadopters. Reductions in per bushel or per acre labor hours are unlikely to be mainly attributable to VRT adoption, however. Larger operations adopt VRT at greater rates than smaller farms, but these operations also tend to adopt other labor-saving practices (e.g., conservation tillage) and labor-saving equipment that directly embeds DA technologies (e.g., modern, everwider implements).
There are a lot of other dynamics occurring with a farm that adopts technology that are enabled by the mentality and culture, but it is fascinating to see how variable it is when it comes to time saved.
Then there is the other big factor: yield.
Farmers that use various technologies increase their yields drastically compared to other farmers. Again, there is likely more nuance to the why here (eg: they also likely have larger fertilizer bills), but it’s interesting to note:
And as technology increases, we can again see the utilization of crop consultants increase. I have long been bullish on independent crop consultants, along with the strong agronomists at retails or equipment dealerships, and this data reinforces the opportunity for these individuals moving into the future:
Overall, there is great insight in the report. It is 52 pages and I just scratched the surface of some of the interesting take aways.
I think it's important to note that the original study of hybrid-corn adoption in Iowa was by Everett Rogers who wrote the book Diffusion of Innovations. And Rogers is quoted as saying "there is no chasm in the adoption curve." (see link below for reference)
Also, a recent research study shows there is massive confusion and misunderstanding of the innovation-adoption model. The biggest mistake is people think the adoption curve applies to technologies, when in fact it only applies to "applications" of technology:
Great article Shane! I'd only make one note as food for thought that may explain part of the different adoption rates of the listed technologies.
In my opinion, it's difficult to compare those that include fisicall products and those that imply a change in processes. A new ht hybrid or even a guidance mechanism are kind of binary / "plugg and play" once you buy the technology you start using it and thats it .
When it comes to VRA applications for example its a huge change in procesess challenging the "we've always done it this way" and requires a combination of factors, includding, technology, knowledge, training of operators, machine softwares compatibility , etc, etc.
Its true that once a farmer gets into that world it changes they way they farm and move into a possitive spiral of more tech better results, more tech, but also barriers of entry are higher and results are seen more mid/long term thats why addoption rates crawl rater than run. Cheers