If we are to believe the literature, all innovations have an S–Curve that models and describes their market acceptance. The S-Curve is a generalized model that plots product acceptance over time. This S-Curve also integrates to a normal curve.
The normal curve that integrates the S-Curve models the average and standard deviation of the delays in the acceptance of the innovation. This normal curve with its average and standard deviation of the timing of the adoption of the innovation can be used to characterize the separations made between the innovation acceptance categories.
The standard deviations generally identify the location of chasms that separate these acceptance, called adopter, categories, and the regions that are separated illuminate and identify the characteristics of that adopter category. Therefore, the categories define the adopter segments of the marketplace, and the chasms identify the solution needs required to overcome the acceptance of the innovation by users in the next adopter category.
Digital Farming has achieved the role as an innovation with many early stage product offerings. The users in this early stage category are called innovators; and the offerings in this category are accepted by these few, more technically, astute users. The offerings in this category must transition in the market to the second major category of users, the early adopters. Generally, there are many offerings at this innovator stage. Some of these offerings will become viable in the early adopter category; the rest will die.
MORE BY MICHAEL R. COLLINS
In Digital Farming, there are incremental differences in the characteristics of the offerings from category to category. These differences define the chasm between the categories. The needs of the early adopters will differentiate the viable offerings of the innovation from those that are just an initial splash in the market. The viable solutions will cross over to the early adopter category; and, then when the characteristics of the next category are incorporated in the innovation, the viable solutions of that category will cross over to the next adopter category called the early majority.
There are five characteristics of the viable offerings that will cross over the first chasm to be early adopters. These characteristics define the technical deployability needs that ensure the early adopters in the market will incorporate digital farming in their farm management practices. It will also allow adopters to embrace and integrate it in their best practices. These characteristics will be represented principally in the digital platform and its architecture. Fundamentally, they will surface these characteristics as best practices for all stakeholders. The five characteristics are: economic, scalability, agility, latency, and holistic grower usability.
Economic: Leave no doubt Digital Farming is about maximizing profit, and the most important characteristic is the ability of the innovation to monetize its acceptance for the developer/ deployer. If it does not generate an economic profit, it will eventually die. The architecture of the digital platform will dominate the determination of its economic viability.
Digital Farming should demonstrate an ability to illuminate the optimum yield and determine the best profitability for the grower. Ultimately, viability is about cost; and, unless these costs are visible and accumulated continuously, they are immediately forgotten as the glamour of the new capability is demonstrated. Transparency must reveal the entire cost as an audit trail. The economics of the solution must be embedded, visible, and transparent.
The digital platform must monetize the results for all of the stakeholders, but especially to those that develop, deploy, and maintain the capability. It must pay for itself; and deliver an accounting profit while it also provides solutions for security, administration, data management, and access control to millions of stakeholders. The scale is enormous.
The digital platform must be economic to the grower. It must provide the grower the same security, administration, and access as the deployment stakeholders, and the grower must be able to use the tool throughout the cultivation to generate decision quality information. The digital platform must be as automated as possible to eliminate human intervention, bias/ uncertainty, and delay, and it must provide consistent proven results. This must not adversely add costs that detract from the margin the grower seeks from that harvest.
Each analytical process accessed and executed in the digital twin during the cultivation must monetize and then integrate these results to a farm management business system so that an informed cost decision is developed in the context of the cultivation. This integration, or incorporation, informs the grower of the intrinsic costs and inherent risks. During the cultivation each productive asset of the farm is continuously assessed. It is critical that the cost, revenue, and historical impacts of every process, or data need, are visible and directly traceable to the Profit and Loss of the farm.
Scalability: Scalability must be an intrinsic characteristic of the architecture of the digital platform. It must scale to thousands of concurrent users with multiples of cultivations running many instances of analysis across a wide geographic deployment. This digital platform must enable a plethora of synchronous, asynchronous, and asymmetric instances to run seamlessly and concurrently.
The most obvious use of scale is the ability to digest, process, and manage the vast amount of information. For example, we know that data from imagery, even from a five-band, multi-spectral sensor, will range into the multiples of gigabytes. This data needs to be geo- and ortho-rectified and then aligned for spatial, spectral, radiometric, and temporal resolutions. This needs to be automated, embedded in the solution, and immediate.
Depending upon the threats presented during the cultivation, data will be compounded, over the duration of the cultivation, by multiples of instances of that similar imagery; and not just for a single year but for multiple years, the temporal resolution. Every step of this process must be inside the decision cycles of that cultivation. How the system scales to perform the analytics, manage the data, and integrate it for alarming and decision making and how it does all of this automatically, and economically, are a critical part of the scale characteristic. Scale is often a cost driver, the antithesis of economy, but scale in the automation, management, and processing of data is critical.
Agility: Agility is a new term often used only in software development. The term agile in agriculture needs detailed definition; but, as is, it highlights an important analogue in this market. Each cultivation is different. From year to year a new crop or new variety is cultivated, and with each year a new insight to that instance may be developed. While the field shape is the same, the crop/variety is different, and there will always be new conditions that carry over from the previous cultivation. The digital twin must be continuously updated to new cultivation processes and analytical routines.
Each cultivation is often more than a different input. One of the major benefits of a viable digital farming solution is that it learns, and new approaches and algorithms can be implemented to take advantage of this learning. The ability to quickly adapt and innovate the solution must be intrinsic in the technology selected for implementing the solution. A monolithic structure will be difficult, and is ultimately impossible, to maintain, enhance, and improve. Agility in concept must be embedded in the technology of the solution. This enables scale and economy by applying the lessons learned at the user interface, crop/variety, field, and farming condition.
The deployed Digital Farming solution must continue to mature in its user interface to support “out-of-the-box” application, with minimal configuration, and the continued development required to support the early and late majorities. All aspects of the system need to be simple and easy to understand and use. The equipment in the digital farming platform and in the rest of the digital ecosystem, in the hardware and communications, must be interoperable.
Latency: The fourth characteristic, “Latency in the Digital Farming decision cycle,” is frequently overlooked. The viable solution must deliver alarms and decision information inside the decision loops of the cultivation. This means that system latency is critical. In general, the longer it takes for the system to return a result and a decision option, the less relevant the action to be taken.
The entire global solution will reside in the cloud, but in the local agricultural instance, the field and that cultivation, the cloud may not be immediately, or consistently, available. When it is not available and the decision loop demands an immediate answer, it is possible to execute a “cloud” at the edge with near zero latency processing and storage.
Edge computing provides a local private cloud where latency is minimized by the local communications and the virtualization of processing and storage is done in the local hardware network. It demands a data architecture that incorporates the latency of the data and its inputs. All decision data does not require cloud-based storage and processing for immediate alarming and decision making. This can be done locally at the edge and later integrated to the global cloud.
Agriculture, like no other market, must deal with the concept of the edge. This single issue is true for anything in the Industrial Internet of Things, but it is compounded by the farm ecosystem that is remote, communications and power constrained, and processing and storage limited. While much will be done in the cloud, in constrained environments, this system must be designed to work locally, i.e., for both the edge and the cloud. This characteristic demands the solution define processes at the edge and in the cloud, and provide the design to operate in both. Further, as the digital platform improves this solution, it must be forward, current, and backward interoperable and compatible in technology.
Most cultivation issues like weather, stress/crop threats, or planned activities can be predicted and scheduled. Decision information must be generated to support these schedules. Latency can have significant impact if the notice of an action is delayed and a crop is impaired or destroyed due to stress caused by hydration, weed, insect, fungus, etc. These stresses can be predicted and the required sensing and timing can be surgically determined and processes executed when needed. Once executed, this step cannot wait days or weeks for a result. The latency of the step needs to be minimized. It can be economic. It can scale, and it can be agile.
Holistic grower usability: The final characteristic of a viable solution is holistic grower usability. Every Digital Farming solution is a digital twin of the entire cultivation processes, even the farm management processes, and it requires user interface. The solution processes must be as simple and complete as possible, while returning explicit decision quality information.
This solution is also a part of a larger system that includes other growers with the same crops, in the same region, and with similar concerns. It is a part of a broader digital ecosystem, and can integrate regional inputs on emergent or detected stresses or environmental conditions.
As a part of this larger system, this solution should enable crop alerts, scouting, and other media. As a minimum, it must be complete in addressing all of the needs of the cultivation and translating this to the impact on the farm business. Beyond the local systems, the digital farming system must be able to access futures markets, upstream data from seed production files/genetics/small plot trials/ greenhouse data, and downstream data like harvest information for equipment/storage costs, capacities, availabilities, and warehousing/labor.
In summary, the viable Digital Farming solution, the solution that will cross over from the innovator to the early adopters, and then to the early majority, and enjoy the broadest acceptance and deployment, will be the one that supports these five characteristics: economic, scalability, agility, latency, and grower usability.