Despite the promise of data, the Global and South African agricultural sectors face an industry-wide challenge of data fragmentation. At this industry scale, data value is subject to six dimensions of data quality: timeliness, accuracy, consistency, completeness, uniqueness, and granularity. Suppliers often use multiple recordkeeping platforms, each with different standards and metrics. With multiple recordkeeping systems in place for primary agriculture, data overload becomes a possibility, risking data quality at the industry level. This not only results in duplication and inconsistencies, but also raises the transaction costs of accessing the data for data-consumers — the effort, time, and cost to access and align data.
For example, one supplier may input the same data into several platforms using different units or timeframes. This can lead to errors, inconsistencies, increased time burdens, and undermine the data’s value and credibility.
These transaction costs of data capturing, such as platform fees, motivational costs (time and effort spent on capturing data across platforms), search costs (effort to collect and prepare data for multiple platforms), and coordination costs (ensuring consistent data across platforms), actively discourage recordkeeping.
Optimising the value of data
Data is widely referred to as the world’s most valuable resource. However, what gives data its value? The key is in the quality, rather than the quantity, of the data. Given the requirements of the Carbon Tax Act of 2019 and the Climate Change Act of 2024, it is important for agricultural stakeholders, like farms and packhouses, to understand how to extract the greatest value from their data via the effective use of digital recordkeeping programmes, such as SIZATrack360. It is equally important for industry-wide alignment in data-capturing efforts to enhance data quality at an industry level.
A common misconception is that more data equates to more value and better insights. In reality, data overload can result in duplication, inefficiencies, and higher costs. Typically, the “value” or “quality” of data is dependent on six dimensions: timeliness, accuracy, consistency, completeness, uniqueness, and granularity (see below). Optimising these characteristics can greatly enhance the value of one’s data.
- Timeliness: Refers to data being as up-to-date and relevant for analysis. The value of data declines as it becomes increasingly outdated, misrepresenting the present reality and thus resulting in incorrect conclusions and recommendations. Therefore, it is important to ensure that data is updated regularly, or in real time where possible.
- Accuracy: Refers to the extent to which data accurately reflects reality. Enhancing data’s accuracy involves correcting inaccurate entries, ensuring correct units, etc. Data accuracy may be improved by including data validation as a risk-mitigating step. SIZA can play an essential role in validating data of its members within the MySIZA data platform and with the SIZATrack360 programme through direct validation and stakeholder engagement.
- Consistency: Also called “objectivity”, refers to the extent to which data is consistent and coherent across systems and datasets. This may be improved by ensuring a data standardisation process to ensure that all data is interoperable across systems and datasets
- Completeness: Refers to whether the data contains all necessary records, where applicable, without missing values or data gaps. A more complete dataset enables more comprehensive analysis and provides more in-depth evidence for decision-makers. Dataset completeness can be improved by ensuring all information from multiple information sources is utilised, with missing values accurately imputed where possible. This reduces areas of data fragmentation.
- Uniqueness: Refers to the absence of duplication in a dataset, which may skew the data’s representation of reality. Uniqueness of a dataset may be enhanced by identifying and removing duplicated or redundant entries where possible. This may also refer to duplication of the entire dataset or a selection thereof.
- Granularity: Refers to the level of detail of the dataset to allow for specific and relevant analysis. Insufficient detail within a dataset may inhibit analytical insights at a desired level (i.e., block-level, cultivar-level, etc). However, one must be vigilant of excessive granularity resulting in unnecessary overcomplexity.
Data for evidence-based decision-making
High-quality data results in informed decisions. This is because high-quality data not only accurately depicts reality but also accurately informs decision-makers regarding the real-world impact of interventions. This is even more important in highly complex interrelated business systems, such as primary agriculture.
For example, high-quality data with accurate insights allows management to influence operational efficiencies. This is because high-quality data enables accurate analysis and diagnosis of resource utilisation, allowing for optimisation thereof via setting targets and continuous monitoring to improve resource-use efficiency.
Additionally, high-quality data may improve stakeholder relationships. By enhancing the transparency of relevant data to data consumers, the transaction costs of accessing and cleaning the data for the end-user are lowered, resulting in improved rapport throughout the value chain.
It is therefore important for the industry to prioritise a holistic and coordinated approach to recordkeeping and, where possible, enhance synergies, avoid duplication, and reduce transaction costs of data capturing in agriculture as a whole. This would incentivise data capturing via a digital recordkeeping tool, such as SIZATrack360. Thus, it is paramount that recordkeeping platforms in agriculture collaborate for the benefit of the industry.
Overall, a holistic, collaborative approach among digital recordkeeping programmes in South African agriculture may reduce the transaction costs of data capturing via monetary and time incentives and, therefore, greatly enhance the quality and value of the industry’s data. This is increasingly important given the requirements of the proposed Carbon Tax Act of 2019 and the Climate Change Act of 2024.
