Categories
Blog

Digital data in the agri-food supply chain

Established companies and technology start-ups are all racing to create solutions that better manage agricultural data in supply chains. Does this mean radical new transparency for farmers and producers? And have we thought through the implications for data collection, management, and ownership?

In this post:

What is agricultural data?

Definition of agricultural data;

The facts, metrics, and statistics that describe elements of one or more farming or agriculture operations.

Most farms collect data in some form. Much of this may be in personal notebooks and mandatory compliance forms. Precision farming equipment, machinery, and mobile and desktop apps also collect agricultural data.

Most farmers collect information to:

  • Support improvements to farm management;
  • Follow government directives; or
  • Have something interesting to talk about in the pub.

So why are processors, food service and retailers, and dozens of internet start-ups becoming more interested in on-farm data?

How rich digital data can benefit agri-food supply chains

Three ways that digital data can benefit consumers, retailers and processors in the agri-food supply chain are:

  1. Traceability and tracebacks;
  2. Forecasting and efficiency; and
  3. Supporting product claims.

1.      Traceability and tracebacks

Tracking animals and crops through the supply chain helps the entire chain to respond to concerns about food safety or disease. This is especially important in livestock industries where animals move between farms, often through shared facilities.

Even for crops, on-farm records can establish linkages between the fertilisers and slurries, pesticides and herbicides used, and the resulting product.

2.      Forecasting and efficiency

Purchasing and processing goods from biological systems carries uncertainty and risk. Crop yields, dry-matter, or flavour will vary from the sector “average”. Animals may not be ready when first predicted or vary in how they meet processing specs.

If on-farm data were available before harvest or delivery, processors and retailers could predict the likely quality, timing, and specification of supply.

With enough lead time, processors and marketers could better match demand and processing capacity to supply. A dairy processor might vary the mix of UHT, cheese, and powder products based on expected quantities, fat, protein, and calcium levels. A fruit marketer could negotiate different market commitments based on predicted ripeness and flavour profiles.

Connected data may allow market signals to flow the other direction also. With the right information, producers could adjust harvest dates or livestock delivery to achieve target specifications and match market demand.

3.      Supporting product claims

Consumer interest is driving the creation of differentiated products, which make claims about what they do or do not contain. Examples might include:

  • “free from x”,
  • “organic”,
  • “naturally produced”,
  • “grass-fed”,
  • “local”,
  • “A2 beta-casein only”, or
  • “higher welfare”.

Consumers can see differentiation like “chocolate flavour” or gold kiwi fruit. “Credence attributes” are types of differentiation that can’t be seen. Consumers can only evaluate these based on trust and the story that supports the claims.

Small-scale producers can single-source from one or two farms that they own and closely control. For supply at scale, the evidence and controls to support credence attribute claims must be based on data and audits. And even audits make substantial use of agricultural data collected on farm.

Challenges of data in agri-food supply chains

Making effective use of agricultural data to benefit the supply chain is a worthy goal. In our experience, it is not necessarily straightforward. If you intend to use on-farm data to support an agri-food supply chain, there are four key challenges to consider:

  1. Data collection effort and methods;
  2. Data quality and completeness;
  3. Data flow between organisations; and
  4. Data ownership or control.

1.      Data collection effort and methods

With some exceptions, farmers have not traditionally been proponents of formal data collection. A few agribusinesses have built a culture of data gathering and analysis, but many farms would collect the minimum possible.

Recording has often been informal. Data to support a decision might appear on paper, in a notebook, or on an embedded device. After the on-farm decision, data may be discarded, having never been transcribed or centrally stored.

Apps are a great improvement over desktop software for data collection. But, collecting agricultural data is not as simple as rolling out a new app. Design effort needs to go into establishing when, how, and why data will be collected. You need to consider appropriate incentives and support.

A powerful data collection incentive is to immediately return useful insights to support on-farm decisions. For instance, a tool tracking mobs of animals for a processor might graphically show small changes the producer might make to improve their returns.

Remote sensing, image processing, and Internet of Things (IOT) devices promise to take farmer effort out of data collection. In our opinion, this could be transformative. At present the cost of some devices (compared to their perceived benefits) is still a challenge, as is network connectivity. Rollout of 5G networks may improve this!

2.      Data quality and completeness

Data quality issues in agricultural data don’t always arise from insufficient validation of input boxes. Sometimes just the opposite! Issues include:

  • Software and tools that are too clumsy to use or take too long, so don’t get used.
  • Overly tight validation that forces farmers to lie or “fudge” data to get it accepted.
  • Farmers who record results they believe that they should be getting, rather than what is really occurring. A farmer once told me about lamb growth rates that matched industry best benchmarks: I only to discovered later that they did not own any weigh scales.
  • Farmers recording data “just to tick the boxes”, so records are abbreviated, approximated, or (potentially) fabricated.

Transcription errors are another common cause of problems with data quality. We can understand this where data is captured on paper and later transcribed (and certainly in-field data collection can reduce errors). We have also seen real cases of manual transcription between software systems – with an advisor placing their laptop by the farmer’s computer so they can manually re-enter data from one screen to the other.

For supply chain data to be timely and useful to all parties, careful attention needs to be paid to the underlying design issues that cause missing and inaccurate data.

3.      Data flow between organisations

Supply chain networks face potential challenges in managing the flow of data between organisations. For example, farmers may potentially make use of several similar-but-different tools that capture data on farm. Or supply chain partners may request that a grower or farmer use their preferred tool – which can be challenging if the grower sends produce to multiple markets with different preferences!

In an ideal world, producers would not be locked into a single software tool or equipment manufacturer. Use of global standards would allow farmers, growers, processors and retailers to “mix and match”, selecting the best tool for their circumstances with confidence of compatibility. 

Such e-commerce standards have existed between large supply chain partners for many years. Consider electronic ordering, ship notifications and invoices exchanged in the automobile supply chain, for instance. Equivalent progress in the agricultural market has been slow and fragmented, although initiatives such as ICAR, DataLinker, and AgGateway are changing this.

4.      Data ownership or control

As supply chains start to leverage agricultural data, a key question that needs to be asked is “who owns or controls this data?”. Is it the producer, the manufacturer of on-farm equipment, a software vendor, or the processor or market partner who receives data?

It may be tempting to take the approach of “possession is nine tens of the law”. If the data has made it into our database, surely it is ours to use?

With some exceptions, rights to control data fall under copyright law. This leaves the “ownership” decisions about who can use data, and for what purpose to the party who invested time or money to create it – unless changed by a contract.

Surveys show that farmers worry about who controls and uses their data. Surveys of US farmers from 2014 and 2016 showed that 77% of farmers were concerned or very concerned about which entities could access their data, and whether it could be used for regulatory purposes. The November 2018 Farm Credit Canada survey showed similar results.

These concerns motivated the US Farm Bureau to draft its Privacy and Security Principles for Farm Data, and the NZ pastoral farming industry to create the NZ Farm Data Code. The position of these codes has been that organisations and farmers should explicitly agree what data is shared, and for what purposes, and that the starting point should support farmers rights to data about their businesses.

When we work with supply chain and agritech companies, we recommend that organisations are definite about the uses to which they will put data, and that they communicate this clearly and trustfully with producers.

In summary

There are compelling reasons why supply chain organisations in procurement, processing, marketing and retail, are looking to make greater use of agricultural data. Effective use offers greater forecasting accuracy and supply chain efficiency, as well as supporting differentiated product claims. If this is your vision, you’ll also want to consider how you will tackle the challenges of agricultural data – collection, quality, connectivity between organisations, and rights to data.

Rezare Systems helps organisations collect and make sense of supply chain data. We focus on your intended outcomes, rather than a single technology. We use design-led processes to collaboratively look across the issues of collection, quality, connectivity and rights – to identify what must be tackled, and when. If this resonates with you, let’s discuss.