Time for Agriculture to Think Bigger on Data
Agriculture generates more data just at a field level than almost any other industry does across its entire supply chain. As such, you’d think the ag sector would be leading the pack on big data.
As we know, the reality is very different.
Despite large scale investments in agtech, and the subsequent creation of a lot of great products and services, I think most would agree we’re still struggling to find industry-wide success when it comes to data.
Integration Friction
So, what is holding us back?
MORE BY MATT WAITS
The nature of this industry requires collaboration between input manufacturers, ag retailers, growers, grain merchants, and others. Given that no single company can provide the digital tools for the entire supply chain, integration between systems is key to creating successful outcomes. This is where we continue to fall down as an industry.
Due to a lack of agricultural data standards, integration between market participants remains costly and is prone to error. There is as yet no single source of truth for a grower’s data. It is spread between many systems, some controlled by the grower and some by their trusted advisors. Some are geospatial in nature, others aren’t, and all have different underlying data models.
Moving data between systems is difficult enough, but maintaining synchronicity between them is nearly impossible. Other than in isolated pockets of cooperation, scalable insights that could create supply chain efficiencies and improve grower outcomes are yet to be realized.
Mitigating the Human Factor
The issue of non-standard data is amplified when growers and agronomists are able to enter data manually. What makes sense to the person doing the data entry, can be very different for someone else having to interpret it! For example, the individual that typed ‘RU’ into his software knows which version of Roundup he used, but given that there are around 10 different versions of Roundup, the rest of us will have to guess.
Humans are sloppy, moreover we can be guilty of making assumptions about what data we do or don’t need to enter into the system. It is for this reason that all tech providers to the ag sector need to provide tools with clearly defined data models and we must train users on the value of data completeness.
Start with the Basics
Eventually the industry must settle on common standards that ensure that we all define growers, their farms, and the underlying field data in the same way, but that will be a long journey.
The first step towards a better functioning digital ecosystem is to create systems that provide users with consistent picklist-driven data collection. Today, as mentioned, too many systems allow users to create their own picklists or worse they allow them to type whatever they want into the software. This simple issue makes integration and analytics exceedingly difficult.
The very real promise of digital agriculture remains tantalizingly close but frustratingly distant. It will continue to be so without consensus on data collection methodologies. Scalable analytics can only go so far until we make life easier for those charged with interpreting the data.
The solution should be straightforward — to agree on standard picklists for items like crop, seeds, crop protection products, and fertilizers.
However, that’s easier said than done when there are simply so many products on the market, many of which are variations on a theme. It’s all very well having a standard picklist, but can we really expect users to scroll through long and unwieldy lists of similarly named herbicides, for example, from just one brand. It’s much easier just to pick the first one – and then we’re back to square one.
Until it’s quick and easy for those at the field level to complete the data entry accurately, we might as well go back to the drawing board.
Time to Step Up
Let’s not admit defeat quite so easily though. One answer lies in QR codes, there have already been successful trials conducted across markets. These have generally been in relation to recalls and reducing input fraud, but these prove the model can work.
Change has to start somewhere and it has a tendency to cascade out – we just need someone to take that first step. In this case, it’s the input manufacturers and agricultural software businesses that can lay the groundwork for standardization.
This is an important time for the ag sector, there are undoubtedly new challenges on the horizon and meeting these means closer collaboration, particularly in relation to data. Thinking bigger on agriculture means starting small though. Having common definitions for data might sound minor, but the implications for what it could lead to are so much more.
Subscribe Today For
