In many sectors the problem is that data is scattered across departments and systems, but still originates within one organisation. In the agricultural sector, the fragmentation starts before that point. A large part of the most requested sustainability data — land use, water consumption, crop rotation, animal feed, emissions from manure — does not originate at the reporting company itself, but at growers, suppliers and cooperatives that have no accounting or digital routine for it. Where an office-based organisation can retrieve data from an ERP system, an agricultural company often first has to make inquiries with a chain of third parties who do not see the question as their problem.
On top of that, the data that is available internally is usually recorded per business unit or per growing season, not per reporting period. A financial year does not align with a harvest cycle, and an emission factor that applies today may already have changed by next season due to a different crop or different tillage. Most of the time spent on sustainability reporting does not go into filling in figures, but into finding out who those figures should actually come from.
The first place is the supply chain itself. A large part of scope 3 emissions in the agricultural sector sits with suppliers who do not carry out sustainability reporting and whose data — if it exists — exists in a different format, a different language, or not digitised at all. Anyone who wants to link a data point to a source here discovers that the source is sometimes a person, not a system.
The second place is the operation on the company's own site: consumption of water, energy and crop protection products is often tracked on paper or in separate spreadsheets per plot or barn. This data exists, but has not been assigned to anyone as a responsibility for reporting purposes — it is farm bookkeeping, not CSRD bookkeeping, and the two rarely align seamlessly.
The third place is the translation into measurable units. Land use and soil carbon are recorded in very different ways by different actors, and converting this to a uniform reporting unit is itself already a source of errors if no one records which assumption was used in doing so.
In the ICT sector the sustainability data usually sits in digital systems that already exist, and it is mainly a matter of access. In the energy sector much data is already measured because measurement obligations and monitoring are part of the operations. The agricultural sector more often lacks that foundation: the data is not hidden in a system, it has never been systematically recorded. That does not make the problem bigger or smaller than in financial services or the real estate sector, but it is different in nature: there, the problem is usually integration between systems; here, it is often first registration at the source.
The temptation is great to buy a software package that automates reporting for the agricultural sector. But a tool that retrieves data from systems that do not contain the right data only produces a neatly formatted report based on the same loose ends. Before a system can add anything up, it must be established which data point belongs to which source, who inside or outside the own organisation is responsible for it, and which quality rule determines whether an entered figure is plausible. That is not a software question but a question of organisation, and it precedes any reporting system.
The Data Readiness Scan records for each data point where it comes from — from the company's own operation, from a grower, from a cooperative or from an external data source — and who within the chain can be held accountable for it. For a sector where much data originates outside the company's own walls, that ownership is often the missing piece: not the figure itself, but who stands behind it. To determine which data points are actually relevant for an agricultural company, and which are reported without serving a function, the distinction made in which data points you actually need when there are multiple business units is useful, as is the starting point that a data point often already exists at one of the existing business units before it is collected again.
Once it is clear which data points in the agricultural chain need to come from where and who supplies them, a second question arises: how much of the work of collecting, checking and tracing that data still requires human actions, and which part is repetitive enough to leave to AI. The work scan from FTE TO AI calculates this per task, not based on an estimate for the entire sector but based on the tasks that are actually carried out in a specific chain. For a sector where much time goes into tracing sources rather than interpreting figures, this is a starting point for seeing where automation delivers something and where it does not.
The scan that maps these data points, sources and ownership for the agricultural sector is under development. Anyone who wants to be notified when the scan becomes available can sign up for the waiting list.
Vraag maar waar een datapunt vandaan komt. Dat is meestal de hele vraag.
Answers come from this site’s knowledge base. Not tailored advice, and not a scan of your company.