Diagnostic
Data and loss diagnostic
Sample mill, 2024/25 season
Delivered in May 2025
Agnaris team, Recife
Sample document. The numbers are illustrative.
Consulting and diagnostic
The work starts with an analysis of the data the operation already has. Each phase ends with a deliverable and with your decision on whether to move on to the next one.
-
Phase 14 to 6 weeks
Diagnostic
Analysis of one or more seasons of your history, under a confidentiality agreement.
What we do
- We map where the data is: systems, spreadsheets and paper records.
- We organize the history of crushing or production, stoppages, weather and quality.
- We measure how much time and raw material the operation lost, and why.
- We estimate the possible gain in reais, with the assumptions written out.
What you receive
- Report on the causes of stoppage and loss, in hours and in reais
- Gain estimate with open assumptions
- Data inventory: what exists, where it is and what is missing
- Presentation of the results to management and the technical team
What we need from you
Exports or read access to the data and a few hours a week from someone in the operation.
At the end of the phase
With the report in hand, you decide whether a field test is worth running.
-
Phase 26 to 12 weeks
Field test
The forecast runs on part of the operation, with the success criterion agreed beforehand, and is compared with what happened.
What we do
- Together with the team, we choose the areas or fronts that go into the test and agree on the success criterion.
- We run the forecast every day in those areas.
- We compare forecast with actual, week by week.
- We tune the model together with the people in the field.
What you receive
- Daily forecast for the chosen areas
- Weekly report of forecast against actual
- Record of the decisions made from the forecast and the result of each
- Test summary with the gain measured, not estimated
What we need from you
One person in the field to record what was done and give feedback on the recommendations.
At the end of the phase
With the measured gain, you decide whether to extend it to the whole operation.
-
Phase 3ongoing
Implementation
Integration with the existing systems and use across the whole operation.
What we do
- We connect the platform to the systems the operation already uses.
- We set up dashboards and alerts for each role.
- We train the teams that will use the recommendations.
- We track results every season and recalibrate the models.
What you receive
- Platform in use, integrated with weighbridge, laboratory and ERP
- Dashboards for management, agricultural and industrial areas
- Results report every season
- Support from the technical team
What we need from you
A person responsible for the platform on the operation's side.
At the end of the phase
The contract is reviewed every season based on the measured result.
What a diagnostic report looks like
A complete sample report, with illustrative numbers, like the one delivered at the end of the first phase. You can browse it here or save it as a PDF.
Summary
In the 2024/25 season, the mill lost 519 hours of crushing. Close to 40% of them, 218 h, could have been avoided with rain forecasting and planning of available cane. One in every five loads went over 36 hours between burning and crushing.
- 519 h
- of crushing lost
- 218 h
- avoidable
- R$ 2.2 to 3.3 million
- estimated gain per season
- 2 of 5
- complete data sources
We recommend testing the non-workable day forecast for 12 weeks in the next season, comparing forecast with actual every week.
Crushing hours lost in the season
519 h
hours lostavoidable with forecasting and planning
Rain and stoppages by month
Rain stoppages were concentrated at the start and end of the season, when available cane was lowest.
62 41
Sep 38 22
Oct 21 9
Nov 18 6
Dec 44 28
Jan 71 53
Feb 118 89
Mar Rain (mm)Hours stopped by rain
Time between burning and crushing
One in every five loads went over 36 hours between burning and crushing. Those loads arrived with ATR about 4% lower, which matches the loss of about 2% of the sugar per day of cut cane left waiting.
Loads in the season
Gain estimate
R$ 2.2 to 3.3 million per season
Assumptions
- Recovery of 40% to 60% of the 218 avoidable hours
- Crushing of 400 t/h and revenue of R$ 250 per tonne, from sugar and ethanol
- 25% of the revenue of a stopped hour treated as lost: fixed cost and the sugar in cane left waiting
Data inventory
| Data | Where it is | Status |
|---|---|---|
| Hourly crushing | Plant system | Complete |
| Stoppages and causes | Maintenance spreadsheet | Partial |
| Rain | 3 rain gauges, manual notes | Partial |
| Quality (ATR) | Laboratory system | Complete |
| Burning and cutting by front | Agricultural area spreadsheet | Not recorded |
Recommendations and next steps
- Record burning and cutting by frontBefore the season
Today they live in a spreadsheet. Without them, a load cannot be linked to its farm and the time to crushing cannot be measured.
- Automate rain readingsBefore the season
Two automatic stations cover the areas with the most stoppages.
- Test the non-workable day forecastPhase 2
12 weeks, comparing forecast with actual every week.
- Replace the fixed yard target with available canePhase 2
Raise available cane ahead of forecast rain, counting cut and loaded cane, not just the yard.
Each step has its calculations and assumptions in the appendix, for the mill team to check.
Agnaris team, Recife
Page 1 of 8
How we measure the field test
Every week, the forecast number of non-workable days is compared with what happened. The gain goes into the summary only when it is measured.
- Crushing hours preserved
- 41 h
Numbers from an illustrative test
How we work
- One decision per phase
No phase starts until the previous one has been delivered and approved.
- Assumptions in writing
Every estimate comes with the calculations and hypotheses used, so your team can check them.
- Your systems stay
We work with the systems and spreadsheets the operation already uses.
- The analysts are your contacts
You always talk to the technical team that does the work.
Talk to the team
Tell us what your operation produces and what data it already has. The first conversation is with the technical team.