The paper pond record, replaced by an app that works offline.
An aquaculture feed and farm-services group whose field technicians support shrimp and fish farmers across several Asian markets, collecting pond data on paper during farm visits.
What wasn't working.
Every pond cycle was recorded on a paper culture record form. A technician filled it in during a visit, and the data then travelled by hand to an office to be rekeyed, so by the time anyone could analyse a cycle it had usually finished. Farmers had no way to log their own feeding or spot a problem between visits, technicians had no consistent routine to work to, and head office could not compare performance across countries. Connectivity ruled out the obvious fix: many farms sit where mobile data is expensive, intermittent, or absent, so anything requiring a live connection would have been abandoned in week one.
What we built.
Offline-first capture
The full record can be completed with no signal and syncs when a connection appears, which is what made adoption possible on farms with little or no coverage.
Pond-level culture records
Stocking, species, area, days of culture, daily feeding, water quality, and sampling all recorded against a specific pond, so every entry reconciles to a cycle instead of floating free.
Water quality and safe ranges
On-site readings captured against per-species safe ranges, so a technician sees immediately whether a pond is drifting out of tolerance rather than finding out at the next visit.
Sampling with photo diagnostics
Specimen photographs and visual check-ups captured in the app, giving remote experts something concrete to assess and building a visual history per pond.
Feed and survival analytics
Feed conversion ratio and approximate survival calculated from what was actually logged, turning daily data entry into the two numbers the business runs on.
Per-country administration and alerts
Each subsidiary has its own local admin who can issue alerts, for disease season or an outbreak nearby, to farmers and technicians in their own country and language.
What it is worth when it is used.
A record-keeping app does not grow shrimp. It shortens the gap between something going wrong in a pond and somebody knowing about it, and that gap is where the value sits. Modelling that conservatively, on a one-hectare pond, gives a sense of what the shortened gap is worth.
modelled gross benefit per pond, per crop
modelled annual benefit across 100 ponds at 70 percent adoption
technician reporting time returned per year at that scale
Assumptions: a one-hectare pond running a 120-day crop, 6,000 kg baseline harvest at 75 percent survival and a 1.45 feed conversion ratio, shrimp at 4 dollars per kg and feed at 1.20 dollars per kg, two crops a year. The expected case moves survival to 78 percent and feed conversion to 1.38, which adds 240 kg of shrimp and saves 89 kg of feed. The FAO cultured species profile for whiteleg shrimp puts intensive feed conversion at 1.4 to 1.8, so the 1.45 starting point is a normally run pond rather than a weak one.
These are modelled outcomes on stated assumptions, not measured results from this deployment. The model also depends entirely on field behaviour: roughly 70 percent active usage, 75 percent daily feeding-log completion, 80 percent same-day water-quality logging, and technician response to red alerts within 24 hours. Without those, what remains is the administrative saving rather than the biological gain, which is about a tenth of the case.
Built with.
What powered this build.
Have a system that's holding you back?
Walk us through it. We'll tell you what we'd build, what we'd keep, and what it would take.
Start the conversation