Operations software for an agricultural supply business working across hundreds of thousands of acres
An agricultural group that supplies farmers across Punjab — farm inputs, dairy, crop production, machinery, veterinary services and lab testing — operating at the scale of hundreds of thousands of acres and hundreds of locations.
I built the public web platform and the internal operations dashboard behind it.
Two things make this different from ordinary commerce software.
The first is the user. This is not a customer browsing at a desk on good broadband. It is a business whose customers and field staff are outdoors, on phones, on connections that come and go — which changes what you can assume about page weight, session length and how much a form is allowed to demand in one go.
The second is the business itself. Supply, dairy, machinery, veterinary and testing share customers and almost nothing else: different units, different cycles, different definitions of what a transaction even is. Software that models one of them well and forces the rest into the same shape produces an operations team quietly keeping the real records in a spreadsheet.
Full-stack: the Next.js front end and the Node.js backend, including the internal dashboard the operations team uses.
The temptation with a multi-line business is to build one abstract 'order' and make everything an instance of it. It demos well and it fails in month three, when the dairy team needs something the model cannot express and starts keeping their own records. Modelling the differences honestly costs more up front and is the reason the software still gets used a year later.
A Next.js front end over a Node.js API. The data model keeps the business lines distinct where they genuinely differ and shares only what is actually shared — principally the customer and the location.
Built for phones and imperfect connections rather than for a desk: short forms, few assumptions about session length, and pages that do not need a fast network to be useful.
Operations software shaped around a business with several genuinely different lines, built for users who are in a field rather than at a desk.
If this sounds familiar
Building something similar?
Most of these problems show up again in different clothes. Describe yours and I will tell you what I would look at first.