The situation
Anant International exports spices, pulses and tea — bought direct from farmers, packed in their own facility in Bhuj, shipped from Gujarat to buyers abroad.
The website problem sounds trivial until you try to solve it. An agri-export price is not a property of the product. It moves with the crop, the season, the currency and the freight, and it moves fast enough that anything printed is wrong within days. On top of that, the margin is not the same for every buyer: a first order and a ten-year relationship are not quoted the same way, and neither party expects them to be.
So the one thing a normal e-commerce site is built to do — show a price and take an order — was the one thing this business could not allow.
The question we asked
When a buyer emails asking for a rate on twenty tonnes, who works out the number, and what do they need in front of them to do it?
The answer was not a system. It was a person, a current cost, and a knowledge of who was asking. That told us the website's job was never going to be pricing. Its job is to get a well-formed question to that person, and to stop the answer getting lost on the way back.
What we decided
No prices anywhere, and no cart. The catalogue exists to let a buyer be specific about what they want, and every product path ends at a quote request rather than a checkout.
The alternative we rejected was the one that gets asked for by default: indicative pricing, a "from $X per tonne" band, updated periodically. It would have been wrong most of the time, it would have anchored every negotiation to a number nobody stood behind, and it would have made per-buyer margins visible in a trade where they are ordinary and private. A stale price is worse than no price, because a stale price looks like a commitment.
What we built
A product catalogue organised so a buyer can specify grade and quantity without a phone call, a quote request flow at the end of every product path, and — the part that does the quiet work — enquiry tracking, so a submitted request carries a reference the buyer can check rather than an email they have to chase.
Around it sits the material an overseas buyer screens on before they will talk to a supplier at all: the certifications, the manufacturing facility, the three locations, and the fact that the packing happens in-house.
What happened
The site is live. A quote request now arrives as a structured enquiry with a reference attached, instead of as free text in an inbox, and the buyer can check its status instead of sending "did you get my email?"
We are not quoting enquiry volumes or conversion rates here. Those are Anant's numbers, not ours to publish, and we would rather show you the mechanism than a figure you cannot check.
What we would not build
A shopping cart. It was the obvious thing to build for something described as an export e-commerce site, and it would have been actively harmful. Adding a cart means committing to a price, and committing to a price is precisely the thing this business cannot do.
We also left out live currency conversion. It looks like a courtesy to an overseas buyer and it is a trap: converting a number you have not quoted yet just produces a second wrong price in a second currency.