Nexora

This is not a startup looking for a market, it's a well-established trade doing roughly ₹100M a month, and every rupee of it moves offline. Sellers, buyers, traders, and the logistics partner meet in WhatsApp groups: sellers post stock lists, buyers bid in the thread, and the winning deal leaves the chat entirely, payment, quality check, and delivery are all settled one-to-one, on trust.

The logistics partner is the quiet centre of the whole operation. It warehouses every seller's stock, runs the quality checks, collects and releases payments, and fulfils the orders. Yet none of that work is visible anywhere, it lives in chat scrollback, phone calls, and paper.

For the business owner this means zero oversight of a nine-figure monthly flow: no deal history, no price discovery, no audit trail when a dispute lands. Solving it mattered because the business had outgrown the medium it runs on, and every competitor is one WhatsApp export away from its entire market.

This wasn't a demand question, the trade exists and thrives. The research job was to map how the floor actually works before moving it: shadowing the WhatsApp groups, tracing real deals end-to-end, and interviewing each role about where deals stall or die. The deal lifecycle that came out of it, list, bid, negotiate, confirm, QC, pay, release, became the backbone of the entire IA.

Two findings shaped the design more than anything else. First, the floor has its own vocabulary, WTS and WTB lots, bids and counter-bids, stock grades, and any interface that renamed those concepts would read as foreign. Second, generic B2B marketplace patterns don't fit: marketplaces assume buyer and seller settle fulfilment between themselves, while here logistics is the market's trust anchor, it holds the stock, certifies the quality, and moves the money. The platform had to be designed around that role, not bolt it on.

1. I.

   Keep the floor's vocabulary.

   WTS lots, WTB requests, bids, counters, stock grades, the platform speaks exactly the language the WhatsApp groups already speak. Adoption is a rename away from failure.

2. II.

   Make the ledger visible.

   Every deal is a stepper, QC → payment → release, and every role sees the same truth from their own side. Transparency isn't a report the owner pulls; it's the interface everyone works in.

3. III.

   Logistics is the spine, not a service.

   The IA routes every confirmed deal through the logistics workspace, stock verification, payment confirmation, release. The platform's trust model is the business's trust model, formalized.

One platform, four role-scoped workspaces, seller, buyer, logistics, admin, sharing a single design system and a single source of truth for every deal. Two artifacts below: the sitemap of the full platform, and the lifecycle a deal travels from listing to release.

What went well:leading with the floor's own vocabulary. The moment stakeholders saw their WTS lots, grades, and bids on screen, instead of generic marketplace language, the conversation changed from “will traders use this?” to “when can my group get on?”

What I'd improve:I'd bring the owner's analytics forward. The admin dashboard came late in the sequence, but oversight was the founding motivation, designing the owner's view first would have sharpened what every other role needed to record.

The lesson: when a business already works, the design job isn't to reinvent it, it's to move its rituals somewhere transparent without breaking their speed. The WhatsApp group was never the problem; the invisibility was.