Jhotpot — E-commerce Suite

The build

Bangladeshi online shopping has a shape the global SaaS builders never modeled: most orders are cash on delivery, most couriers remit that cash on their own schedule, and a real slice of orders simply never receive anyone. A store platform that ignores this is a demo; one that embraces it is a business tool. Jhotpot was built from that truth: COD as a first-class flow with a handling fee and a confirmation call to blunt fake orders, prepay that actually saves the customer money, courier integrations that reconcile every taka remitted against every order expected, and a seller experience that runs from a phone between prayers. It sells this studio's own merch today; a merchant store on it is a configuration, not a rewrite.

What we handled

• COD-first: handling fees, confirmation-call prompts, per-buyer fake-order risk history on the order card
• Prepay incentives wired to bKash/Nagad/SSLCommerz — discount math closes on product page, checkout and ledger
• Courier integrations: pickup booking, labels, tracking sync, and a COD settlement ledger that finds the mismatch
• Campaign engine: flash sales, bundles, coupon ledgers, cart-reserved inventory and SMS/email nudges with countdowns
• Seller phone app: new-order pushes with accept timers, one-tap print-and-pack, stock edits in the queue
• Merchant-owned everything: own domain, own customer list, documented exports, no platform commission on sales

Platforms are designed for the country's habits, not the customer's credit card

Strip global e-commerce lore of its optimism and the local reality is boringly specific: a customer adds three items, chooses COD because trust is earned at the doorstep, the seller calls to confirm because roughly one in five such orders is imaginary, the courier delivers, collects cash, and remits it whenever the courier feels like. Somewhere in that chain — a lost parcel here, an unrecorded remittance there — the shop's margin dies quietly. Every hosted platform we evaluated treated COD as a checkbox on a settings page. For a Bangladeshi merchant it is the whole business.

So Jhotpot was scoped in reverse: design for the COD path first, make the exceptions (fake orders, remittance gaps, doorstep refusals) first-class states in the software, then bolt on the parts everyone demos anyway — storefront polish, campaigns, analytics. That inversion is the entire product thesis, and it is why the suite can be offered as a Shopify alternative rather than another Shopify.

A store that closes, not just displays

The product page ends in a money decision, so the payment block gets the hero position: COD with its honest handling fee next to a bKash option that saves exactly that fee — the same ৳60 the merchant would spend on a phone call and a courier return. It is a pricing lever, not a payment dropdown, and merchants tune it themselves: discount percentage, flat saving, or free delivery over a threshold.

The checkout keeps its arithmetic perfect — subtotal, delivery per selected courier, COD fee, total — and states the confirmation-call policy in one amber line, because customers behave better when the rules are visible. Catalog needs are covered with the unglamorous basics that decide store speed: variant matrices that map to real stock counts, bulk CSV import for shops that already keep an Excel, and Bangla content that renders correctly at every size instead of the tofu boxes a foreign theme drops on them.

The seller's office is a pocket

Merchants do not live in admin panels; they live in WhatsApp, and Jhotpot meets them there. A new order arrives as a push with an accept timer, the buyer's verification status on the card face, and a print-and-pack button that produces the courier label before the customer changes their mind. Stock edits, price nudges and campaign starts all run from the same phone — the desktop admin has every feature, not because anyone wants more clicks, but because the accountant needs the ledger.

The fake-order machinery hides in plain sight: repeat rejectors accumulate a risk chip on the order card, COD is gated behind confirmation-call prompts above a basket threshold, and a merchant can require bKash advance on the sketchiest categories. None of this needs a fraud vendor or a data-sharing agreement — it is the shop's own history, counted.

Reconciling the courier, every single day

The signature screen is the COD settlement ledger: what couriers should have remitted versus what actually arrived, per order, mismatches tinted red, exports ready for the accountant's argument. It is not a sexy feature and it is the reason shops stay on the platform — money leaks stop when a system counts them daily. Delivery booking, label printing and tracking sync ride along through the courier integrations, so the fulfillment checklist (picked, invoiced, labeled, booked) is the whole desk ritual, digitized without changing anyone's muscle memory.

Payout-side money works on real rails too: bKash and Nagad merchant payments settle into the shop's own account with SSLCommerz as the card route, refunds follow the same ledger discipline, and none of it transits us. The platform's cut is a flat license — no per-sale commission, ever, because percentage models and merchant success stop agreeing at scale.

Eid-grade campaigns, planned like software

The campaign engine treats a flash sale the way an operations person should: a timeline strip for live and scheduled campaigns, discount rules bound to categories with maximum-amount ceilings, coupon ledgers that show claimed versus redeemed, and the load-bearing toggle nobody else ships — inventory reserved at cart and released on a timer, so a viral post cannot sell the same three units eleven times.

Nudges ship as templates that read like a human wrote them: a cart-saver at 12 percent with a countdown, a bundle invite for the second and third item, a restock SMS the day the size comes in. Every message renders identically in the preview pane and the sent inbox, because marketing software that lies about its own preview is just a gambling site for merchants.