Every channel has its own stock figure
Own site, two marketplaces, three stores. The moment they disagree you are either overselling or holding back sellable stock.
D2C storefronts, omnichannel POS, distributor ordering portals and marketplaces. Built around one inventory truth, with GST and logistics wired in from the start.
Own site, two marketplaces, three stores. The moment they disagree you are either overselling or holding back sellable stock.
A cancelled marketplace order damages the seller rating. That costs future orders long after the refund is done.
Marketplace and gateway payouts arrive in different formats on different cycles. Differences get written off because tracing them takes longer than they are worth.
Retail software development covers what a retailer or brand sells through and runs on. Online storefronts, point of sale in stores, inventory and warehouse management, order management across channels. Plus the payment, logistics and accounting integrations that turn an order into a delivered, invoiced, reconciled transaction.
Almost every retail software problem is an inventory problem wearing a disguise. Stock counted in six places will eventually disagree, and then you are cancelling orders or sitting on stock you could have sold.
The second problem is that the sale is only the start. An order becomes a picked shipment, a GST invoice with the right place of supply, a tracking number, a delivery confirmation, then a settlement line. Retailers rarely lose money at the till. They lose it in returns and settlement gaps.
So we build order management and inventory as the centre. A good storefront on a broken order system just produces angry customers faster. Once one system owns stock and order state, every channel becomes a way into the same truth.
Every card links to the practice that does the work. Follow one if you want the detail.
Fast, search-friendly storefronts with product, variant and bundle handling. Headless, so one catalogue serves web, app and marketplace.
ExploreOne stock ledger across warehouses, stores and channels. Allocation rules, reservations, transfers and cycle counts, so the site shows what exists.
ExploreBilling counters that keep working offline, loyalty recognised across stores and online, and buy-online-pickup-in-store handled as one order.
ExplorePrice lists per customer tier, credit limits and outstanding visibility, schemes and slab discounts, and reorder built for somebody ordering weekly.
ExploreSKU-level forecasts using seasonality and promotion history, turned into reorder points and purchase suggestions instead of a chart nobody acts on.
ExploreGoogle Merchant Center and Meta catalogues kept in sync with live stock and price. Campaigns judged on margin after returns, not on ROAS.
ExploreSearch that understands how customers describe products, recommendations from real basket behaviour, and assistants handling returns within policy.
ExploreThe situation it answers, the architecture, every module, the systems it plugs into and how long it takes. These describe our work, not one client's project. Named clients with measured results are in the case studies.
A brand selling on marketplaces that wants its own channel, better margin and a direct customer relationship. The catalogue has variants, bundles and frequent promotions. Marketing needs to launch campaign pages without waiting for a developer.
A headless commerce backend holds catalogue, pricing and orders. An Astro or Next.js storefront renders statically wherever possible, for speed and search. Content and campaign pages come from a headless CMS, so marketing ships them alone.
Invoices carry the correct place of supply and HSN codes, raised through the NIC IRP where turnover requires it. Return and refund timelines are stated on the product page under e-commerce consumer protection rules.
A retailer with several stores and an online channel, where each store runs its own billing software and stock is counted per location. Online orders cannot pull from store stock. A customer cannot return in a different store. Head office sees consolidated numbers after month end.
An offline-capable POS built local-first. It queues transactions and syncs to a central order and inventory service, with conflict handling for stock movements recorded on both sides during a disconnect.
Each store bills under the correct GSTIN with the right state tax treatment, and interstate stock transfers are documented as required. Cash is recorded per shift down to denomination, which makes a till difference traceable.
A manufacturer or importer selling through distributors and retailers. Orders arrive by phone and WhatsApp. Prices differ by customer tier and region. The sales team spends its week confirming stock and outstanding balances instead of selling.
A React portal and an Android app for field sales, on an order service that resolves customer-specific pricing, credit limits and scheme eligibility at the point of order. Nothing waits for a later approval step.
E-way bills are generated for consignments over the value threshold, with vehicle and transporter details captured at dispatch. Credit notes for scheme claims follow the GST treatment for post-sale discounts.
An operator building a category marketplace where independent sellers list and fulfil their own stock. The hard parts are seller quality, catalogue consistency when several sellers list the same product, and settlement sellers trust enough to keep listing.
The catalogue service separates the product from the seller offer, so several sellers compete on one product page instead of creating duplicate listings. A settlement engine computes commission, shipping recovery and returns per order line.
Seller settlement runs through split settlement on a licensed payment aggregator, never a pooled operator account. Seller identity, GSTIN and grievance officer details are held and displayed as e-commerce rules require.
A retailer buying from last season sales pulled into a spreadsheet, with no view of which SKUs tie up working capital and which went out of stock during their best week. Season-end markdowns are the recurring symptom.
A warehouse consolidates POS, e-commerce, marketplace and purchase data nightly, modelled to SKU-store-day grain. Forecasting models feed reorder suggestions back into the ordering workflow instead of stopping at a report.
Marketplace settlement data is reconciled against invoiced sales, so GST returns are filed against what was actually settled. Return credit notes match the original invoice instead of being adjusted in aggregate.
Get these wrong and the build is not late, it is unusable. They are design inputs on day one.
Above the notified turnover, invoices must go through the NIC Invoice Registration Portal and get an IRN before they are valid. Consignments over the value threshold need an e-way bill with transporter and vehicle details. Both work far better generated at billing or dispatch than reconciled later.
Interstate and intrastate orders attract different tax components, and place of supply follows the delivery address, not billing convenience. Getting it wrong is quiet. It surfaces at return filing, by which point hundreds of invoices need correcting.
Selling on the Open Network for Digital Commerce means implementing the network catalogue, order, fulfilment and issue-resolution flows. Build the catalogue and order model to that structure from the start and joining is configuration. A bespoke model means a rebuild.
Seller identity, country of origin, return and refund policy, and grievance officer details must be displayed, and stated cancellation timelines honoured. These are workflow requirements, so they belong in the product data model, not a policy page nobody updates.
A marketplace collecting for sellers cannot hold those funds in its own account. Split settlement through a licensed payment aggregator is the compliant route, and it changes how the order ledger and payout cycle get designed.
If the catalogue is straightforward and you sell through one channel, a platform is usually right and we will say so. Custom earns its cost in three cases. Inventory shared across stores and marketplaces. B2B pricing and credit rules the platform cannot model. A checkout that must behave in a way the platform forbids.
One stock ledger owns the true quantity. Each channel gets an allocation instead of the full figure, with reservations held from the moment an order is placed. Channel updates push on every movement instead of a periodic sync, and a per-SKU buffer is configurable where a cancellation costs more than a missed sale.
Yes. Sales invoices, credit notes, purchases and stock movements post to Tally on a schedule, with a reconciliation report showing anything that did not transfer. If your accounting system is something else, the pattern is the same. The operational system stays the source of truth and posts summarised entries.
Modelling the catalogue, order lifecycle, fulfilment states and issue resolution the way the network specifies, then writing the protocol adapter and completing onboarding. Building the underlying model that way costs nothing extra during a fresh build, so we default to it whether or not you plan to join.
Ten to fourteen weeks for catalogue, checkout, payments, logistics, GST invoicing and returns. Shorter if you stay on a platform and we build the front end and integrations only. The timeline usually moves for content and photography, not engineering, so starting catalogue work in parallel is worth the coordination.
Yes. Billing runs locally and queues transactions, syncing when the connection returns, because a store cannot stop billing. Stock movements made offline reconcile on reconnect with conflict handling. The shift close is computed locally, so a disconnected day still balances.
Tell us the workflow that keeps breaking. We will say honestly whether software fixes it, and what it would take.