A retail ERP I build and deploy for multi-store retailers in India and the UAE: fast billing, live stock across branches, purchase and supplier management, GST or UAE VAT compliance, and an owner’s dashboard that answers “aaj kitna hua?” from your phone. Status: deployed and demonstrable.
What does multi-store retail actually run on today?
Eleven spreadsheets, three WhatsApp groups, and one accountant doing jugaad at month-end — that’s the honest stack in most growing chains. Each store’s billing is its own island. Stock exists as totals nobody trusts: the sheet says 40, the shelf says 32, and which store the difference walked out of is a genuine mystery with a genuine cost. Purchases follow instinct rather than sell-through. The udhaar khata lives in the manager’s memory, which resigns when he does. And the owner reconstructs the day’s truth by phone call — store by store, every evening, forever.
Add a second city and it compounds; add a second country and it fractures. The Dubai store closes VAT returns under the FTA while Mumbai files GST, and the owner reconciles an empire across two tax regimes with a calculator and patience. Retail’s margins don’t forgive this: the leaks — walked stock, dead inventory, unclaimed supplier terms, forgotten udhaar — are individually chhota and collectively the difference between a good year and a shrug.
What does the Retail ERP include?
- POS billing that survives the market’s realities. Fast — barcode or quick-key, because kirana-style unlabelled items deserve seconds too. Offline-capable — power cuts are a feature of the market, so bills queue locally and sync later. GST e-invoicing native for India (mandatory above the ₹5-crore threshold — cbic.gov.in), FTA-compliant VAT for UAE stores (5% since January 2018).
- One stock truth, all branches. Store-wise levels, inter-branch transfers tracked like the movements they are, variance reports after every count — the sheet-versus-shelf gap named per store, per item. The same event-sourced discipline as my ledger work: every movement a fact, totals as projections you can trust.
- Purchase with sell-through brains. Reorder points from actual movement, supplier ledgers with payment ageing, scheme and margin tracking — buying discipline that survives the buyer’s day off.
- Customer khata, dignified. Udhaar tracked in-system with sendable statements — credit on bharosa, records on Postgres. Schemes and loyalty basics included; the notebook retires with honour.
- Owner’s dashboard, phone-first. Sales, margins, dead stock, dues, per branch and consolidated — because owners live on the floor and in the car, not at a desk. “Aaj kitna hua?” becomes a glance, not three phone calls.
- India and UAE in one system. Same core, two tax brains, per-entity books. Your Deira store and your Mumbai supplier finally agree on a number. Khalas.
How does deployment protect the business while it changes it?
Pilot first, waves after, parallel-run throughout. One store — ideally the one whose team complains most honestly — runs the system for two weeks alongside the old process. Counts verified, staff feedback applied, the offline behaviour tested by actually pulling the cable. Then branches convert in waves, each inheriting the pilot’s fixes. The old system stays readable until the last wave closes; nothing is bet on a weekend.
Masters and balances migrate up front — items, customers, suppliers, stock, udhaar balances — verified per store before its cutover, using the same staged method that moves factories without stopping them. Counter staff train in a day because the counter screen is deliberately simple; the deeper back-office training goes to the two people who need it. And the accountant keeps Tally via the bridge, fed clean vouchers instead of manual entry — the CA’s blessing collected early, as always, because that veto sinks more rollouts than any bug.
What’s the honest pitch?
This ERP makes a chain’s truth arrive without being asked for. It will not choose your merchandise or charm your customers — it makes sure the stock you bought is the stock you sold or still have, that the udhaar you extended comes back, that the supplier terms you negotiated are the terms you got, and that the day’s number reaches your phone before you reach home. Most chains find the system pays for itself in the first stock-take — the variance report has a way of ending arguments that registers never could, seedhi baat.
The trade context is at /industries/retail and the systems-buyer’s view at /industries/retail-systems; the distribution sibling — schemes, van sales, beat collections — is /products/distribution-erp. The WhatsApp layer that turns the customer khata into polite automated reminders is /products/whatsapp-automation.
What does the owner’s day become?
Morning, in the car: yesterday’s number per store, already on the phone — sales, margin, the two stores that dipped, the dead-stock alert on winter inventory in March. No calls made, no calls needed. Midday: a supplier negotiation armed with actual sell-through — “yeh item chal nahi raha, scheme do ya wapas lo” said with a report instead of a feeling. Evening: the udhaar list that crossed thresholds, three customers nudged politely by WhatsApp without anyone’s awkward phone call, and tomorrow’s transfer approvals queued — the Borivali store sends slow stock to the Thane store where it moves. Month-end: GST filings drawn from clean data while the accountant checks rather than reconstructs, and the stock-take variance report that ends the oldest argument in retail — where the difference walked out, per store, per item. The chain runs on truth arriving by itself; the owner’s attention goes back to buying well and serving well, which is the actual business.
Related reading
/industries/retail — the trade page · /products/distribution-erp — the wholesale sibling · /products/whatsapp-automation — reminders and order updates · /work/erp-migration — the migration method
The demo that ends the argument
One store or twenty — bring your worst stock-take variance and your udhaar notebook to the demo, and we’ll run both through the system live. Does what it says on the tin: /contact.
Questions I actually get
How is it priced?
Deployment plus annual support, sized by branch count — not per-seat rent that punishes you for growing or training more staff. The quote is fixed after the first call and the pilot store scope is priced separately, so you can prove the fit for one store's worth of risk before committing the chain.
How does migration from Tally, Excel or a legacy POS work?
Masters and balances import first — items, customers, suppliers, stock, udhaar khata — then a parallel-run month where the old system stays available while the new one earns trust. Counts are verified per store before cutover. The old system retires with dignity, and nobody's Diwali season is bet on a big-bang switch.
What exactly happens when the power or internet goes?
Billing continues offline at the counter — bills queue locally and sync when the connection returns. Stock conflicts from the offline window resolve by rule with a variance trail, not by whoever shouts first. Power cuts are a feature of the market, so they are a feature of the product; the demo includes pulling the network cable.
What is the right rollout order for a multi-branch chain?
Pilot store first, two weeks, with the store team that complains most honestly. Then waves of branches with the pilot's learnings applied — never big-bang, because retail cannot pause for software. The chain's owner dashboard becomes truthful branch by branch, which is more useful than a simultaneous fiction.
How much training do counter staff need?
Productive in a day — the counter screen is deliberately simple: bill, stock lookup, customer khata, done. If they use WhatsApp they will manage this. The complexity lives in the back office where it belongs, and the back-office training is a separate, deeper session for the two people who need it.
Does it need special hardware?
No — standard barcode scanners, standard thermal printers, ordinary Android devices or PCs at the counter. No proprietary terminals, no hardware lock-in, and existing printers usually survive the migration. Where a device genuinely needs replacing I will say so with a price, before you commit rather than after.