If you came looking for a Shopify retail roundup for 2026, here is the part that actually affects your P&L: most of the headline retail features arriving in Shopify are built on top of Shopify Payments, and Shopify Payments still isn't offered to merchants in India. So the unified card reader, the tap-to-pay-on-phone story, the automatic reconciliation between till and admin — those land in the US, UK, Canada and Australia and stop at our border. What you do get, and what's genuinely worth acting on, is the inventory, order-routing and customer-record side of retail. That part works here, and most Indian D2C brands running two or three stores are using maybe a fifth of it.
This is being read in India right now because a lot of D2C brands crossed into physical retail in 2025 — a Bengaluru flagship, three kiosks in a mall chain, a shop-in-shop at a department store — and the reconciliation is now somebody's Monday problem. Below is what we'd tell that person.
POS runs fine here. The payment hardware doesn't.
Shopify POS Lite and POS Pro both work on an Indian store. You can ring up a sale on an iPad, pull the customer record, check stock at the warehouse, and push the order into the same Orders list as your web sales. What you cannot do is take the card payment through Shopify. The first-party readers and phone-based tap-to-pay are tied to countries where Shopify Payments operates.
So every Indian store we set up runs a split: the swipe happens on a Pine Labs or Razorpay terminal, or on a static UPI QR taped to the counter, and the POS records it against a custom payment type. Create separate custom types for UPI, Card – terminal and Cash rather than one lazy "Other". It costs nothing and it means your end-of-day cash-up actually tells you something. The MDR reconciliation still happens outside Shopify, in your accounting tool, once a day.
Is that annoying? Yes. Is it a reason to buy a separate retail POS and sync it back? Almost never. The moment you split the product catalogue across two systems you have created a permanent job for a human being. Keep one catalogue.
GST is the part every demo skips
Shopify's native invoice is not a GST tax invoice. It has no HSN code, no split of CGST/SGST/IGST, no place of supply, no GSTIN on it. For online orders you have probably already patched this with an invoicing app or by pushing orders into Zoho Books or Tally. For retail, the gap is louder, because a walk-in customer buying for their company will ask for a B2B invoice at the counter, and your associate will not be able to produce one.
Two workable answers. Either you use an invoicing app that hooks POS orders as well as online orders — check that specifically, several handle web orders only — or you keep a lightweight billing tool at the counter for the handful of B2B walk-ins and let Shopify record the sale for inventory. The second is ugly and it works.
The other GST detail that catches multi-city brands: if you hold stock in a state, you need a GSTIN registered there, and stock moving between your own locations across state lines needs an e-way bill above the threshold even though nothing has been sold. Shopify's transfer feature will happily move 400 units from Bengaluru to a Delhi store and generate no document your transporter can show at a checkpoint. Build the e-way bill step into your transfer SOP or your goods sit at a border.
One inventory pool, and the three places it cracks
The real prize in a retail-plus-online setup is a single stock number. Shopify does this well. The cracks are predictable:
- Overselling the last unit. Your Meta ad drives a sale at 9:47pm for a size you sold off the shop floor at 7:20pm and the associate forgot to ring properly. Set a buffer of one unit on store locations for fast-moving SKUs. Yes, you'll leave a little stock unsold. It's cheaper than a cancellation email.
- Fulfilment priority. By default Shopify will route an online order to whichever location can fill it by your priority list. If your store is high on that list, you will end up asking a retail associate to pack and hand over to a Delhivery pickup during shop hours. Sometimes that's the point. Usually it isn't, and priority should sit warehouse-first with stores as fallback only.
- RTO restocking. A refused COD parcel comes back to the warehouse, not the store. If nobody marks it restocked, that unit is invisible for a week. This is the single most common source of "the system says we have it" in Indian retail setups, and it's process, not software.
COD, RTO and the arithmetic of a prepaid nudge
Everyone knows COD drags returns. Fewer people have done the sum on the fix, so here it is with numbers you can swap for your own.
Take a brand doing 1,000 orders a month, 55% of them COD, with a 22% RTO rate on that COD volume. That's 550 COD orders and 121 that come back. Forward leg roughly ₹120, reverse roughly ₹110, so ₹230 per failed order. Call it ₹27,830 a month burned on parcels nobody paid for, before you count packaging and the units that come back unsellable.
Now the standard fix: offer 5% off for prepaid. Say it converts 15% of your COD orders to prepaid — 82 orders. At an average order value of ₹1,800, that discount costs you ₹90 an order, or ₹7,380. The RTO you avoid is 82 × 22% = 18 orders, worth 18 × ₹230 = ₹4,140. You're down ₹3,240 for the month.
A flat ₹50 prepaid incentive on the same shift costs ₹4,100 against ₹4,140 saved. Roughly a wash on shipping, and now the gain is real but different: you have the cash 12 days earlier and your working capital cycle tightens. That's the honest case for prepaid nudges. Not shipping savings. Cash timing. Anyone selling you a 10% prepaid discount as a cost-saving measure has not run the numbers.
Where retail changes this: a store gives you a free returns counter. Let online customers return or exchange in-store, and you convert a reverse-logistics cost into footfall. Exchanges done at the counter convert to an upsell often enough that we now build the exchange flow before the returns flow.
The shop-floor tablet is a search problem
Endless aisle sounds like a feature. In practice it is one associate, one iPad, one impatient customer, and a catalogue of 3,000 SKUs. If the associate takes forty seconds to find the black one in a 38, the sale walks. The POS search is fine for a small range and becomes a liability past a few hundred products, because it matches on title and not much else.
The same problem shows up on your collection pages, where mobile users on patchy 4G bail after two taps. If your catalogue is deep — jewellery, apparel with size-and-colour variants, anything with 20+ filterable attributes — filtered navigation and a search that understands synonyms and misspellings is the highest-return thing you can fix this quarter. We build this with FilterPro on stores where the native filters have run out of road.
Festive 2026: lock dates, not intentions
Diwali falls in early November 2026, which is late compared to 2025. That's a longer runway and a very short post-festive window before end-of-season clearance. Two practical consequences.
First, your peak-load work has to be done by mid-September, not mid-October. Theme changes, app installs and checkout edits during the last week of October are how stores go down. We freeze deploys ten days before the peak and we do not make exceptions, including for the client who really wants a countdown timer.
Second, if you're opening a store before festive, your inventory transfer and state registration paperwork needs a six-week head start. A GSTIN in a new state is not a same-week thing, and neither is a mall's onboarding.
What we'd ignore this year
Shopify Plus for the sake of retail. If your reason for upgrading is POS Pro across a handful of locations, POS Pro is priced per location on any plan and you don't need Plus to get it. The Plus case is checkout customisation, B2B, high order volume or wholesale channels — we've written about when the upgrade actually pays for itself in India, and "we opened two stores" is not on that list.
ONDC as a retail channel. Worth a look if you're in grocery or a category where price is the only variable. For a brand with any margin structure and any brand equity, the integration effort in 2026 still costs more than the orders return. Revisit in a year.
Also: buying a second analytics tool because POS reporting confused you. Reporting is usually fine. The data going into it isn't, because of the custom payment types and the unrestocked returns above.
Where to start
Pull last month's POS orders and count how many are sitting under a generic "Other" payment type. If it's more than a handful, that's your first hour of work — set up proper custom types and brief the counter staff. Then check your fulfilment priority list and make sure your Bengaluru store isn't quietly packing web orders during peak shopping hours.
If you'd rather have someone look at the whole setup — inventory routing, GST invoicing on POS orders, storefront speed before festive — our team in Bengaluru does a free store audit and will tell you the two things worth fixing rather than the twenty that aren't.