If you're setting up Google Shopping ads for a Shopify store in India, the campaign is not where you'll win or lose. The feed is. Roughly four out of five accounts we're handed have the same three problems: a price in the feed that disagrees with the price at checkout because of how GST is configured in Shopify, product titles written for a brand deck instead of for what people actually type, and a ROAS target set by gut feel rather than by contribution margin. Fix those in that order, then spend money.
Here's the sequence we run, and the arithmetic that tells you when the account is actually profitable rather than just busy.
Merchant Center setup for an Indian Shopify store
Do this before you touch Google Ads. It takes an afternoon.
- Business details, in INR, targeting IN. One feed, country India, language English. Don't add a second target country yet even if you ship to the UAE — a multi-country feed with one price list is the fastest route to disapprovals.
- Claim and verify the domain. This is where Shopify stores stall. Verify the custom domain (yourbrand.com), not the myshopify.com address. If Search Console is already verified for the domain, Merchant Center picks it up in seconds.
- Returns policy and contact details. Merchant Center asks for both, and a store with no returns page, no phone number and no address gets flagged for misrepresentation. In India that flag is common on new stores because founders launch with a Terms page copied from a template and nothing else.
- Shipping settings that mirror Shopify. If your storefront gives free shipping above ₹999 and charges ₹69 below it, build exactly that in Merchant Center. Add handling time of one business day and transit time by region — pull real numbers from your Delhivery or Shiprocket lane reports rather than guessing. Metro-to-metro surface is usually 2–4 days; tier-3 non-metro is 5–8. Overstating it gets you a bad first order; understating it costs you clicks to competitors showing a faster date.
- Skip tax configuration. Unlike the US, there is no tax table to build. Indian prices are quoted GST-inclusive, and the feed price must be the number the customer pays. Which brings us to the thing that breaks most often.
The price mismatch error is a GST setting, not a feed bug
"Mismatched value (price)" in Merchant Center means Google's crawler visited your product page and saw a different number than the feed submitted. In Indian Shopify stores, these are the usual causes, in the order we find them:
- Tax-exclusive pricing. In Settings → Taxes and duties, if "All prices include tax" is off, your ₹1,000 product page becomes ₹1,120 at checkout for a 12% HSN. The feed says 1000. Google sees 1000 on the page, then the checkout adds tax, and the offer gets suspended for misleading pricing. Almost every Indian catalogue should be tax-inclusive. Set it once, then re-check your margins, because turning it on changes your net realisation per order.
- Discount apps that rewrite the displayed price. Cart-level automatic discounts, quantity breaks, "extra 10% off on prepaid" scripts. The crawler sees the struck-through price or the discounted one depending on how the theme renders it, and the feed sees neither.
- Stale sync during a sale. Feeds refresh on a schedule. Drop prices at midnight for an end-of-season sale and you'll spend the morning with half the catalogue disapproved. Schedule the price change and let the feed pick it up, or use sale price effective dates.
- Theme structured data disagreeing with the feed. Google trusts what it crawls. If your JSON-LD outputs the variant's compare-at price, or a price without tax, you get a mismatch even when the visible number is right. This is a theme fix, not a Merchant Center fix — one of the more common reasons merchants end up needing to hire a Shopify developer for what looks like an ads problem.
- Markets price adjustments. If you run Shopify Markets with a price adjustment for India or a geo-based currency, the crawler's IP decides what it sees.
One rule that saves hours: the price on the landing page must be visible without login, without a pincode check, and without a popup. Gate anything and the crawl fails.
Which feed app you actually need
The free Google & YouTube channel that Shopify ships with is enough for a single-country INR store with a straightforward catalogue. It syncs products, handles variants, pushes availability. If that's you, don't pay for anything.
You need a dedicated feed app when you want feed rules — rewriting titles in bulk without changing your storefront titles, mapping a metafield to product_type, excluding SKUs below a margin threshold, splitting the feed by custom label so you can bid differently on your top 30 sellers. Simprosys is the one we see most in Indian accounts because it handles GST-inclusive pricing and Indian shipping tables without fuss; DataFeedWatch and Feedonomics are for catalogues in the tens of thousands where a mapping error costs real money. Pick on rules, not on price. A feed app that saves one suspension during Diwali week has paid for itself for the year.
On GTINs: if you manufacture your own products, declare no identifier and move on. If you resell branded goods and leave GTINs blank, your reach on those SKUs quietly halves.
Write titles for what people type, including the Hinglish version
You cannot add keywords to a Shopping campaign. The title is the keyword. So the first 60–70 characters have to carry the words a real buyer uses, and in India that word is frequently not the English one your brand book prefers.
People search kurti far more than "A-line tunic". They search chappal, saree and sari, mixer grinder, bedsheet double bed, dupatta set, gold plated jhumka. Transliterated queries are a large slice of Shopping volume and they don't show up in your Search Console data because your blog doesn't rank for them.
A title pattern that works: product type + key attribute + brand + variant. "Cotton Anarkali Kurti Set with Dupatta – Brandname – Navy, M" beats "Brandname Meera Collection Set 03" by a distance. Check the Search Terms Insights report after two weeks, find the query clusters you're matching on, and push those exact words into titles for the SKUs that earn them. This is the same discipline as writing collection pages that rank, applied to 150 characters.
Signalling COD and delivery when there's no attribute for it
There is no cash-on-delivery field in the Google feed. COD is more than half of orders for a lot of Indian D2C brands, and buyers who need it will bounce off a PDP that doesn't say so. Since the Shopping tile can't carry it, everything you do here is post-click:
- A COD badge and a pincode serviceability check above the fold on the product page, not buried in an accordion.
- The words "Cash on Delivery available" inside the first line of the product description, which Google also uses for matching.
- Merchant Promotions for offers that genuinely apply — a prepaid discount is a good one, and it shifts mix in your favour.
Do not stuff "COD" into the title. It costs you the character budget you need for the category term, and it doesn't render usefully.
The harder truth: Google Ads cannot exclude by pincode. You can exclude cities and states, and that's a blunt instrument for RTO control. Manage refusals at the fulfilment layer instead — order confirmation on WhatsApp, an address-quality check, prepaid nudges on high-risk pincodes.
What Google Shopping ads cost in India
In the accounts we run, broad apparel, home and accessories Shopping clicks land somewhere around ₹4–₹12. Niche high-ticket categories with few competitors sit higher, ₹25–₹60 is normal for things like professional equipment or B2B supplies. Branded queries on your own name are cheap and, if a reseller is bidding on you, worth defending.
Budget floors matter more than CPC. Below roughly ₹500–₹700 a day, an automated campaign gets so few conversions per week that the bidding model never leaves guess mode. If that's your ceiling, run Standard Shopping with manual CPC on your 15–20 best SKUs and turn on free listings, which cost nothing and will send some traffic on their own. That's the honest answer for a small business: don't start with the full-funnel setup, start with the catalogue subset that already sells, and don't let a ₹300/day budget get spread across 4,000 SKUs.
The ROAS number that actually matters
Google reports revenue including GST, and it counts every COD order that a customer later refuses. Both inflate the number. Work out your break-even from delivered contribution instead. Here's a real calculation on a mid-market apparel catalogue:
AOV ₹1,600 inclusive of 12% GST, so net revenue per order is 1600 ÷ 1.12 = ₹1,428.57. COGS ₹520. Packaging ₹15. Forward freight ₹80. Mix is 60% COD, 40% prepaid; RTO on COD runs 20%.
Per 100 booked orders: 12 come back, 88 deliver.
- Net revenue: 88 × 1,428.57 = ₹1,25,714
- COGS on delivered: 88 × 520 = ₹45,760
- Freight out on all 100: ₹8,000. Return freight on 12: ₹960
- Packaging on 100: ₹1,500. Write-off on returned stock at 10%: 12 × 52 = ₹624
- Gateway fees, prepaid: 40 × ₹32 = ₹1,280. COD remittance: 48 × ₹40 = ₹1,920
- Total variable cost: ₹60,044
Contribution before ad spend is ₹65,670. Google will report gross booked revenue of ₹1,60,000. So contribution as a share of reported revenue is 65,670 ÷ 1,60,000 = 41%, and break-even ROAS as Google reports it is 1 ÷ 0.41 = 2.44.
That's the whole game. At 2.4 you are working for free. At 3.2–3.5 you're funding overheads and growth. Anyone quoting you a universal "good ROAS for India" without asking your COD share and RTO rate is guessing. Cut RTO from 20% to 12% in that same model and break-even drops to roughly 2.2, which is a bigger lever than any bid adjustment.
Performance Max, or Standard Shopping first
Performance Max for a Shopify store in India works well once the account has consistent conversion volume — call it 30 to 50 orders a month attributable to Google. Below that, it burns budget on Display and Discovery inventory while it learns, and you won't be able to tell whether the feed or the creative was at fault.
Our default sequence: Standard Shopping first, segmented into a high-margin SKU group and everything else, manual or target-ROAS bidding once there's data. Then a feed-only PMax campaign with no image or video assets, which behaves close to Shopping and gets you the extra surfaces. Then add assets. Always add a brand exclusion list to PMax so it isn't buying your own name and reporting it as new demand — this single setting has made accounts look 30% worse and be 30% better.
Google Shopping vs Meta ads for Indian D2C
Different jobs. Shopping harvests demand that already exists, so it's capped by how many people search your category. Meta creates demand, so it scales further but starts colder and costs more per first order in most categories we work in.
Two practical points. If your product has a name people already search — cotton bedsheet, whey protein, wall clock, running shoes — start with Shopping, because the CAC is usually lower at small budgets and the setup is cheaper. If you're selling something nobody searches for because the category is new, Shopping has nothing to harvest and Meta is the right first channel.
Second, and this is worth money: Shopping traffic tends to pay prepaid more often than interruption traffic from a Reels placement. Someone who typed the query is further along. A better prepaid mix means less RTO, which means a lower break-even ROAS on the same margin structure. When you compare channels, compare delivered contribution, not platform-reported ROAS.
The landing page is part of the feed
Most of your Shopping clicks arrive on a 4G phone in a tier-2 city with three bars of signal. A product page with a 4.2 second LCP loses a chunk of those clicks before the image paints, and you paid ₹8 for each of them. Google's ad system doesn't publish a Shopping quality score you can stare at, but slow landing pages absolutely show up as a lower conversion rate and a higher effective CPA.
Before scaling spend, get the PDP under 2.5 seconds LCP on a throttled mobile connection. Usual culprits: a hero image served at 2000px, four review apps loading their own scripts, and a chat widget booting on page load. We wrote up the approach in more detail on Shopify speed optimization, and if you want the score tracked continuously rather than checked once a quarter, SwiftStore scans and monitors it from inside the admin.
Where to start this week
Open Merchant Center, sort the diagnostics by affected offers, and fix the price mismatches first. Then rewrite the titles of your top 20 SKUs with the category word first. Then compute your own break-even ROAS with your real COD share and RTO rate, so the number you type into a target-ROAS bid strategy comes from your P&L rather than from a webinar.
If you'd rather have someone else read the diagnostics, our free store audit covers the feed, the tax configuration and the PDP speed together, since in practice they're the same problem showing up in three places.