The fastest way to sort a fashion-literate Shopify partner from a generalist is to ask one question in the first call: how would you structure size, colour and fit for a 400-SKU apparel catalogue running out of two warehouses, and where does the size chart live? If the answer is "options and variants" and nothing else, you're going to end up teaching them your business on your budget. A partner who has actually shipped fashion stores will start talking about metafields, per-variant inventory, whether colour should be a variant or a separate product with linked swatches, and what happens to your Meta feed when a size sells out.
Fashion is the category where Shopify's defaults hurt the most. Everything else in eCommerce sells one SKU per product. You sell one product in nine rows, half of which are out of stock by week three of a drop, and a fifth of what you ship comes back.
Look at their collection pages, not their homepages
Every agency portfolio leads with a homepage: big hero video, a lookbook grid, some parallax. Homepages are the easiest page to make beautiful and the least important page for revenue on a repeat-purchase fashion store.
Open the collection pages instead. On a phone. Count how many taps it takes to get from "Women" to "black midi dresses, size M, under ₹3,000, in stock". If the filters reload the whole page, if out-of-stock variants still appear as available, if the sort options are "Featured / Price / Alphabetical" and nothing else, that's the level of thought you're buying.
Then open a product page and check the size guide. Is it a modal with real garment measurements per size, or a JPEG someone exported from Illustrator in 2021 that's unreadable at 390px wide? The JPEG tells you nobody on that project ever looked at a returns report.
Where fashion actually breaks a Shopify build
Shopify has raised its per-product variant ceiling, so the old hard wall is less of a problem than it was. The operational maths hasn't changed. Three fits × four sizes × six colours is 72 rows per style, each with its own inventory, barcode, weight and image assignment. Get the model wrong at the start and you're re-uploading the catalogue at month nine.
The decisions we spend the most time on with apparel and footwear clients:
- Colour as variant vs. colour as separate product. Separate products give you cleaner collection merchandising, better SEO on "black leather jacket", and per-colour imagery that doesn't confuse the gallery. They also fragment your reviews and make "same style, other colour" a manual link job unless you build the metafield relationship properly. There's no universally right answer. There is a right answer for your catalogue depth and your ad strategy, and a partner should argue for one.
- Where fit data lives. Garment measurements belong in metaobjects, keyed to a size-chart type, reused across hundreds of styles. Not pasted into each product description by an intern.
- Back-in-stock at variant level. Notifying someone that "the dress" is back when only XXL returned is how you train people to ignore your emails.
- Multi-location. This breaks the moment you add a second warehouse or start shipping from a retail store. Ask them directly how many multi-location fashion stores they've configured, and what they do about split shipments on a two-item order.
Returns and exchanges are your P&L, and most agencies won't touch them
Here's arithmetic you can run against your own numbers tonight. Say you ship 1,000 orders a month at an average order value of ₹2,200, and 18% come back. That's 180 returns. In our experience with Indian apparel brands, roughly half of those are fit, not fault, so 90 returns a month are pure sizing loss.
Reverse logistics at ₹110 per shipment plus warehouse handling and re-tagging at around ₹60 gives ₹170 a return. 90 × ₹170 = ₹15,300 a month, or about ₹1.84 lakh a year, spent moving clothes that never should have shipped in that size. That's before the refund processing fee and before the units you can't put back on the shelf.
Now: if a proper size guide, a "true to size / runs small" flag pulled from reviews, and an exchange-first returns flow cut fit returns by a quarter, you've saved about ₹46,000 a year and kept a chunk of the revenue as an exchange instead of a refund. That's a small project paying for itself. A partner who has never asked to see your returns reasons report can't have this conversation with you, and won't.
Ask what returns app they'd recommend and why. Ask whether they'd push exchanges over refunds by default, and how they'd handle a COD order where the money never really landed cleanly in the first place.
Search and filters on a deep catalogue
Fashion catalogues grow faster than any other category, because every season adds inventory nobody deletes. By the time you're at 800 live SKUs, the native Shopify filter set and the default search are the ceiling on your conversion rate. Shoppers who use on-site search convert well above the site average, and shoppers who search, get nothing useful and bounce are the most expensive traffic you buy.
What to check: does search handle misspellings and Hinglish ("kurti", "kurta", "kurtha"), does it search colour and fabric metafields, does filtering update without a full page reload, and are out-of-stock variants excluded from size filters. If your partner's answer to a large catalogue is "we'll add more collections", they're solving a merchandising problem with URLs. We built FilterPro partly because we kept hitting this on apparel projects and wanted filters and AI search that understood fashion attributes rather than generic tags.
Speed, and the fact that your best asset is also your heaviest
Fashion sells on imagery, so fashion stores are heavy by nature. A partner who tells you to use fewer photos has misunderstood the brief. The job is to serve 14 images on a product page and still hit an LCP under 1.5 seconds on a mid-range Android over 4G, which means responsive srcsets, correct image dimensions at upload, lazy loading below the fold, and a hard audit of every app script the marketing team has installed.
That last one is usually the biggest win. We routinely find six to nine apps loading JavaScript on every page, three of which nobody at the company remembers signing up for. Ask a prospective partner what their process is for auditing app scripts, and whether they'll actually remove things or just add a caching app on top. Our speed work starts with deletion, not addition.
The India-specific things a fashion partner has to know cold
COD is not a payment method here, it's a business model, and it changes the build. Prepaid discounts, partial COD advance collection via Razorpay, pincode-level COD availability, RTO risk scoring, address confirmation before dispatch: these are theme and app decisions, and a partner who's only built for US brands hasn't made any of them.
Tax is the other one. Apparel in India sits across two GST slabs split by price point, and the threshold changed in 2025, so a store selling a ₹1,200 top and a ₹6,000 jacket cannot run on one flat rate. Your partner should be setting up price-based tax overrides and checking them against what your CA tells you the current slabs are, not guessing. If you sell food, cosmetics or anything FSSAI-adjacent alongside apparel, that's another set of rules on the same store.
And festive peaks. A brand that does a normal ₹8 lakh month can do ₹40 lakh across a Diwali fortnight. Ask what they do differently for that: cart and checkout load testing, a freeze on theme deploys during the sale window, a rollback plan, and someone actually reachable at 11pm on the day the sale goes live. "We'll monitor it" is not a plan.
What it costs, and when Plus is a waste of money
A serious fashion build on a well-chosen premium theme, with proper variant architecture, size-chart metaobjects, filters, returns flow and a payment stack configured for COD, is not a ₹60,000 project. It's also not a ₹15 lakh project unless you're integrating an ERP and a WMS. We've written up the ranges in detail on what Shopify development actually costs in India, and the short version is that theme choice and integration count drive the number far more than page count does.
On Shopify Plus: if you're under roughly ₹4–5 crore a year, single market, single currency, and your main complaint is that you want a nicer homepage, Plus is the wrong purchase. What Plus buys a fashion brand is checkout customisation, Shopify Functions for complex discounting, B2B if you wholesale, and headroom for flash-sale traffic. Those matter when you're running drops that do a few thousand orders in twenty minutes. A partner who recommends Plus before asking about your peak-hour order concurrency is selling, not advising. If you have genuinely outgrown the standard plan, that conversation is worth having properly.
Red flags in the sales call
- They quote a fixed price before asking how many SKUs and variants you have.
- No question about your returns rate, or your top return reason.
- Their portfolio fashion stores all use the same theme with the colours changed.
- They can't tell you who does the work. A named developer and a way to reach them beats an account manager relaying messages. If you'd rather build the relationship directly, hiring a dedicated Shopify developer is often cleaner than a fixed-scope project.
- Post-launch support isn't in the proposal. Fashion stores change weekly. A build with no ongoing arrangement is a build you'll abandon by season two.
A cheaper option that is sometimes the right call
If you're pre-₹1 crore, running one warehouse and 150 SKUs, you probably don't need an agency retainer. Buy a good premium theme, pay a competent developer for two weeks to set up variants, metaobjects, size charts and the payment stack properly, and spend the rest on photography and ads. The build is not your constraint at that stage. Demand is. Agencies that won't say this to you are optimising for their pipeline.
The moment it flips is when catalogue size, returns volume or drop traffic starts costing you money that a build decision could recover. That's usually visible in the data before it's visible in your gut.
Next step
Pull three numbers before you talk to anyone: your fit-related return rate, your mobile collection-page bounce rate, and your COD share of orders. Take them into the first call and see whether the person across the table knows what to do with them. If you'd like us to look at your store first, our free audit covers exactly those three, and we'll tell you if the answer is "leave it alone and spend the money on marketing".