Marketplace search is not a mystery box. Meesho, Flipkart, Amazon, Myntra, AJIO and Nykaa run different code, but they optimise for the same outcome — the listing most likely to convert a search into a delivered, un-returned order — and they read the same five signals to find it.
Relevance. Whether your title, category and attributes match what the buyer typed. This is the cheapest thing on the list to fix and the most commonly left broken: a listing filed under the wrong category or missing its fabric, sleeve-length and occasion attributes cannot rank for the searches those attributes describe, no matter how good the product is.
Click-through rate. Of the buyers shown your listing in a grid, how many tap it. In fashion the grid is almost entirely image — a thumbnail, a price and a rating. Your first photo is doing nearly all of this work.
Conversion rate. Of the buyers who tap through, how many buy. This is price, reviews, size information and — heavily — whether the additional photos answer the questions the thumbnail raised: how it falls, how it looks from behind, what the fabric actually is.
Seller reliability. Dispatch inside the SLA, low cancellations, low returns. Marketplaces demote unreliable sellers regardless of product quality, because a cancelled order costs them a customer, not just a sale. Return rate is the one most sellers underestimate: a listing whose photos flatter the product converts well and then gets returned, and the engine reads that pattern faster than you do.
Paid amplification. Ads buy you the impressions to generate the signals above. They do not fix a listing that converts badly — running ads on a poor listing simply pays to prove it converts badly, faster.
The order matters. Relevance and reliability are hygiene: get them wrong and nothing else helps. Click-through and conversion are where fashion sellers actually win or lose, and both are decided by imagery long before they are decided by copy.
Same five signals, different weights. Knowing which weight dominates tells you where to spend your effort on each platform.
Meesho weights price band position and return rate unusually heavily, because its buyers filter on price and its reseller layer is calculating margin. Win it on conversion and RTO, not on further price cuts. Full guide
Flipkart weights sales velocity and ratings, which creates a cold-start trap for new listings. Full guide
Amazon India weights listing completeness more mechanically than the others, which is why unused image slots are the single largest unclaimed asset on most accounts. Full guide
Myntra is browse-led rather than search-led, so your photograph is the entire proposition rather than a supporting asset. Full guide
AJIO curates its seller base, trading a harder entry for far thinner competition inside each category. Full guide
Nykaa Fashion goes further still — premium positioning, selective onboarding, and a context where discounting works against you. Full guide
Shopify has no demand of its own. It is an arithmetic problem where conversion rate is the multiplier on every rupee of ad spend. Full guide
Etsy rewards specific long-tail listings and craft detail, which is a genuine structural advantage for Indian textile sellers. Full guide
Instagram rewards posting frequency, which makes content supply the real constraint for most small businesses. Full guide
WhatsApp has no algorithm at all — just a list, a catalogue, and whether your images look trustworthy enough on a small screen. Full guide
When a listing is not selling, sellers tend to reach for the lever they know — usually price or ads. Both are expensive guesses. The cheaper approach is to find out which stage is actually broken, because each stage has a different fix and applying the wrong one costs money without changing anything.
No impressions? That is relevance. The listing is in the wrong category, missing attributes, or titled in language nobody searches. Costs nothing but an afternoon to fix, and it is the most commonly skipped step.
Impressions but no clicks? That is your thumbnail. In a fashion grid the buyer sees an image, a price and a rating, and yours is losing to something visually stronger. No amount of description rewriting fixes this.
Clicks but no orders? That is conversion. Price, reviews, size information, and whether your remaining images answer the questions the thumbnail raised. This is where most fashion listings actually leak.
Orders but returns? That is accuracy. Size ambiguity and imagery that flatters rather than represents. Expensive twice over, because the return costs margin and the return rate costs rank.
Run this before spending on ads, on any platform. Advertising amplifies whichever of these is currently true.
Most listing advice treats photography as a presentation detail. On a marketplace it is a ranking input, and it is the only input that touches three of the five signals simultaneously.
A better first image raises click-through, because the search grid is a wall of thumbnails and yours is competing on visual quality alone. A complete set of secondary images raises conversion, because every unanswered question — how does this drape, what does the back look like, is this the same shade in daylight — is a reason to close the tab. And an accurate set lowers returns, because the most expensive return in fashion is the one where the buyer says the product looked different online.
That last point is where flat-lay-only catalogues quietly bleed money. A flat-lay is a truthful photo of a garment lying down, and buyers cannot read fit or fall from it, so they guess — and a meaningful share of them guess wrong. On-model imagery does not flatter the product, it removes the guess.
This is the honest case for AI catalogue photography rather than the promotional one. It is not that AI photos are magic; it is that most small sellers ship two flat-lays because a studio shoot costs ₹15,000–₹50,000 a day and cannot be justified per SKU. Generating a full on-model set from one flat-lay at ₹10 a photo removes the reason the catalogue was thin in the first place.