Myntra — subscription-classified SAMPLE (n=50, no API)

Wishlist→purchase discovery engine · last 12 months · generated 2026-08-26 05:32 · Myntra (subscription sample · n=50)

Business metric: % of users who purchase ≥1 wishlisted item within 30 days of adding it. Solution constraint: non-monetary (no discounts).

39Myntra (subscription sample · n=50) — relevant items (of 50 scraped)

Top non-monetary opportunity

NON-MONETARY: attack fit/size + return-confidence at the wishlist→purchase moment — a fit/size-certainty aid (size guidance + true-to-size signals) paired with a visible "easy, free exchange" reassurance. In the sample this covers sizing uncertainty (blocker+residual, ~3 items) plus return-hassle fear (4 items) — the largest addressable, non-price cluster. Price is the biggest raw blocker but is MONETARY and off-limits per the brief.

Opportunity areas — ranked & quantified

Each area is sized by how many relevant items map to it and its share of that brand's relevant corpus. Needs money? flags whether fixing it would require a monetary lever (a discount / price cut) — the brief bars monetary solutions, so “no” areas are the addressable ones.

Myntra (subscription sample · n=50) — 39 relevant items

Opportunity areaDimensionn% of relevantNeeds money?
delivery-concern purchase-blocker 5 12.8% non-monetary
return-hassle-fear purchase-blocker 4 10.3% non-monetary
price purchase-blocker 4 10.3% money
value-for-money residual-uncertainty 4 10.3% non-monetary
reliable, easy returns and exchanges unmet-need 3 7.7% non-monetary
dependable delivery and accountability unmet-need 3 7.7% non-monetary
competitive pricing and rewards unmet-need 3 7.7% money
fit-size residual-uncertainty 2 5.1% non-monetary
accurate sizing and fit assurance unmet-need 2 5.1% non-monetary
premium browsing and clear product imagery unmet-need 2 5.1% non-monetary
fit-size-uncertainty purchase-blocker 1 2.6% non-monetary
quality-material residual-uncertainty 1 2.6% non-monetary
color-accuracy residual-uncertainty 1 2.6% non-monetary

Largest addressable (non-monetary) area: delivery-concern — 5 items (12.8% of relevant).

“This app has no safety for customers' orders. I neither received my parcel nor was I contacted by the delivery person. But the status of my parcel shows 'Delivered'. I did not get any OTP and no one called me to get one, so how can this app show that the parcel was delivered? I have raised this query, and now they are ”
playstore · 2026-08-22 · 3★ · gp_06744991-be3c-4f23-a8cd-fe22b409bb26
“Delayed Delivery. Pathetic app, I ordered a bag worth 6.5k. Payment is done and all. I placed my order on 4th august. Its 21st August today, my bag has not arrived yet. The initial delivery was supposed to happen on 11th, then 13th then they raised a ticket on 14th asking until 18th, on 18th they asked for 48 hours and”
appstore · 2026-08-21 · 1★ · as_14454001996

The ten discovery questions

Myntra (subscription sample · n=50)

*** n = the number of real feedback items behind each row (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

Why do users add fashion products to their wishlist?

All 39 relevant items, each assigned one wishlist reason. Percentages are of 39, so the rows sum to the whole.

All 39 items appear in the rows below — each in a specific group, nothing dropped (rows sum to 39).

Themen%Detail
other 34 87.2%
price-watch-wait-sale 4 10.3%
gift-idea 1 2.6%
“They sell defective products with extra price. And returning an defective item is almost impossible. I had bought belt, which was defective and they were not ready to take it back. After that uninstall the app”
playstore · 2026-08-23 · 1★ · gp_376a7c76-a57b-4186-8d92-014a52b307fe
“I have never had such a bad experience buying a product. The product itself is good, but first of all, it was not delivered to my home or anywhere near my house, so I had to travel 4 km to collect it. Also, the T-shirt I received was much larger than a normal L size. I requested an exchange, and when I travelled 4 km a”
playstore · 2026-08-22 · 2★ · gp_07015f27-c191-486b-8759-30b46d07d8c3
“Delayed Delivery. Pathetic app, I ordered a bag worth 6.5k. Payment is done and all. I placed my order on 4th august. Its 21st August today, my bag has not arrived yet. The initial delivery was supposed to happen on 11th, then 13th then they raised a ticket on 14th asking until 18th, on 18th they asked for 48 hours and”
appstore · 2026-08-21 · 1★ · as_14454001996
Reviews used to validate this answer (39)
Page

*** n = the number of real feedback items behind each row (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

What prevents wishlisted products from being purchased?

All 14 relevant items that name a purchase blocker (of 39 relevant), grouped by the single strongest blocker.

All 14 items appear in the rows below — each in a specific group, nothing dropped (rows sum to 14).

Themen%Detail
delivery-concern 5 35.7%
return-hassle-fear 4 28.6%
price 4 28.6%
fit-size-uncertainty 1 7.1%
“This app has no safety for customers' orders. I neither received my parcel nor was I contacted by the delivery person. But the status of my parcel shows 'Delivered'. I did not get any OTP and no one called me to get one, so how can this app show that the parcel was delivered? I have raised this query, and now they are ”
playstore · 2026-08-22 · 3★ · gp_06744991-be3c-4f23-a8cd-fe22b409bb26
“Delayed Delivery. Pathetic app, I ordered a bag worth 6.5k. Payment is done and all. I placed my order on 4th august. Its 21st August today, my bag has not arrived yet. The initial delivery was supposed to happen on 11th, then 13th then they raised a ticket on 14th asking until 18th, on 18th they asked for 48 hours and”
appstore · 2026-08-21 · 1★ · as_14454001996
“Very Poor Delivery timelines. They are always delay in delivering items to locations in South. Ekartel people don’t even call or answer our calls”
appstore · 2026-08-22 · 2★ · as_14461074154
Reviews used to validate this answer (14)
Page

*** n = the number of real feedback items behind each row (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

What uncertainties remain after a user identifies a product they like?

All 8 relevant items that voice a residual uncertainty after liking a product, grouped by uncertainty.

All 8 items appear in the rows below — each in a specific group, nothing dropped (rows sum to 8).

Themen%Detail
value-for-money 4 50%
fit-size 2 25%
quality-material 1 12.5%
color-accuracy 1 12.5%
“very bad,iska saman bahut mahenga hai or mesho shopsy se dubble rate hai ,isse badhiya mesho shopsy hai”
playstore · 2026-08-21 · 1★ · gp_d4a04135-bfef-4ab6-baa6-0375227a5687
“worst app ever so expensive not good productss”
playstore · 2026-08-23 · 2★ · gp_564985dc-d630-4f3f-ac60-b00b485d5d19
“Good app to find good quality trend wears also can't find that mutch offers that's a bad point”
playstore · 2026-08-23 · 3★ · gp_1f14962f-58aa-4dad-aea9-8dda5e575157
Reviews used to validate this answer (8)
Page

*** n = the number of real feedback items behind each row (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

What causes users to postpone a purchase?

All 2 relevant items that state a reason for postponing the purchase, grouped by cause.

All 2 items appear in the rows below — each in a specific group, nothing dropped (rows sum to 2).

Themen%Detail
waiting-sale 2 100%
“very bad,iska saman bahut mahenga hai or mesho shopsy se dubble rate hai ,isse badhiya mesho shopsy hai”
playstore · 2026-08-21 · 1★ · gp_d4a04135-bfef-4ab6-baa6-0375227a5687
“Good app to find good quality trend wears also can't find that mutch offers that's a bad point”
playstore · 2026-08-23 · 3★ · gp_1f14962f-58aa-4dad-aea9-8dda5e575157
Reviews used to validate this answer (2)
Page

*** n = the number of real feedback items behind each row (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

How do users compare multiple shortlisted products?

All 2 relevant items that describe how they compare shortlisted products, grouped by comparison behaviour.

All 2 items appear in the rows below — each in a specific group, nothing dropped (rows sum to 2).

Themen%Detail
across-apps-compare 1 50%
in-app-compare 1 50%
“very bad,iska saman bahut mahenga hai or mesho shopsy se dubble rate hai ,isse badhiya mesho shopsy hai”
playstore · 2026-08-21 · 1★ · gp_d4a04135-bfef-4ab6-baa6-0375227a5687
“Used to be great. Now its just stupid. This was probably my favourite app for shopping. Now its just bad in terms of user experience. The product image quality are quite low resolution. There are stupid sponsored ads shoved so aggressively that its almost stupid. There stupid buttons here and there (why do i need a cli”
appstore · 2026-08-23 · 3★ · as_14461824398
Reviews used to validate this answer (2)
Page

*** n = the number of real feedback items behind each row (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

What information do users seek outside the platform before purchasing?

All 0 relevant items that mention seeking information outside the app before buying, grouped by source.

No rows — this question had no usable signal in this corpus.

*** n = the number of real feedback items behind each row (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

What role do fit, size, styling, price, reviews, occasion and social validation play?

All 28 relevant items that name at least one factor (of 39 relevant). A single item can touch several factors, so the percentages are shares of 28 and sum to more than 100%.

All 28 items appear in the rows below — each in a specific group, nothing dropped (rows sum to 46).

Themen%Detail
quality 13 46.4%
delivery-returns 13 46.4%
price 9 32.1%
size 4 14.3%
fit 3 10.7%
occasion 1 3.6%
styling 1 3.6%
reviews 1 3.6%
brand-trust 1 3.6%
“They sell defective products with extra price. And returning an defective item is almost impossible. I had bought belt, which was defective and they were not ready to take it back. After that uninstall the app”
playstore · 2026-08-23 · 1★ · gp_376a7c76-a57b-4186-8d92-014a52b307fe
“Used to be great. Now its just stupid. This was probably my favourite app for shopping. Now its just bad in terms of user experience. The product image quality are quite low resolution. There are stupid sponsored ads shoved so aggressively that its almost stupid. There stupid buttons here and there (why do i need a cli”
appstore · 2026-08-23 · 3★ · as_14461824398
“worst app ever so expensive not good productss”
playstore · 2026-08-23 · 2★ · gp_564985dc-d630-4f3f-ac60-b00b485d5d19
Reviews used to validate this answer (28)
Page

*** n = the number of real feedback items behind each row (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

When is the wishlist genuine purchase intent vs a bookmarking mechanism?

All 39 relevant items, each assigned a wishlist nature. Percentages are of 39.

All 39 items appear in the rows below — each in a specific group, nothing dropped (rows sum to 39).

Themen%Detail
unclear 21 53.8%
genuine-intent 16 41%
price-watch 2 5.1%
“They sell defective products with extra price. And returning an defective item is almost impossible. I had bought belt, which was defective and they were not ready to take it back. After that uninstall the app”
playstore · 2026-08-23 · 1★ · gp_376a7c76-a57b-4186-8d92-014a52b307fe
“Size issues. Clothesline is good but order processing is badly handled.Size issues are mostly there forcing you to exchange”
appstore · 2026-08-21 · 2★ · as_14454508909
“I am facing a lot with my exchange order, it's getting frustrated now. how much I have to wait. once they attempt n the delivery agent didn't receive my order as it was his first day don't know much. Calling myntra customer care service is also not helping much ..!”
playstore · 2026-08-25 · 2★ · gp_98c2b13d-1402-4d59-9f71-ee69f0c36a08
Reviews used to validate this answer (39)
Page

*** n = the number of real feedback items behind each row (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

How do these behaviours differ across user segments?

All 39 relevant items grouped by shopper segment (segments with <3 items are pooled). % = the segment's share of the relevant corpus; the note gives that segment's top purchase blocker.

All 39 items appear below (rows sum to 39): 79% fall into a named theme, the remaining 21% are grouped in the “other / not stated” row so nothing is dropped.

Themen%Detail
general shopper 21 53.8% top blocker: delivery-concern (1) · 9 genuine-intent
frequent returner 5 12.8% top blocker: return-hassle-fear (4) · 1 genuine-intent
deal / offer seeker 5 12.8% top blocker: price (3) · 2 genuine-intent
smaller / unlabelled segments 8 20.5% accounted for, but with no single dominant theme
“delivery parter are very bad”
playstore · 2026-08-22 · 1★ · gp_9643705c-2abb-40dd-ad1f-b088ecc1f93b
“the quick exchange experience made it give it 5 STAR. Thank you for providing such great PRODUCT and Service.”
playstore · 2026-08-23 · 5★ · gp_007b70ff-5523-40c3-985b-0d6948be2cab
“"Myntra is my absolute go-to app for fashion and lifestyle shopping! The user interface is very smooth, search filters are highly accurate, and the sizing guides are mostly perfect. Delivery is usually on time, and refunds are processed swiftly. Highly recommended!"”
playstore · 2026-08-21 · 5★ · gp_51aba27e-3f91-4d42-a157-a32b94d53713
Reviews used to validate this answer (39)
Page

*** n = the number of real feedback items behind each row (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

What unmet needs emerge consistently across conversations?

All 13 relevant items that state an unmet need, grouped by theme. Themes in 2+ independent sources are the durable ones.

All 13 items appear in the rows below — each in a specific group, nothing dropped (rows sum to 13).

Themen%Detail
reliable, easy returns and exchanges 3 23.1% seen in 2 of 4 sources
dependable delivery and accountability 3 23.1% seen in 2 of 4 sources
accurate sizing and fit assurance 2 15.4% seen in 2 of 4 sources
competitive pricing and rewards 3 23.1% seen in 1 of 4 sources
premium browsing and clear product imagery 2 15.4% seen in 2 of 4 sources
“They sell defective products with extra price. And returning an defective item is almost impossible. I had bought belt, which was defective and they were not ready to take it back. After that uninstall the app”
playstore · 2026-08-23 · 1★ · gp_376a7c76-a57b-4186-8d92-014a52b307fe
“I am facing a lot with my exchange order, it's getting frustrated now. how much I have to wait. once they attempt n the delivery agent didn't receive my order as it was his first day don't know much. Calling myntra customer care service is also not helping much ..!”
playstore · 2026-08-25 · 2★ · gp_98c2b13d-1402-4d59-9f71-ee69f0c36a08
“Refund not received. I am disappointed with Myntra refund don’t buy with this app”
appstore · 2026-08-22 · 1★ · as_14459333595
Reviews used to validate this answer (13)
Page

Synthesis — written from the aggregates only, every claim cited

*** n = the number of real feedback items this claim rests on (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

Why do users add fashion products to their wishlist?

SAMPLE (n=50): wishlisting intent is mostly genuine, with a clear price-watch minority

n=4 — Buyers who saved/added items mostly show genuine purchase intent; a distinct minority hold off to wait for offers or a better price (price-watch).
“very bad,iska saman bahut mahenga hai or mesho shopsy se dubble rate hai ,isse badhiya mesho shopsy hai”
playstore · 2026-08-21 · gp_d4a04135-bfef-4ab6-baa6-0375227a5687
“Good app to find good quality trend wears also can't find that mutch offers that's a bad point”
playstore · 2026-08-23 · gp_1f14962f-58aa-4dad-aea9-8dda5e575157
n=1 — Occasion/gift buying appears explicitly, tying the save→buy window to a date (a birthday), which makes delivery reliability decisive.
“This app has no safety for customers' orders. I neither received my parcel nor was I contacted by the delivery person. But the status of my parcel shows 'Delivered'. I did not get any OTP and no one called me to get one, so how can this app show that the parcel was delivered? I have raised this query, and now they are ”
playstore · 2026-08-22 · gp_06744991-be3c-4f23-a8cd-fe22b409bb26

*** n = the number of real feedback items this claim rests on (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

What prevents wishlisted products from being purchased?

What blocks purchase: delivery reliability, return-hassle fear, price, and sizing

n=5 — Delivery reliability is the single most-named blocker in the sample (late, undelivered, or falsely "delivered" orders).
“Delayed Delivery. Pathetic app, I ordered a bag worth 6.5k. Payment is done and all. I placed my order on 4th august. Its 21st August today, my bag has not arrived yet. The initial delivery was supposed to happen on 11th, then 13th then they raised a ticket on 14th asking until 18th, on 18th they asked for 48 hours and”
appstore · 2026-08-21 · as_14454001996
“This app has no safety for customers' orders. I neither received my parcel nor was I contacted by the delivery person. But the status of my parcel shows 'Delivered'. I did not get any OTP and no one called me to get one, so how can this app show that the parcel was delivered? I have raised this query, and now they are ”
playstore · 2026-08-22 · gp_06744991-be3c-4f23-a8cd-fe22b409bb26
“Very Poor Delivery timelines. They are always delay in delivering items to locations in South. Ekartel people don’t even call or answer our calls”
appstore · 2026-08-22 · as_14461074154
n=4 — Fear of a hard return/exchange is the next-strongest blocker — defective items, refused returns, unprocessed refunds.
“They sell defective products with extra price. And returning an defective item is almost impossible. I had bought belt, which was defective and they were not ready to take it back. After that uninstall the app”
playstore · 2026-08-23 · gp_376a7c76-a57b-4186-8d92-014a52b307fe
“I have never had such a bad experience buying a product. The product itself is good, but first of all, it was not delivered to my home or anywhere near my house, so I had to travel 4 km to collect it. Also, the T-shirt I received was much larger than a normal L size. I requested an exchange, and when I travelled 4 km a”
playstore · 2026-08-22 · gp_07015f27-c191-486b-8759-30b46d07d8c3
“Refund not received. I am disappointed with Myntra refund don’t buy with this app”
appstore · 2026-08-22 · as_14459333595
n=5 — Price (vs Meesho/Shopsy and fewer offers) is a real but MONETARY blocker; sizing uncertainty also blocks purchase directly.
“very bad,iska saman bahut mahenga hai or mesho shopsy se dubble rate hai ,isse badhiya mesho shopsy hai”
playstore · 2026-08-21 · gp_d4a04135-bfef-4ab6-baa6-0375227a5687
“Size issues. Clothesline is good but order processing is badly handled.Size issues are mostly there forcing you to exchange”
appstore · 2026-08-21 · as_14454508909

*** n = the number of real feedback items this claim rests on (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

What uncertainties remain after a user identifies a product they like?

Residual uncertainty is dominated by value-for-money and fit/size

n=6 — After liking an item, the doubts that linger are whether it is worth the price and whether the size/fit will be right.
“Size issues. Clothesline is good but order processing is badly handled.Size issues are mostly there forcing you to exchange”
appstore · 2026-08-21 · as_14454508909
“I have never had such a bad experience buying a product. The product itself is good, but first of all, it was not delivered to my home or anywhere near my house, so I had to travel 4 km to collect it. Also, the T-shirt I received was much larger than a normal L size. I requested an exchange, and when I travelled 4 km a”
playstore · 2026-08-22 · gp_07015f27-c191-486b-8759-30b46d07d8c3
n=1 — Low-resolution product images add colour/quality uncertainty that a shopper cannot resolve on-platform.
“Used to be great. Now its just stupid. This was probably my favourite app for shopping. Now its just bad in terms of user experience. The product image quality are quite low resolution. There are stupid sponsored ads shoved so aggressively that its almost stupid. There stupid buttons here and there (why do i need a cli”
appstore · 2026-08-23 · as_14461824398

*** n = the number of real feedback items this claim rests on (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

What causes users to postpone a purchase?

Postponement is mostly waiting for a sale/offer (small n)

n=2 — The clearest postponement signal is waiting for a sale or better offer; sample n is small, so read as directional.
“very bad,iska saman bahut mahenga hai or mesho shopsy se dubble rate hai ,isse badhiya mesho shopsy hai”
playstore · 2026-08-21 · gp_d4a04135-bfef-4ab6-baa6-0375227a5687
“Good app to find good quality trend wears also can't find that mutch offers that's a bad point”
playstore · 2026-08-23 · gp_1f14962f-58aa-4dad-aea9-8dda5e575157

*** n = the number of real feedback items this claim rests on (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

How do users compare multiple shortlisted products?

Comparison is cross-app (price) and, when blocked, in-app by brand

n=2 — Shoppers compare Myntra prices against other apps (Meesho/Shopsy); one power user wants to browse a whole brand in-app but cannot.
“very bad,iska saman bahut mahenga hai or mesho shopsy se dubble rate hai ,isse badhiya mesho shopsy hai”
playstore · 2026-08-21 · gp_d4a04135-bfef-4ab6-baa6-0375227a5687
“Used to be great. Now its just stupid. This was probably my favourite app for shopping. Now its just bad in terms of user experience. The product image quality are quite low resolution. There are stupid sponsored ads shoved so aggressively that its almost stupid. There stupid buttons here and there (why do i need a cli”
appstore · 2026-08-23 · as_14461824398

*** n = the number of real feedback items this claim rests on (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

What information do users seek outside the platform before purchasing?

Off-platform info-seeking is thin in this store-review sample

n=0 — Store reviews rarely state where users look before buying; off-platform info-seeking (YouTube/Instagram/Google) is not evidenced here and needs a community/social source to measure.

*** n = the number of real feedback items this claim rests on (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

What role do fit, size, styling, price, reviews, occasion and social validation play?

Delivery/returns, price, quality, and size are the factors that dominate

n=5 — Delivery-returns is the most frequently mentioned factor, followed by price and quality; size/fit recurs wherever apparel is discussed.
“Delayed Delivery. Pathetic app, I ordered a bag worth 6.5k. Payment is done and all. I placed my order on 4th august. Its 21st August today, my bag has not arrived yet. The initial delivery was supposed to happen on 11th, then 13th then they raised a ticket on 14th asking until 18th, on 18th they asked for 48 hours and”
appstore · 2026-08-21 · as_14454001996
“Size issues. Clothesline is good but order processing is badly handled.Size issues are mostly there forcing you to exchange”
appstore · 2026-08-21 · as_14454508909
“I have never had such a bad experience buying a product. The product itself is good, but first of all, it was not delivered to my home or anywhere near my house, so I had to travel 4 km to collect it. Also, the T-shirt I received was much larger than a normal L size. I requested an exchange, and when I travelled 4 km a”
playstore · 2026-08-22 · gp_07015f27-c191-486b-8759-30b46d07d8c3

*** n = the number of real feedback items this claim rests on (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

When is the wishlist genuine purchase intent vs a bookmarking mechanism?

Wishlist is genuine intent for buyers; a price-watch subset treats it as "wait"

n=4 — Most decisive items read as genuine intent; a price-watch subset uses saving as "buy later when cheaper".
“This app has no safety for customers' orders. I neither received my parcel nor was I contacted by the delivery person. But the status of my parcel shows 'Delivered'. I did not get any OTP and no one called me to get one, so how can this app show that the parcel was delivered? I have raised this query, and now they are ”
playstore · 2026-08-22 · gp_06744991-be3c-4f23-a8cd-fe22b409bb26
“very bad,iska saman bahut mahenga hai or mesho shopsy se dubble rate hai ,isse badhiya mesho shopsy hai”
playstore · 2026-08-21 · gp_d4a04135-bfef-4ab6-baa6-0375227a5687

*** n = the number of real feedback items this claim rests on (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

How do these behaviours differ across user segments?

Segments: frequent returners, deal-seekers, occasion/premium, tier-2

n=5 — Frequent returners carry the return/size friction; deal-seekers carry the price signal; occasion/premium buyers make delivery reliability existential; a tier-2/South shopper flags regional delivery gaps.
“Size issues. Clothesline is good but order processing is badly handled.Size issues are mostly there forcing you to exchange”
appstore · 2026-08-21 · as_14454508909
“very bad,iska saman bahut mahenga hai or mesho shopsy se dubble rate hai ,isse badhiya mesho shopsy hai”
playstore · 2026-08-21 · gp_d4a04135-bfef-4ab6-baa6-0375227a5687
“Very Poor Delivery timelines. They are always delay in delivering items to locations in South. Ekartel people don’t even call or answer our calls”
appstore · 2026-08-22 · as_14461074154

*** n = the number of real feedback items this claim rests on (the evidence count). Higher n = stronger; n=1–2 is a weak anecdote. ***

What unmet needs emerge consistently across conversations?

Consistent unmet needs: return confidence, delivery trust, sizing certainty

n=8 — The needs that recur across sources are: easy/reliable returns, dependable delivery with accountability, and sizing certainty — all NON-MONETARY confidence levers.
“They sell defective products with extra price. And returning an defective item is almost impossible. I had bought belt, which was defective and they were not ready to take it back. After that uninstall the app”
playstore · 2026-08-23 · gp_376a7c76-a57b-4186-8d92-014a52b307fe
“Delayed Delivery. Pathetic app, I ordered a bag worth 6.5k. Payment is done and all. I placed my order on 4th august. Its 21st August today, my bag has not arrived yet. The initial delivery was supposed to happen on 11th, then 13th then they raised a ticket on 14th asking until 18th, on 18th they asked for 48 hours and”
appstore · 2026-08-21 · as_14454001996
“Size issues. Clothesline is good but order processing is badly handled.Size issues are mostly there forcing you to exchange”
appstore · 2026-08-21 · as_14454508909

Cross-product

Single brand in this sample (Myntra). Cross-brand comparison requires the AJIO and Nykaa Fashion runs, which are pending API credits.

Analysis quality (how we know the insights hold)

ProductCoverageRelevance yield Avg confidenceConsistency Human agreementTriangulationCitation validity
Myntra (subscription sample · n=50) 100% 78% 66% 70% 97%

Coverage = items that classified into a valid object · Relevance yield = share that is substantive shopping feedback · Consistency = label stability when a sample is re-classified · Human agreement = model vs a hand-labelled gold set (npm run label) · Triangulation = top themes corroborated by ≥2 independent sources · Citation validity = synthesized claims whose quote ids resolve to real rows.

Source coverage

Myntra (subscription sample · n=50) playstore: 27appstore: 13youtube: 10

Low or zero counts reflect real source availability and are reported as-is.

Limitations

SMALL-N SAMPLE (n=50), hand-classified by Claude Code subscription reasoning (no Anthropic API) because API credits are exhausted — NOT the full 1704-item cleaned corpus. Source skew: store reviews are largely post-purchase/app complaints, so wishlist reasons and off-platform info-seeking are under-observed; YouTube comments were off-topic; Reddit/communities were not collected (403). Counts are directional, not statistically representative. Re-run the full corpus across all 3 brands via the API for representative figures.

Method — the four demonstrations

1. How this workflow gathers and analyses data

Every stage below ran for this report. The numbers are this run's, not an illustration.

StageWhat happensMyntra (subscription sample · n=50)
1 · ResolveCompany name → verified store ids (never hard-coded)1 brand
2 · CollectPlay Store · App Store · YouTube, into one schema2265
3 · CleanNormalise, then drop empty / too-short / spam / bot / duplicate1704
4 · ExtractOne structured AI call per item → 16 fields, schema-enforced50
5 · ClusterFree-text fields → labelled themes carrying their member ids27 themes
6 · AggregatePlain counting — the 10 questions + money-tagged opportunities, no model opinion39 relevant
7 · SynthesiseNarrative written from the aggregates only10 answers

Cleaning — what was removed, and why

RuleDroppedRuleDropped
tooShort463 exactDuplicate54
nearDuplicate33 empty6
spam4 noLetters1

2265 collected → 1704 kept (75%), 561 dropped. Normalised: repeatedPunctuation: 113emoji: 915usernames: 4stretchedWords: 6urls: 3 Every dropped row is kept in dropped.jsonl with the rule that removed it.

2. How themes are identified

Anything we count uses a fixed vocabulary, so counts are reliable. Anything that needs nuance is free text and is clustered — never matched by string, which would produce a list of count-of-1 "themes". Each cluster keeps the ids of the items it was built from, so the taxonomy is derived from this data rather than chosen in advance.

FieldMethodRaw linesThemesLargest theme
blocker_notefree text → clustered → label-merged 155 returns and exchanges are hard, slow or refused (4)
info_neededfree text → clustered → label-merged 32 accurate size and fit before ordering (2)
segmentfree text → clustered → label-merged 389 general shopper (21)
painfree text → clustered → label-merged 206 return and exchange ordeal (4)
unmet_needfree text → clustered → label-merged 135 reliable, easy returns and exchanges (3)
wishlist_reason · purchase_blocker · residual_uncertainty · factors · wishlist_nature · sentimentfixed enum — chosen by the model from a closed listcounted directly
3. How insights are generated

The narrative model never sees the review corpus. It receives only the aggregates above — counts, percentages and quote ids — and must return each claim with the evidence count it rests on and the quote ids it drew from. Those ids are then checked against real rows; the citation validity figure in the quality panel is that check.

GuaranteeHow it is enforced
No claim without a numbern is a required field in the output schema
No claim without evidencequote_ids required; rendered inline above
No invented quotesIds validated against the corpus → 97% valid
No reading of raw reviewsThe synthesis payload excludes the corpus by construction
4. How the quality of these insights was validated

Seven independent checks. Figures are in the Analysis quality panel above.

CheckQuestion it answersThis run
CoverageDid every item classify into a valid object?100%
Relevance yieldHow much of the corpus is substantive shopping feedback?78%
Model confidenceHow sure was the model of its own labels?66%
Re-run consistencyDoes the same item get the same label twice?
Human agreementDoes the model agree with a hand-labelled gold set?not yet run
Source triangulationDo independent communities surface the same theme?70%
Citation validityDid the narrative invent any evidence?97%

All reviews analysed

Myntra (subscription sample · n=50) — 50 items collected from Play · App Store · YouTube, newest first
Show per page Page