Amazon contributes 62% of combined net realization but retains ~30% as fees before the brand sees it. D2C keeps ~94% of gross but carries discount cost (~5.7%) and RTO risk (~19.1% of orders). The brands that win are the ones actively growing D2C share while using Amazon as an acquisition channel.

Both channels contribute meaningfully but Amazon GMV is ~2x D2C net. The gap narrows on net realization once Amazon’s ~30% fee is stripped out.

Amazon accounts for 62% of combined net realization; D2C 38%. Growing D2C share is a structural margin play — every rupee shifted keeps ~97% of its value vs ~70% on Amazon.

Amazon AOV (ex-GST) ~₹1,783 vs D2C gross AOV ~₹1,695 — a real ₹88 gap on a like-for-like basis. The true gap is modest; D2C customers buy slightly smaller baskets.

The honest cost comparison: Amazon charges ~29.6% as marketplace fees. D2C costs are discounts (~5.7% of gross) + RTO (~19.1% of orders) + gateway fees (~2.7%, not yet built in). D2C keeps more per transaction but carries more operational risk.

COD is dominant on both channels: 46% on Amazon, 50% on D2C. COD on D2C drives the ~19% RTO rate. Prepaid nudges on the website are the highest-leverage short-term action.

Across 13 weeks, effective take-rate holds near 29.6% while weekly GMV slid -48.0% from Feb to late Apr. Net realization is the number to defend; the decline is volume-led, not pricing.

Orders (bars, left axis) and units (line, right axis) move in near-perfect lockstep — confirming most orders are single-unit. The basket-size opportunity is visible: multi-unit orders would expand units faster than orders.

AOV (ex-GST) has been strikingly stable near ₹1,780/wk — pricing and product mix are consistent. Rules out discounting as the GMV-decline cause; it is a volume/traffic problem.

Cumulative curve flattening in later weeks visualizes the deceleration — each week adds less than the last.

Effective take-rate (line, right) held flat at ~29-30% all quarter while net realization (bars) fell with GMV. Fee compression isn't available at this scale; volume is the only lever for net realization.

Selling fees (referral + closing) dominate the fee stack, FBA fees a distant second. Referral-rate category is where to negotiate or re-tier.

Total weekly fees shrink with GMV but the proportion stays fixed — no fee efficiency is gained at lower volume.

Revenue is heavily concentrated in a handful of SKUs. Concentration is both a strength and a risk.

A small fraction of SKUs drive the majority of revenue. Inventory and ad spend should follow this curve.

Top-5 SKU weekly mix shows whether the GMV decline is broad or concentrated — useful for triage.

Take-rate spread across SKUs reveals where pack-size or pricing changes recover margin.

Revenue clusters in one or two HSN codes, confirming a focused catalog — relevant for GST treatment and category strategy.

Revenue bars (left) show where customers are; take-rate line (right) shows that Amazon's fee structure is geographically uniform — Karnataka earns the same margin % as Tamil Nadu. Concentration, not margin variance, is the geographic story.

Bengaluru, Hyderabad, Chennai lead. Useful for targeted regional ad spend and same-city delivery-speed expectations.

A striking 82.8% of sales ship interstate (IGST). The single Karnataka warehouse serves the whole country — a second regional warehouse could cut transit time and freight cost.

Stacked bars show absolute COD vs prepaid order counts; the line shows COD share %. A rising COD share forecasts rising RTO risk. The share has been persistently high (~46%) — nudging prepaid improves both cash cycle and return rate.

COD orders carry a slightly higher take-rate (30.4%) than prepaid (29.0%) — Amazon charges more on COD. Combined with higher return risk, the prepaid rupee is more profitable.

13,807 Seller Flex orders dispatch from the LOT warehouse (VVVG, coded as FBA by Amazon) — 81.3% of GMV. True FBA is only 18.2% (3,172 orders), Merchant (MFN) is a sliver at 0.5%. Treat Seller Flex and True FBA as distinct fee/economics buckets going forward.

B2C dominates; B2B is a thin sliver. B2B is an untapped lane if the brand wants it.

GST collected tracks GMV at a flat 18% slab — useful for reconciliation and GST filing sanity checks.

Joining each order to its own fees reveals take-rate is not flat — it ranges from ~24% to ~40% per order. The headline ~29.6% average hides real variance.

Margin-drag index = each SKU’s share of total fees minus its share of revenue. Positive means the SKU consumes more fees than its revenue warrants.

Beyond order fees, refunds and service fees are the biggest negative events. Reimbursements partially offset — audit all eligible claims.

D2C net sales (bars) and order volume (line) on one view. Unlike Amazon, D2C volume is spiky — campaign-driven peaks rather than steady marketplace flow. Each order here builds first-party customer data Amazon orders never capture.

Discounts are real on D2C — 5.7% of gross sales. Tracking discount-to-revenue yield is important as this scales.

D2C SKU mix may differ from Amazon — gaps could indicate listing or pricing mismatches.

D2C geography vs Amazon reveals whether the owned channel reaches the same states or has a different customer base.

MTR includes 18% GST; settlement product sales is ex-GST principal. After stripping tax the two sources reconcile to within 0.9% — the residual is timing from 1,781 prior-period orders.

17,059 orders match across MTR and settlement (99.98%). The 1,781 settlement-only orders are January shipments that settled within Feb-Apr — a timing difference, not a data problem.

MTR refund value (₹78.2L) vs settlement refund (₹66.3L) — the ₹11.9L gap is the 18% GST component. Both figures are correct for their respective purposes.

The full settlement event waterfall. Reimbursements, SAFE-T claims, and fulfilment fee refunds are positive offsets worth auditing to ensure all eligible claims are filed.

Shopify D2C is split ~50/50 COD vs Razorpay prepaid, with a small 2.6% “Other” bucket (Draft Orders, C2P, ZOP) flowing through separate payment flows not captured in the Razorpay gateway export. COD orders have zero gateway economics — they settle in cash at delivery or are lost on RTO.

The Razorpay gateway export covers only 1,217 payments (₹23.1L) against 4,524 Shopify prepaid orders (₹83.8L) — a ₹60.6L gap explained almost entirely by incomplete Feb/Mar exports (only 27 and 80 payments vs thousands expected). April alone reconciles to ₹22.3L gateway vs ₹29.7L Shopify prepaid — the remaining ₹7.4L gap is C2P/ZOP orders going through a different payment flow. Gateway fee rate is a consistent 2.69% on all captured payments.