Settlement Reconciliation: Are Marketplaces Actually Paying You Right?

Most brands treat the marketplace settlement report the way they treat a salary slip. The money lands, the number looks roughly right, and nobody reads the line items. This is exactly the assumption the platforms are built on. A settlement is not a payment. It is a calculation, run by software you do not control, against a fee schedule that changes without notice, applied to thousands of transactions you never audit. Some of those calculations are wrong. Some sales never get paid out at all. And because the errors are small per order and buried in a spreadsheet with forty columns, they leak margin quietly for months before anyone notices. The question is not whether your marketplaces make settlement errors. They do. The question is whether you are catching them or eating them.

A settlement is a claim, not a fact

When Amazon or Flipkart pays you, what arrives in your account is the net of a long subtraction. Gross sale value, minus commission, minus fulfilment fees, minus closing fees, minus shipping, minus payment gateway charges, minus returns, minus reserves held against future returns, minus tax deductions. Each of those is a separate computation with its own rate, its own rounding, and its own room for error. You are handed the final number and asked to trust the working.

The trouble is that the working is opaque and the rates are not static. Commission percentages vary by category and sometimes by price band. Fee structures get revised. Promotional fee waivers are supposed to apply and sometimes silently do not. Weight-based shipping charges depend on a dimensional weight the platform measured, not the one you declared. Every one of these is a place where the number you were paid can drift from the number you were owed. Treating the settlement as a fact rather than a claim you should verify is the original mistake.

Where the money actually leaks

After enough reconciliations you start to see the same failure modes repeat across platforms. The leaks are rarely dramatic. They are small, systematic, and they compound across volume.

Commission charged at the wrong rate. Your SKU is categorised one way for the listing and another way for the fee engine, so you are billed a higher commission slab than your contract specifies. On a few hundred orders a month this is real money.
Fees on cancelled or returned orders that were never reversed. The order came back, the customer was refunded, but the commission or shipping fee on the original sale was never credited back to you.
Settlements that simply never arrive. A batch of orders marked delivered, eligible for payout, and then absent from every settlement cycle that follows. Not delayed. Missing.
Shipping charged on the wrong weight. The platform’s measured dimensional weight bumps you into a higher slab than your actual product, and you pay the difference on every unit until someone disputes it.
Reserves held longer than policy. Money parked against potential returns that should have released weeks ago and quietly did not.
Promotion and ad-fee double counting. A deal-day fee deducted at settlement that was supposed to be covered by the campaign you already funded.

None of these will bankrupt you in a month. That is precisely why they survive. A single order shorted by twelve rupees is beneath notice. The same error on every order in a category, every month, for a year, is a number that would change how you feel about that category.

Marketplace fee errors are not loud enough to notice and not large enough to chase, which is exactly why they are worth chasing.

Why nobody catches it

Reconciliation is unglamorous, manual, and easy to defer. The settlement file is enormous, the order export lives in a different report with different identifiers, and matching one to the other by hand is the kind of task that gets pushed to next week forever. So most brands never do it. They look at the headline payout, compare it loosely to last month, and move on.

This is understandable and expensive. The platforms are not run by villains running a scam. They are running imperfect software at enormous scale, and imperfect software at scale produces a steady drip of errors that always, by the structure of the situation, favour the party doing the calculating. There is no malice required. There is only the fact that an unaudited counterparty computing your revenue will, on average, compute it in its own favour more often than yours. The only correction is to check.

What real reconciliation looks like

Doing this properly means rebuilding the settlement from your own data and comparing it to theirs, line by line. You take your order export, you apply the fee rates your contract actually specifies, you compute what each order should have netted, and you match that against what the settlement report says it did net. The gaps are your evidence.

The matching key matters. Reconcile at the order or transaction level, not the summary level, because a summary that nets out to roughly the right total can hide an overcharge in one place cancelling an undercharge in another. You want to know that every single delivered order produced a settlement line, that every fee on that line is at the contracted rate, and that every return reversed the fees it should have. Orders that appear in your sales data but never appear in any settlement are the highest-value find, because a missing payout is one hundred percent leakage, not a few percent.

This is fundamentally a data exercise, which is why it belongs in the same system as the rest of your marketplace numbers. If you are already running a real reporting dashboard leadership actually reads, settlement variance is one more panel on it, computed automatically instead of chased manually every quarter.

It is connected to numbers you already care about

Reconciliation is not a side quest. It feeds directly into the decisions you make with the rest of your data. The fees you actually paid, as opposed to the fees you assumed, are an input to profitability per SKU. If your true commission is higher than the slab you modelled, the SKU you believe is your margin hero may be quietly underwater, and you would never know because you costed it on the wrong fee.

It connects to liquidity too. A settlement that arrives late, or a reserve held past policy, is cash trapped exactly where you cannot afford it. We have argued at length that working capital is the real constraint on marketplace growth, and unreconciled settlements make that constraint worse in two directions at once. You are short the money that was miscalculated, and you are blind to the timing of the money that was merely delayed.

And it overlaps with tax. The deductions on your settlement include TCS and TDS, and those have to reconcile not just against your expectation but against what gets reported to the authorities in your name. The mechanics are their own discipline, which we cover in the TCS, TDS and reconciliation nightmare, but the principle is the same. If you are not checking what was deducted, you are trusting a number that has compliance consequences attached to it.

What changed recently

The last year handed marketplace operators two reasons to rebuild their reconciliation baseline rather than coast on last year’s assumptions, and both are the kind of thing that quietly breaks a model nobody updated.

The first is fees. From March 16, Amazon India overhauled its seller fee schedule, expanding zero referral fees to over 12.5 crore products priced under 1,000 rupees across more than 1,800 categories, cutting referral fees on several high-volume categories above that band, and reducing Easy Ship and closing fees for low-price items, with the closing fee on sub-300-rupee products dropping from 45 rupees to 20 according to Amazon India. This is good news for margin, but it is a reconciliation trap. A fee cut only reaches your bank account if the fee engine actually applies the new rate to your category, and the single most common settlement error we find is a SKU that keeps getting billed at the old slab after a schedule change. The brands that benefit are the ones reconciling every payout against the current fee table, not the table they memorised eighteen months ago. A lower headline fee you are still being charged the old rate on is not a saving, it is a dispute waiting to be filed.

The second is tax. The GST 2.0 reform that took effect on 22 September 2025 collapsed the old four-slab structure into mainly 5 and 18 percent bands, which forced a wave of repricing and changed the tax math sitting inside every settlement line, as Unicommerce details. When the GST on a product moves, the gross-to-net working on your settlement moves with it, and any reconciliation rebuilt on pre-reform rates will throw false variances or, worse, hide real ones. The deadline pressure is real too. From December 2025, GST returns more than three years past their due date can no longer be filed, per Cashfree, which means a settlement mismatch you ignore today can become a permanently unrecoverable tax position. Reconciliation used to be about clawing back fees. It is now also about not letting a tax window close on money the platform already deducted in your name.

Make it a process, not a panic

The brands that recover this money do not do a heroic one-time audit and then stop. They turn reconciliation into a monthly cadence that runs against every payout, flags variances above a threshold, and produces a clean list of disputes to file with the platform. The recovery is real. Filing a well-evidenced fee dispute, with order IDs and the contracted rate attached, is usually paid out, because the platform’s own data confirms the error once you point at it. What you cannot do is dispute what you never measured.

This is the kind of work our Analytics & Reporting practice is built for, because it is pure data plumbing with a direct rupee return. Pulling settlement and order exports across platforms, normalising them to a common ledger, computing expected versus actual at the transaction level, and surfacing the gaps. From there our Marketplace Account Management team turns the variances into filed disputes and chases the recoveries, and our D2C & Marketplace Strategy Consulting folds the true fee picture back into pricing and catalogue decisions so you are not just clawing back the past but costing the future correctly.

The short version

Your marketplace settlements are calculated by the counterparty, against a fee schedule that moves, on data you do not audit. Under those conditions the errors will not be random. They will, on balance, favour the platform, not because anyone intends it but because that is what unchecked computation does. The leak is small per order and invisible per month, which is exactly why it runs for years.

Reconcile every payout against your own expected numbers, at the transaction level, on a fixed cadence. Treat a missing settlement as a fire and a mis-rated fee as recoverable cash, because both are. Assume you are being shorted until your own ledger proves otherwise. On a marketplace, that assumption is not cynicism. It is just arithmetic.

SKU Rationalization: Killing the Long Tail That Is Bleeding You

Open your listing count and feel the pride. Two thousand SKUs. Four thousand. A catalogue that looks like a serious operation. Now ask a harder question. How many of those SKUs sold more than a handful of units last quarter, and of the ones that did, how many made money after fees, returns, and the ad spend it took to move them. The honest answer, in almost every catalogue we have audited, is that a small head carries the business and a vast tail just sits there. The tail does not feel expensive because each dead SKU costs almost nothing on its own. Added up, it is one of the most expensive things you own.

SKU rationalization is the unglamorous discipline of cutting that tail on purpose. Not because pruning is virtuous, but because every dead listing is consuming something the winners need. Aggregate revenue hides this. Roll everything into one GMV number and the tail disappears into the average. You have to break the catalogue apart to see what it is actually doing to you.

The long tail is not free inventory, it is a tax

The seductive lie about a long tail is that it costs nothing to keep. The listing is already live. The photos are already shot. Why not leave it up in case someone wants it. The problem is that a SKU is never just a listing. It is a slice of working capital tied up in stock that turns once a year. It is a forecasting line nobody can predict. It is a row in every report that makes the real signal harder to read. It is operational attention every time it stocks out, gets a return, or throws a pricing error.

Multiply that across a thousand near-dead SKUs and you are running a second, invisible business whose only product is drag. The capital frozen in slow tail stock is capital you cannot put behind the SKUs that actually compound, which is the whole argument we make about working capital being the real constraint on marketplace growth. The tail does not lose money loudly. It loses it by denial of resources.

A dead SKU rarely shows up as a loss. It shows up as a winner you could not afford to stock deeper. The cost is the order you never placed.

Why aggregate revenue protects the tail

The reason most teams never cut is that the number they watch is built to hide the problem. Total revenue, total GMV, total units. At that altitude the tail and the head are blended into one comforting line that only ever goes up. Nobody looks at a rising top line and thinks half the catalogue should be deleted.

The tail only becomes visible when you change the unit of analysis from the catalogue to the SKU. The moment you rank every SKU by contribution rather than revenue, the bimodal shape appears. A steep head of SKUs that make real money, a long flat tail that makes almost nothing or actively loses. That ranking is exactly the output of measuring profitability per SKU, the number that reorders your whole catalogue. Rationalization is not a separate project. It is what you do with that list once you have it.

How to decide what dies

Cutting on gut feel is how good SKUs die and sentimental ones survive. The decision has to run on data, and the inputs are not exotic. For each SKU, you want a small set of honest signals over a trailing window:

Contribution, not revenue. Net of referral fees, fulfilment, returns, and ad spend. A SKU with healthy GMV and negative contribution is the first to go.
Velocity. Units per week. Slow-but-profitable is a different decision from slow-and-loss-making.
Inventory held. A dead SKU sitting on deep stock is freezing capital right now, not in theory.
Return rate. A high-return tail SKU costs far more than its refund line suggests once reverse logistics and write-offs are counted.
Strategic role. Some low-contribution SKUs earn their place as range fillers, search-coverage plays, or deliberate loss leaders. Name that reason explicitly, or cut.

Put those side by side and the catalogue sorts itself into keep, fix, and cut. The cut bucket is rarely small, and that is the point. You are not trimming a handful of mistakes. You are removing a structural drag the aggregate number was built to conceal.

Cut, merge, or fix

Killing is not the only move. A long tail often hides duplication. Six near-identical variants that split demand six ways, each looking weak alone, strong if consolidated into one or two. Merge those and velocity per SKU jumps without losing a single sale. Other tail SKUs are not dead, they are neglected. A broken title, missing attributes, or thin imagery suppresses them, and the fix is a listing problem, not a deletion. Telling the genuinely dead from the merely starved is where a catalogue data quality score your whole team can rally around earns its keep. Score the listing before you sentence it.

The forecasting dividend nobody mentions

Here is a benefit of rationalization that rarely makes the business case. A shorter catalogue is a more forecastable one. Every SKU you carry is a demand line someone has to predict, and tail SKUs are the least predictable lines you own. Spiky, sporadic, statistically hopeless. They add noise to planning and buffer stock you cannot justify.

Cut the tail and your forecasts get sharper, not just smaller, because you are now predicting demand that actually has a pattern. That compounds directly into better buying and fewer stockouts on the head, which is the entire premise of inventory forecasting for marketplaces when demand is spiky. A leaner catalogue is an easier catalogue to plan, and an easier catalogue to plan is a more profitable one.

What changed recently

Two forces in 2025 turned SKU rationalization from a quarterly hygiene task into a survival skill, and both came out of quick commerce.

First, the channel got more expensive to be average on. Through late 2025, FMCG brands reported quick-commerce margins falling three to five percentage points over six months as peak-slot ad rates nearly doubled and platform charges stacked toward forty percent of product price, according to Business Standard. When the channel taxes you that hard, a low-velocity SKU is not break-even, it is a guaranteed loss every time you pay to surface it. The contribution maths the whole article argues for is now the difference between a profitable shelf and a subsidised one.

Second, the platforms themselves are rationalizing. Blinkit ended the September 2025 quarter with 1,816 dark stores and is targeting 3,000 by March 2027 while moving to an inventory-led model, as covered by YourStory. A dark store holds a few thousand SKUs, not a few hundred thousand, and when the platform owns the inventory it has every incentive to stock only what turns. Your long tail does not just cost you. It gets you delisted by a buyer optimising the same shelf you should be. The winning brands on these platforms are running tight, hero-led ranges rather than sprawling portfolios, a shift Inc42 has tracked across the category. The discipline that used to be optional on Amazon is now mandatory on quick commerce, and it works the same way. We go deeper on this in pruning slow movers on quick commerce.

Run it as a standing discipline, not a one-off purge

The classic failure is to do this once, feel virtuous, and watch the tail grow straight back. SKUs accumulate the way clutter does, one reasonable addition at a time. A new variant here, a seasonal experiment there, a launch that never landed but never got removed. Without a recurring review, you are back to a bloated catalogue within a year.

So make rationalization a cadence, not an event. A quarterly pass that re-ranks every SKU by contribution and velocity, flags the bottom of the tail, and forces an explicit keep-fix-cut decision on each one. Pair it with a rule that new SKUs come in on probation. Earn velocity and contribution inside a window or get pruned automatically. That turns the catalogue from a thing that only grows into a thing that is actively curated.

The short version

A long catalogue is not a sign of strength. It is usually a sign that nobody has looked hard enough to cut. Aggregate revenue lets a tail of loss-making, capital-freezing, forecast-wrecking SKUs hide behind the winners, and the longer it hides the more it costs in orders you could not afford to place. With quick-commerce fees climbing and platforms stocking only what turns, the tail is no longer just a drag, it is a liability. Rank by contribution, decide cut, merge, or fix on data, free the working capital, and run it every quarter so the tail never grows back. The discipline is dull. The dividend, in capital and in clarity, is not.

Building the per-SKU view that makes these cuts defensible, and turning it into a standing review leadership trusts, is what our Analytics & Reporting work exists for. Cut the tail to fund the head. The catalogue gets shorter and the business gets stronger at the same time.

Customer LTV on Marketplaces: Estimating What You Cannot Directly See

Lifetime value is the number that decides how much you are allowed to spend to win a customer. On your own store you can calculate it almost cleanly. You know who bought, how often they came back, and what each order was worth. On Amazon or Flipkart you know none of that at the person level. The platform owns the buyer and hands you aggregates. So most marketplace brands set LTV aside as a D2C luxury, cap their acquisition cost at the first order, and quietly underspend on the categories that would have paid them back twice over.

That is a mistake, and it is the expensive kind, because it is invisible. You never see the second order you failed to fund. The honest position is not that marketplace LTV is unknowable. It is that you cannot measure it directly, so you have to estimate it deliberately, from signals the platform leaks, and treat the estimate as a range you can defend rather than a number you can prove.

Why LTV is worth estimating even when it is fuzzy

The point of LTV was never precision. It was a spending ceiling. If a buyer is worth eight hundred rupees over their life with you, you can rationally pay more to acquire them than if they are worth two hundred. That decision does not need a perfect number. It needs a defensible band and the discipline to act on it.

The brands that refuse to estimate LTV do not avoid the decision. They make it implicitly, by capping acquisition cost at the first-order margin. That cap is a hidden LTV assumption of exactly one order. For a consumable that reorders every month, that assumption is wrong by a wide margin, and the competitor who modelled it correctly will outbid you on every keyword that matters.

Refusing to estimate LTV is not caution. It is choosing the worst possible estimate, one order, and pretending you did not choose.

The signals that proxy lifetime value

You will never get a customer ID you can build a clean LTV model on. You will get enough to build a coarse one. Three signals carry most of the weight for marketplace brands in India.

Subscribe-and-save enrolment and churn. If you sell anything repeatable, this is the closest thing to a real retention curve a marketplace gives you. Enrolments and how many survive three and six months later are a near-direct read on how long a buyer stays. Treat the subscribe cohort as your highest-confidence LTV input.
Repeat-purchase rate at the SKU level. Amazon and, less generously, Flipkart report repeat behaviour in aggregate. The repeat rate is the probability of a next order. Stack a few of those and you have an expected order count, which is the spine of any lifetime-value estimate.
New-to-brand share against total orders. This tells you how much of each period is fresh acquisition versus returning demand. A buyer base that keeps returning is a base with real lifetime value. One that is all new-to-brand every month has none, no matter how good the topline looks.

None of these names a single customer. Together they describe the curve, and the curve is what LTV actually is. The work of turning these into a number is cohort work, which is why the two disciplines are inseparable. We lay out the proxy method in full in building cohorts without first-party data, and an LTV estimate is the natural output of those cohorts rather than a separate exercise.

A method you can defend in a meeting

Keep the model crude on purpose. A crude model you can explain beats a precise one nobody trusts. The estimate has three inputs and one honest caveat.

Expected number of orders. Start from the SKU repeat-purchase rate. If roughly a fifth of buyers place a second order and a smaller share a third, you can build an expected-orders figure without ever knowing who those buyers were. For subscribe SKUs, use the survival curve instead, it is cleaner.
Contribution per order. Use contribution margin, not revenue. Revenue-based LTV flatters you into overspending. Net of fees, ad cost, returns, and shipping is the only version that maps to what you can actually afford. This is the same discipline we apply when reading profitability one SKU at a time, and LTV inherits its honesty from that per-SKU margin work.
A decay assumption. Buyers churn. Apply a falling probability to each successive order rather than assuming a flat repeat rate forever. A simple decay is enough. The goal is a band, low and high, not a single confident point.

Multiply expected orders by contribution per order, discount the later orders, and you have a lifetime-value range per SKU or per category. State it as a range every time. The honesty of the range is what makes leadership trust the rest of the analysis.

Where the estimate breaks, say so first

The model has known weaknesses, and naming them yourself is what keeps it credible. Repeat rates from a promo-heavy period overstate loyalty, because deep-discount buyers repeat worse than full-price ones. A short observation window understates LTV for slow-reorder categories. And aggregate repeat data cannot separate one devoted buyer placing five orders from five buyers placing one each. Flag each of these in the room before someone else does, and the estimate survives the scrutiny it deserves.

What the estimate actually changes

An LTV band is useless if it does not move money. It should change three decisions directly.

Your acquisition ceiling. A SKU with a strong subscribe curve and a rising repeat rate can justify a first-order acquisition cost that looks reckless in isolation, because the second and third orders repay it. A SKU that never repeats has to win on the first order or be cut.
Where you scale spend. Two SKUs with identical first-order economics are not equal businesses if one has triple the expected order count. LTV is what separates them, and without it you fund both as though they were the same.
Who you chase back. The cohorts with the highest estimated value are the audiences worth re-engaging. That points budget straight at the demand most likely to return, which is the entire logic of retargeting marketplace shoppers when you do not own their data. You spend re-engagement money where the lifetime value is, not evenly.

Notice that every one of these is an acquisition decision. That is the point of the angle. You estimate LTV not to admire it but to know how aggressively you are allowed to spend at the top of the funnel. Subscribe-and-save and repeat signals proxy that value well enough to set the ceiling, and a deliberately rough estimate beats the silent assumption of one order every time.

What changed recently

The platforms have started handing you a cleaner subscribe signal than they used to, and the quick-commerce players have turned subscription into the explicit retention lever. That is good news for anyone estimating LTV, because it makes the highest-confidence input bigger and more legible.

The clearest proof is the subscription membership data itself. When Zepto piloted its paid membership, subscribers spent more than 30 percent more on the app and their monthly retention improved by roughly 10 percent, with the pass crossing a million members inside a week of launch, per YourStory. Read that as an LTV statement, not a marketing one. A membership cohort that spends a third more and churns less is a directly observable retention curve, the exact thing marketplace aggregates usually deny you. If you sell through a platform with a membership tier, the subscriber cohort is now your highest-confidence band, and you should be modelling it separately from walk-up demand.

The strategic backdrop reinforces the same point. Inc42’s 2026 quick-commerce outlook argues there is still Inc42 virtually no loyalty in the category, with shoppers keeping several apps and switching on availability and discount, and it expects growth to come from getting existing customers to order more often rather than from new cities, with subscriptions and private labels as the main stickiness levers. For a brand, that is a direct instruction. The platforms are now optimising for order frequency, which means the repeat and subscribe signals you build LTV on are about to carry more of the platform’s own attention, and the brands that already have a defensible LTV band will set acquisition ceilings the switchers cannot.

Make the number live somewhere leadership will look

An LTV band that lives in one analyst’s spreadsheet changes nothing. It has to sit beside acquisition cost and per-SKU margin in the same view, so the spending decision is obvious at a glance. The relationship between what a buyer costs and what a buyer is worth is the single most important comparison on a marketplace P&L, and it deserves a permanent line, not a quarterly slide. Getting it there without drowning the reader is its own craft, the one we mean by a reporting dashboard leadership will actually read.

Update the estimate as the cohorts mature. Early repeat data is noisy, and an LTV band built on two months of signal should widen, not pretend to certainty. As subscribe curves lengthen and repeat rates stabilise, the band tightens, and your acquisition ceiling earns more confidence with it. On quick commerce specifically, the same logic governs unit economics once platform fees are netted out, which is why we treat LTV and quick-commerce unit economics after platform fees as two reads on the same P&L.

The short version

You cannot see a marketplace customer twice. You can still estimate what they are worth, from subscribe survival, repeat-purchase rates, and new-to-brand share, assembled into a contribution-based lifetime band with an honest decay. The estimate will be a range, and that is correct. It is still enough to set an acquisition ceiling that beats the silent one-order assumption most brands spend under. And as platforms lean harder into membership and frequency, the subscribe signal at the heart of that band is only getting stronger.

Our Analytics & Reporting work exists to turn those scattered platform signals into an LTV band a brand can spend against, and our Marketplace Performance teams use it to decide where acquisition money actually compounds. The customer is hidden. Their value is not. Estimate it, defend the range, and spend like you mean it.

A Marketplace Reporting Dashboard That Leadership Will Actually Read

Most marketplace dashboards in India are built backwards. Someone exports everything Seller Central and Flipkart will hand over, drops it into a sheet, adds a tab per platform and a tab per metric, colours a few cells red, and ships it. It is thorough. It is comprehensive. And the founder it was built for closes it after eight seconds, because comprehensive is not the same as useful. A report that shows everything forces the reader to decide what matters, and that is your job, not theirs.

The dashboard leadership actually reads is small. It is built around the three or four decisions a founder makes every week, and it ruthlessly hides everything that does not feed one of those decisions. The data availability is not the point. The decision is the point. Start there and the whole thing gets shorter, sharper, and for the first time, read.

Build around decisions, not data availability

The cardinal mistake is letting the export define the report. The platforms give you hundreds of fields, so the dashboard ends up with hundreds of fields. But a founder does not wake up wanting to know the click-through rate on a Sponsored Brands placement. They wake up wanting to know three things. Are we growing. Are we making money doing it. Is anything on fire that I need to act on today.

Every number on a leadership dashboard should earn its place by answering one of those questions. If a metric does not change a decision, it does not belong on the page leadership sees. It can live in the analyst’s workbook, available on request, but it should not compete for the eight seconds of attention you actually get. The discipline is subtraction. Most dashboards are bad because nobody was willing to remove anything.

A dashboard is not a place to store data. It is a place to make a decision. If a number does not change what someone does on Monday, it is taking up space a decision should have.

The three numbers a founder actually wants

Reduce the top of the dashboard to the smallest set that still tells the truth. For most marketplace brands in India, that is three.

Net revenue trend. Total sales across platforms, shown as a trend, not a snapshot. One line, week over week. Not gross merchandise value the platform brags about, but the revenue you keep before ad spend. This answers are we growing.
Contribution after ads and fees. What is left once you subtract marketplace commission, fulfilment, returns, and advertising. This is the number that separates a busy account from a profitable one. It answers are we making money.
The one thing on fire. A single exception flag. Buy box lost on a hero SKU, account health slipping, a stockout on your best seller, ad efficiency collapsing on a key campaign. Not a list of twenty alerts. The one that needs a decision today.

That is the whole top of the dashboard. Everything else is supporting detail that a reader drills into only when one of those three numbers prompts a question. The structure mirrors how a founder actually thinks, which is why they read it.

Why the profit number is the hard one

Revenue is easy to show because the platforms hand it to you. Contribution is hard because you have to assemble it from fees, returns, and ad spend that live in different places and rarely reconcile cleanly. This is exactly why most dashboards skip it and lead with revenue instead. Revenue flatters. It always goes up if you spend enough. The contribution line is where the truth lives, and it is the one number a founder cannot get anywhere else without the work being done for them.

That work compounds when you push it down to the SKU. A blended profit number can look healthy while three SKUs quietly subsidise five that lose money on every unit. We have argued before that profitability per SKU is the number that reorders your whole catalogue, and a leadership dashboard should surface the worst offenders without making the founder hunt for them.

Trends beat snapshots, and exceptions beat lists

A number on its own lies by omission. Forty percent gross margin means nothing until you know whether it was forty-five last month. Lead with direction. Every headline metric should show where it is heading, not just where it sits, because the trend is what triggers a decision and the snapshot rarely does.

The second principle is exceptions over completeness. A dashboard that lists all two hundred SKUs is honest and useless. A dashboard that shows the five SKUs whose margin dropped this week is opinionated and read. The job of the reporting layer is to do the scanning so the human does not have to. If your founder is still eyeballing rows to find the problem, the dashboard has not done its work. It has just relocated it.

This is also where the right efficiency metric matters. A report that leads with a flattering advertising cost of sales while hiding the total picture is built to be skimmed past, not acted on. We make the full case in the piece on the ad metric your agency is probably hiding from you, and the short version is that leadership should see the number that reveals whether spend is building rank, not just the one that looks good on a slide.

Trust the dashboard before you trust the decisions

None of this works if the underlying data is dirty. A dashboard built on a catalogue full of wrong GTINs, mismatched titles, and duplicate listings will produce confident, precise, wrong numbers, which is worse than no dashboard at all, because people act on it. Before you obsess over the chart, fix the inputs. A catalogue data quality score the whole team can rally around is the unglamorous foundation that makes every number above it trustworthy.

Give the dashboard a visible health indicator of its own. A small note on data freshness and coverage, so leadership knows whether they are looking at complete numbers or a partial sync. A founder who once caught the dashboard being wrong will never trust it again. Showing the confidence level alongside the number is how you keep the trust you need for the report to drive action.

One dashboard, three altitudes

The mistake of building one report for everyone is real, but so is the mistake of building three disconnected reports. The answer is one source of truth read at three altitudes.

The founder view. Three numbers and one fire. Read in eight seconds, on a phone, between meetings. This is the default.
The operator view. The same numbers broken down by platform, campaign, and SKU, with the trends that explain the headline. This is where the marketplace team lives day to day.
The analyst view. The full export, the raw fields, the working. Available, never the front page.

Each layer is a drill-down from the one above, not a separate file. The founder sees revenue fell, taps once, sees it was two SKUs that stocked out, taps again, sees the supply note. Same data, more depth on demand. That structure is what lets a single dashboard serve a founder and an operator without either feeling it was built for the other.

The deeper cuts belong in the operator and analyst layers, not the front page. Retention and repeat behaviour is a good example. It matters enormously, but it is a question you investigate, not a number you glance at, which is why cohort analysis for marketplace brands sits a layer down rather than fighting for space at the top.

What changed recently, and what your dashboard must now capture

The cost lines a founder cares about have shifted faster than most dashboards have. Marketplace advertising is no longer a side cost you tuck into a footnote. Amazon and Flipkart together crossed roughly Rs 15,000 crore in India ad revenue in FY2025, with Amazon up about 25 percent to Rs 8,342 crore and Flipkart and Myntra together up around 27 percent, per Exchange4media. That money is your money. Retail media is now growing faster than social and video, which means the gap between revenue and contribution is widening for almost every brand. A dashboard that still leads with revenue is hiding the line that is moving fastest against you.

Quick commerce has made the same problem sharper and harder to reconcile. Platform fees, per-SKU listing charges, and bundled ad wallets on Blinkit, Zepto, and Swiggy Instamart have climbed through 2025, and smaller D2C brands report these costs eating into margin before a single sale is counted, as Storyboard18 has reported. The reporting consequence is concrete. A quick-commerce contribution line that does not subtract listing fees, ad-wallet commitments, and handling charges will read healthy while the real number is negative. We walk through the full breakdown in quick commerce unit economics after platform fees, and the dashboard lesson is simple. If your contribution number predates this fee escalation, rebuild it before you trust another decision it produces.

What this looks like when it is working

You know the dashboard is right when leadership stops asking for more reports. When the question in the Monday meeting shifts from what are the numbers to what are we doing about this SKU, you have built the right thing. The dashboard has stopped being a place where data is stored and become a place where decisions are made, which is the only reason to build one.

Our Analytics & Reporting work is built on exactly this principle of subtraction. We start from the three or four decisions a founder makes every week and design backward to the smallest set of numbers that drives them, then make those numbers trustworthy enough to act on. Paired with our Marketplace Performance teams, who own the metrics the dashboard surfaces, the report stops being a weekly chore nobody reads and becomes the page leadership opens first. Build around the decision. Hide the rest. That is the difference between a dashboard that exists and one that gets read.

Profitability Per SKU: The Number That Reorders Your Whole Catalog

Pull up your marketplace dashboard and sort by revenue. The SKU at the top feels like the hero of the catalogue. Everyone protects it, stocks deep on it, builds campaigns around it. Now do something most teams never do. Take that same SKU and subtract every cost the platform charges to sell one unit. Referral fee, closing fee, weight-based shipping, return handling, the ad spend it took to win the order. What is left is contribution per unit. Run that math across the catalogue and the list almost always reorders. The bestseller drops. A quiet SKU three rows down turns out to be the one paying the rent.

This is the single most clarifying number in marketplace analytics, and it is the one almost nobody reports. GMV is loud and easy. Profitability per SKU is quiet and hard. So teams optimise the loud number and wonder why scale never turns into money.

GMV ranks attention, not value

Gross merchandise value tells you which SKU moves the most rupees through the platform. That is a measure of attention, not of worth to your business. A high-GMV SKU can be a margin sinkhole. It might carry a heavy referral percentage in its category, ship at a weight that eats the contribution, attract returns that quietly double its cost of sale, or only sell at volume because it is propped up by aggressive ad spend that never appears next to the revenue line.

None of that shows up when you sort by GMV. The number looks magnificent right up until you net it down. And because the platform reports revenue prominently and costs in a dozen scattered statements, the flattering view is the default view. You have to go looking for the truth.

GMV tells you what the platform sold. Contribution per SKU tells you what you got to keep. Only one of those pays salaries.

What actually goes into per-SKU contribution

Contribution per unit is not complicated. It is just tedious, which is why it gets skipped. For a single unit of a SKU, start at the selling price and remove everything the sale costs you to deliver:

Cost of goods. The landed unit cost, including inbound freight and any pre-marketplace handling.
Referral fee. The platform’s category commission. This varies enormously by category and is the silent killer on low-price items.
Fulfilment and weight fees. Pick, pack, and shipping. Heavy or bulky SKUs can lose here even at a healthy headline margin.
Returns cost, amortised. Not just the refund. The reverse logistics, the inspection, the units that come back unsellable. A high return-rate SKU carries this on every unit sold, not only the returned ones.
Ad cost per unit. Total spend on that SKU divided by units sold. If it only sells because you pay for every click, that spend is part of its unit economics, full stop.

What remains is contribution per unit. Multiply by units and you have total contribution. Rank the catalogue by that, and you are finally looking at the business instead of the brochure. The returns line alone reorders fashion and apparel catalogues hard, which is why we treat return rate as a margin problem, not a logistics one.

The ad layer is where the bestseller usually dies

The most common reversal happens once you load ad cost onto the unit. A SKU can look profitable on cost-of-goods and fees alone, then turn negative the moment you account for the spend keeping it visible. This is exactly the trap of judging campaigns on the wrong metric. A pretty advertising cost of sales on a SKU that loses money per unit is not efficiency, it is a faster way to lose. We have argued at length that you cannot read ad efficiency without the total view, which is the whole point of looking at TACoS rather than the number your ad team prefers.

Put profitability per SKU and ad cost per SKU side by side and a pattern appears. Some SKUs earn their organic rank and barely need spend. Some are pure ad annuities, profitable only while you keep feeding them, dead the day you stop. Knowing which is which changes where every marginal rupee of budget goes. You stop subsidising vanity volume and start funding the SKUs that compound.

Quick commerce makes the math unforgiving

If marketplace contribution is tight, quick commerce is tighter. The take rates are steeper, the basket economics are different, and the margin for error is thin to non-existent. A SKU that contributes comfortably on a marketplace can go underwater the instant it enters a ten-minute channel with platform fees and darkstore economics layered on. Running per-SKU profitability is not optional there. It is the only thing standing between you and scaling a loss. We walk through that arithmetic in the quick commerce margin reality check, and the short version is that channel-blind unit economics will bury you.

The fee load on these channels is no longer the quiet part. Reporting in 2025 put Blinkit’s listing charge at around twenty-five thousand rupees per SKU per state, refunded as ad-wallet credit, with Instamart and Zepto quoting listing-cum-ad packages running into several lakh rupees, and one spice brand told Storyboard18 it spends ten to fifteen percent of GMV just to stay visible on the channel. A per-SKU model that ignores those fixed fees and the ad tax on top of them is not a model, it is a wish.

Same SKU, different channel, different verdict

Contribution is not a property of a SKU. It is a property of a SKU on a specific channel. The same product can be your best earner on one marketplace, break-even on another, and a loss leader in quick commerce. Averaging across channels hides all of it. The number only means something when it is cut by SKU and by channel together, which is precisely the kind of view a blended report is built to obscure.

What the reordered list tells you to do

Once you rank by contribution instead of GMV, the actions stop being guesswork. The catalogue sorts itself into a handful of honest buckets.

High contribution, high volume. Your real heroes. Protect availability ruthlessly, never let these stock out, and concentrate the budget that compounds here.
High contribution, low volume. Underexposed winners. These deserve more visibility, better listings, and the ad spend you were wasting on vanity SKUs. Fixing the listing often unlocks the volume.
Low contribution, high volume. The dangerous bucket. Loud, busy, and barely profitable or worse. Re-price, renegotiate cost of goods, cut the ad dependency, or accept they are a deliberate loss leader. Never mistake their GMV for health.
Low contribution, low volume. The long tail that quietly bleeds operational attention and working capital. This is the bucket for honest pruning.

That last bucket is where most catalogues are carrying dead weight they have never measured. Cutting it is not failure, it is hygiene, and it is the natural sequel to this analysis. Once profitability per SKU exposes the tail, the next move is rationalising the SKUs that are bleeding you rather than admiring how many listings you have.

What changed recently

Two shifts in the last several months should send every operator back to the per-SKU model to re-run it. The first is good news for low-price catalogues. In November 2025 Flipkart waived seller commission on goods under one thousand rupees, and Amazon India followed by removing referral fees in the same price band, a move Business Standard reported as a direct response to Flipkart. By March 2026 Amazon had expanded zero referral fees to more than twelve crore products under one thousand rupees across some eighteen hundred categories and trimmed Easy Ship fees for sub-three-hundred-rupee items, with YourStory noting sellers could cut fee costs sharply in that bracket. If a chunk of your catalogue sits under one thousand rupees, the referral line on those SKUs may have just gone to zero, and SKUs you had quietly written off as margin-negative can flip back into the black. Re-run the model before you prune them.

The second shift cuts the other way. The visibility tax on quick commerce keeps climbing. Storyboard18 reported that Blinkit, Zepto and Instamart together crossed three thousand crore rupees in advertising revenue in FY25 and are tracking toward roughly four thousand nine hundred crore this year, with advertising now near fifteen percent of Blinkit’s revenue. That money comes out of brand margins one sponsored slot at a time. The lesson is the same one the per-SKU model has always taught. Lower commissions on one channel do not make you profitable, and rising ad costs on another do not have to sink you. Only the contribution number, cut by SKU and by channel and refreshed when the fee structure moves, tells you which way each SKU actually broke.

Make it a number leadership can see

None of this works if the analysis lives in a spreadsheet one analyst opens once a quarter. Profitability per SKU has to be a standing view, refreshed and ranked, sitting where the people who set budgets and stock plans will actually look at it. That is a reporting discipline as much as an analytics one. A contribution-ranked SKU list, cut by channel, beside the GMV list everyone already trusts, is one of the most decision-changing things you can put on a single screen. Getting it there without drowning leadership in tabs is exactly what we mean by a dashboard leadership will actually read.

The short version

GMV ranks your catalogue by how much the platform sold. Profitability per SKU ranks it by how much you kept. Those two lists are rarely the same, and the gap between them is where the money you thought you were making quietly disappears. Net every SKU down to contribution after fees, returns, and ad spend, cut it by channel, and rank by what survives. When a platform zeroes a referral fee or raises an ad rate, the verdict on individual SKUs moves, so the model is not a one-time exercise. The bestseller you have been protecting may be the one you should be re-pricing, and the SKU you have been ignoring may be the one funding the business.

Building that view, channel by channel and unit by unit, is what our Analytics & Reporting work is for, and it is why our Marketplace Performance teams are judged on contribution rather than the GMV slide. Rank by profit, not by attention. The catalogue will tell you the truth the moment you ask it the right question.

ACoS vs TACoS: The Metric Your Agency Is Probably Hiding From You

Open most marketplace ad reports in India and you will see one hero number near the top. ACoS. Advertising cost of sales. It is the figure the ad team leads with on every call, the one that trends down month after month, the one that makes the slide feel like a win. And it is, on its own, almost useless for telling you whether your ad spend is actually building a business.

That is not an accident. ACoS is the metric an ad team reaches for when it wants to look good without being questioned. It can fall while your brand gets weaker. It can look brilliant in the deck while your real cost of selling on the platform quietly climbs. The number that exposes all of this is TACoS, and the fact that it rarely appears in your reports tells you something about who the reports are written for.

What ACoS actually measures, and what it conveniently ignores

ACoS is simple. It is ad spend divided by the revenue that came directly from those ads. Spend 100 rupees on Sponsored Products, get 500 rupees of ad-attributed sales, and your ACoS is 20 percent. Lower is leaner. So far so reasonable.

The problem is the word attributed. ACoS only sees sales the platform credits to a click on your ad. It is blind to everything else happening on the listing. Your organic sales, the orders that came from a buyer searching, finding you ranked well, and buying without ever touching an ad, do not appear in the denominator. So you can drive ACoS down to a flattering number by simply doing less, or by harvesting only the cheapest, easiest converting clicks, while the organic engine that ads were supposed to feed slowly stalls.

ACoS measures how efficiently you rented sales. TACoS measures whether you are building something you will still own next quarter.

TACoS is the same spend measured against the whole business

TACoS, total advertising cost of sales, changes one thing. It divides ad spend by total revenue on the platform, organic and paid together. Spend 100 rupees, generate 500 in ad sales but 1,500 in total sales, and your TACoS is around 7 percent against a much larger base. One number describes the ad campaign. The other describes the account.

That single change in the denominator is what makes TACoS honest. Because total sales include the organic orders ads are meant to influence, the trend in TACoS tells you whether your advertising is doing the job it is actually for. Ads on a marketplace are not just a sales channel. They are a ranking tool. Early velocity from paid placements pushes a SKU up the organic results, where it then sells without you paying for every click. TACoS is how you see whether that flywheel is turning.

How to read the trend, not the snapshot

A single TACoS figure means little. The direction over time means almost everything.

TACoS falling while total sales rise. This is the goal. Ads are seeding organic rank, organic is carrying more of the volume, and your dependence on paid is dropping. The flywheel is working.
TACoS flat while total sales rise. Acceptable during a scaling push. You are buying growth at a steady efficiency, but ads are not yet earning you free organic lift.
TACoS rising while total sales are flat. The warning sign. You are spending more to stand still. Organic is not picking up the slack, and every rupee of growth is rented, not owned.
ACoS pretty, TACoS ugly. The exact pattern a paid-only report is built to hide. The campaign looks efficient while the account leans harder on ads every month.

Why the ad team prefers the number that flatters them

Be fair to the people running your campaigns. ACoS is the metric the platform puts in front of them, the one their tooling optimises toward, and the one most cleanly under their control. It is natural to report on the number you can move. None of this requires bad faith.

But incentives are incentives. An agency paid on ad spend or judged on ACoS has every reason to keep the conversation on ACoS. TACoS implicates the whole account, including the listing quality, the catalogue, and the organic strategy that a pure ad team may not own and would rather not be measured on. When your report shows only the metric that makes the ad team look good and never the one that reveals whether the brand is getting stronger, that is a choice about what you are allowed to see.

This is also why budgets get set badly. A team optimising for ACoS in isolation will often underspend exactly when aggressive spend would buy lasting rank, and overspend on defensive clicks that protect a number rather than build one. The honest answer to how much to actually burn in a new brand’s first month only makes sense once you accept that early ad spend is buying organic position, not just immediate ROAS.

The same blindness shapes which campaigns you run

ACoS tunnel vision does not just distort budgets. It distorts strategy. Defensive, bottom-funnel, brand-keyword campaigns almost always post a beautiful ACoS, because you are paying to convert people who were already going to buy you. They add little organic rank because the buyer knew the brand already. Upper-funnel, category and competitor targeting runs a worse ACoS but is precisely the spend that wins new-to-brand customers and pushes you up the rankings where organic sales live.

Judge those two by ACoS alone and you will defund the campaigns that build the business and pour money into the ones that merely harvest it. This is the real cost of the wrong metric. It is the same trap behind splitting budget evenly without thinking, which is why we argue you should treat Sponsored Products and Sponsored Brands as different jobs rather than two buckets to fill equally. TACoS is what lets you defend the expensive-looking campaign that is quietly doing the heavy lifting.

What to demand in your next report

You do not need to become an analyst. You need to insist the report tells the truth about the account, not just the campaign. Three things make that happen.

TACoS shown beside ACoS, as a trend, every month. Never one without the other. The gap between them, and where each is heading, is the actual story.
The organic-versus-paid revenue split, also as a trend. A healthy account grows the organic share over time. If paid keeps taking a bigger slice, the ads are renting sales, not building rank.
Both cut to the SKU. Account averages hide everything. A blended TACoS that looks fine can sit on top of a few SKUs bleeding and a few carrying the rest.

That third point is where the work gets real. A blended number is comfortable precisely because it hides the variance, and the variance is the whole point. Reading TACoS per SKU is the natural partner to thinking about profitability one SKU at a time, because an efficient TACoS on a SKU that loses money on every unit is not a win, it is a faster way to lose. The two numbers only make sense together.

Getting this in front of the people who set budgets is its own discipline. A reporting layer that surfaces TACoS, the organic split, and per-SKU economics without drowning leadership in tabs is exactly what we mean by a dashboard leadership will actually read. That is the difference between data that exists somewhere in Seller Central and data that changes a decision.

What changed recently

The ACoS-versus-TACoS argument used to be an Amazon conversation. It is now a portfolio conversation, because the places brands buy ads have multiplied and the spend has exploded. Quick commerce is the clearest example. Zepto’s advertising revenue jumped 151 percent to roughly 1,636 crore rupees in FY26, up from about 651 crore the year before, per Storyboard18. Blinkit, Zepto and Instamart together are projected to pull in close to 4,900 crore rupees in ad revenue this year, on a Datum Intelligence estimate reported by Storyboard18. That same piece notes brands are already moving 10 to 25 percent of digital performance budgets onto these platforms for FMCG and impulse categories.

Two things follow for anyone reading these reports. First, this money is seller-funded, and the platforms increasingly lean on it to subsidise delivery economics, which means take rates and cost-per-click only travel one direction. Second, and more important for this article, the flywheel logic breaks on quick commerce in a way most ad teams have not adjusted for. On Amazon, paid velocity buys durable organic rank. On a dark store, shelf space is finite, ranking is thinner, and there is far less organic real estate for ads to seed. A flattering ACoS on Blinkit can sit on top of an account where almost nothing sells without paying for the slot. So the discipline matters more, not less. Demand the organic-versus-paid split on every retail-media platform you spend on, not just Amazon, and judge each one on whether the paid share is shrinking or quietly eating the whole account.

The short version

ACoS is not wrong. It is incomplete in a way that happens to favour the people reporting it. It tells you how cheaply you bought attributed sales and stays silent on whether you are building anything that lasts. TACoS fills that silence. It is the same spend measured against the whole business, and its trend is the closest thing you have to an honest read on whether advertising is making your brand stronger or just propping it up.

If your agency leads with ACoS and you have never once seen TACoS, that is the conversation to have this week. Our Analytics & Reporting work exists to put both numbers, the organic split, and the per-SKU truth on the same page, and our Marketplace Performance teams are measured against the metric that builds rank, not the one that flatters a slide. Ask for the number they are not showing you. The answer usually explains more than the one they are.

A Catalog Data Quality Score Your Whole Team Can Rally Around

Ask three people on a brand team how good the catalog is and you get three answers. The category manager says it is fine. The performance lead says it is the reason ads underperform. The founder has not looked in months. Everyone has an opinion and nobody has a number. That gap is where listing debt lives, quietly, for quarters at a time. The fix is not another audit deck that gets read once and forgotten. It is a single score, calculated the same way every week, that the whole team can rally around.

We are not talking about a vanity metric. We mean a catalog data quality score that is decomposable into fixable parts, owned by named people, and tracked over time like any other operating number. Once you have it, vague complaints about the catalog turn into a backlog with line items. That shift, from feeling to figure, is the entire point.

Why listing debt stays invisible

The trouble with a broken catalog is that nothing throws an error. A listing with a blank material field is live. A product with three images instead of seven still ranks, just lower. A size chart that does not match Indian fit still sells, just with more returns. None of this trips an alarm. The dashboard says complete. So the debt compounds in silence, and the only signal you get is a slow, unattributable drag on conversion and discoverability.

We have walked through this in detail before, because so much of the damage hides in fields buyers never consciously read. If you have not seen how backend attributes and image order quietly bleed performance, start with our breakdown of the catalog mistakes that kill conversion. The scoring system in this piece is the operational answer to that diagnosis. It takes the qualitative problems and makes them countable.

A catalog without a score is not a healthy catalog. It is an unmeasured one, which is a very different thing.

What a good score actually measures

A score is only useful if it maps to things a person can change this week. We avoid a single opaque number that nobody can decompose. Instead we build the score from weighted components, each one a concrete dimension of listing health. The weights shift by category, but the skeleton holds across Amazon, Flipkart, Myntra, and the quick-commerce platforms.

Attribute completeness. What share of the category’s available structured fields are filled and valid. This is the engine of on-platform discovery, so it carries heavy weight.
Image coverage and sequence. Whether the listing has enough images, in the right order, obeying the platform’s main-image rules. A hero shot plus six supporting frames scores higher than two stray photos.
Content depth. Title, bullets, and description present, on-spec, and free of the obvious failures like missing keywords or banned characters.
Variation integrity. Whether parent-child structure is correct so reviews and ranking signals pool instead of fragmenting.
Compliance and stability. GST and GTIN configured, MRP consistent, inventory signals reliable, no suppression flags.
Enhanced content presence. A plus content or rich media where the category and margin justify it.

Each listing gets a sub-score per component, and the components roll up into one catalog-level number. The detail is what makes it actionable. A catalog at 72 is not just a 72. It is 94 on content, 51 on attributes, and 60 on images, which tells you exactly where the week’s work goes.

Keep the rubric ruthlessly objective

The fastest way to kill a scoring system is to make it subjective. If two reviewers can look at the same listing and disagree on its score, the number is dead on arrival. So every check must be binary or counted, never judged. Attribute filled or blank. Image present or not. Seven images or four. Resist the urge to score copy quality on a feel-based scale. You can grade whether keywords from your research are present, which is checkable, but not whether the prose is elegant. Note that on-platform keyword logic is its own discipline, distinct from web search, and your scoring rules should reflect that as we argue in our piece on listing keyword research for Indian marketplaces.

From score to assignable backlog

A number on a slide changes nothing. The score earns its keep when it generates a queue of work. The mechanism is simple. Every listing below the threshold on a given component produces a task, and that task has an owner, a fix, and a point value equal to the score it will recover.

This reframes the whole conversation. Instead of a manager saying the catalog needs improvement, the standup says there are forty listings missing the occasion attribute, that is six points of catalog health, and it is assigned to the content team for Thursday. Listing debt becomes a sprint backlog. People can see what they own and what it is worth. The score moving up each week is the proof that the work mattered.

Prioritisation falls out naturally too. You do not fix the catalog alphabetically. You fix the highest-revenue listings with the lowest scores first, because that is where recovered points convert to recovered sales fastest. A cheap, low-traffic SKU at 40 can wait. A hero product at 65 cannot.

How the score connects to revenue

The objection we hear is fair. Is a catalog score just hygiene theatre, or does the number actually move money. The honest answer is that the score is a leading indicator, not a guarantee. A higher score does not promise more sales the way a discount does. What it does is remove the structural reasons a listing cannot convert, which is a precondition for everything downstream.

This is why the catalog score and your conversion work belong on the same table. Once a listing is structurally sound, the real optimisation begins, and that is a different experiment entirely. We are firm that the highest-leverage test is usually the image, not the bullet, which we make the case for in our argument on conversion rate optimisation for listings. The score gets you to the start line. CRO is the race after it.

The connection to revenue becomes legible when you put the score next to outcomes leadership already watches. Track catalog health alongside conversion rate and ad efficiency on the same view, and the correlation tells its own story over a few months. If you are building that view, the principles for a report executives will actually open carry over directly from our take on a marketplace reporting dashboard leadership will read.

Running the score as a habit, not a project

The most common failure is treating the score as a one-time cleanup. The team rallies, the number jumps from 68 to 88 over a month, everyone celebrates, and then it drifts back down. New SKUs launch with half their attributes blank. Platform schema changes add fields nobody fills. Entropy is the default state of a catalog.

So the score has to be a recurring measurement with a standing owner, not a quarterly heroics exercise. The cadence that holds in practice:

Recalculate the catalog score on a fixed weekly schedule, automatically where the platform data allows.
Set a non-negotiable launch threshold so no new listing goes live below a minimum score.
Review the component breakdown in the weekly operating meeting, not a separate catalog meeting nobody attends.
Convert every gap into an owned task with a point value and a due date.
Track the trendline, not the snapshot, so you catch drift before it becomes debt again.

What changed recently

Two shifts in the last year make the score harder to treat as optional. The first is on Amazon itself. The platform has moved required attributes from a soft suggestion to an enforced gate, expanding the structured fields you must supply to create or edit a listing and tightening attribute usage and enumeration values across product types in its listing requirement changes. Translation for your rubric: attribute completeness is no longer just a discovery lever you choose to pull. It is increasingly a precondition for the listing existing in valid form at all, which means the weight you put on that component should go up, not down.

The second shift is where the money is moving, and it is the stronger argument for taking catalog quality seriously this year. Quick-commerce platforms have turned into serious ad networks, and a listing that is not structurally complete cannot earn the placements brands are now paying hard for. Zepto’s advertising revenue grew about 151 percent to roughly Rs 1,636 crore in FY26, per figures in its draft prospectus reported by Storyboard18, and a Datum Intelligence estimate cited by Storyboard18 projects Blinkit, Zepto, and Instamart together could pull nearly Rs 4,900 crore in advertising revenue in 2026, with FMCG and impulse brands said to be shifting between 10 and 25 percent of their digital performance budgets onto these platforms. When that much spend rides on a listing, a blank attribute or a missing image is not a hygiene problem. It is wasted media against an incomplete product page. The catalog score is what stops you from buying traffic to a listing that was never ready to convert it. If you are deciding where that spend goes first, our view on quick-commerce unit economics after platform fees is the companion read.

This is the unglamorous discipline behind Catalog & Listing Optimization, and it is deliberately mechanical. The score does not need to be clever. It needs to be consistent, objective, and visible enough that the whole team trusts it. Pair it with steady Marketplace Account Management so the launch threshold actually gets enforced, and with Marketplace SEO so the discoverability gains from a complete catalog show up where buyers search.

The teams that win at marketplace catalogs are not the ones with the most opinions about quality. They are the ones who turned quality into a number, gave the number an owner, and watched it climb. Give your catalog a score this week. The debt you have been ignoring will finally have a name.

Cohort Analysis for Marketplace Brands Without First-Party Data

Every retention playbook written for D2C assumes one thing you do not have on a marketplace. The customer. On your own store you have an email, a phone number, an order history tied to a person. On Amazon or Flipkart you have a settlement report and a wall. The platform owns the buyer, guards the identity, and hands you aggregates. So most marketplace brands quietly conclude that cohort analysis is a luxury for people with first-party data, and they stop looking.

That conclusion is wrong, and it is expensive. You cannot run a textbook cohort table keyed to individual customers. But the marketplace leaks enough repeat-behaviour signal that you can build cohorts that are directionally true and good enough to change what you spend, what you stock, and what you launch. The trick is to stop trying to reconstruct the customer and start reading the signals the platform cannot hide.

What you are actually missing, and what you are not

Be precise about the gap. The thing you lose on a marketplace is identity resolution. You cannot reliably say buyer 4471 purchased in January and again in April. Amazon’s brand tooling gives you a new-to-brand flag and some repeat-purchase aggregates, Flipkart gives you less, quick commerce gives you almost nothing at the person level. None of it is a clean customer ID you can build a classic cohort grid on.

What you keep is more than people assume. You keep the new-to-brand percentage on advertised sales. You keep subscribe-and-save enrolment and churn if you sell consumables. You keep total repeat-purchase rate at the account or SKU level where the platform reports it. You keep your own units-per-order and the gap between gross units sold and unique-ish demand. And you keep time. Every one of those is a retention signal. Cohort analysis without first-party data is the discipline of assembling those signals into a defensible view of whether buyers come back, even when you can never name a single one.

You are not reconstructing the customer. You are reconstructing the curve. The curve is what actually drives the decision.

Build cohorts on proxies, not people

If you cannot cohort by customer, cohort by the next best thing. The most useful unit is the acquisition window. Group everything by the month a buyer most plausibly entered the brand, then watch how the brand’s behaviour evolves against that window. You will not have per-person retention, but you will have a brand-level repeat signal that moves with the cohort.

Three proxy cohorts do most of the work for marketplace brands in India.

Launch-month cohorts. Tag the month a SKU or variant went live. Track repeat-purchase rate and subscribe enrolment for that SKU over the following months. A consumable launched in March that shows a rising repeat rate by June is building a base. One that sells hard on launch and flatlines is renting demand from ads.
Promo cohorts versus organic cohorts. Split the months you ran a deep Big Billion or Great Indian Festival push from the quiet months. Buyers acquired during a heavy discount window almost always repeat worse than buyers acquired at full price. The platform will not tell you this per customer, but the repeat-rate trend across promo-heavy and promo-light periods will.
Subscribe cohorts. If you sell anything repeatable, subscribe-and-save is the closest thing to a real customer cohort the marketplace will ever give you. Enrolments by month, and how many are still active three and six months later, is a near-clean retention curve. Guard it like the asset it is.

Each of these is a cohort in the way that matters. It groups demand by when and how it was acquired, then measures whether it persisted. That is the entire point of cohort analysis. Identity is a convenience, not a requirement.

The new-to-brand number is your acquisition denominator

Amazon’s new-to-brand metric is underused as a cohort input. Read alongside total orders, it tells you what share of this period’s sales came from buyers the brand had probably never seen. A high new-to-brand share with flat total sales means you are acquiring and leaking in equal measure. A falling new-to-brand share with rising sales means existing demand is carrying you, which is healthy until it is stagnant. Tracked monthly, new-to-brand becomes the front edge of every acquisition cohort you build.

Repeat-purchase rate is the signal to defend

If you only instrument one thing, instrument repeat-purchase rate at the SKU level and watch its trend. It is the cleanest retention proxy a marketplace gives most brands, and it answers the question that actually pays. Are we building something, or are we buying the same sale again every month.

The danger is reading it as a snapshot. A 22 percent repeat rate means nothing in isolation. The same number rising across three launch cohorts means your product and your post-purchase experience are earning a second order. The same number falling while ad spend climbs is the warning that you are acquiring worse buyers, or that a competitor undercut the reorder. This is the same trap we flag when teams stare at a flattering blended figure instead of the trend, and it is why a per-SKU read matters more than an account average. We have argued the same logic from the cost side in looking at profitability one SKU at a time, and retention and margin are the two halves of whether a SKU deserves its shelf.

Repeat rate also reframes acquisition. A SKU with a strong, rising repeat signal can justify a worse acquisition cost, because the second and third orders pay it back. A SKU that never repeats has to win on the first order or not at all. Without cohorts you cannot tell these two apart, and you end up funding both as if they were the same business.

From cohorts to the number that matters

A repeat curve is not the destination. It is the input to value. Once you have a defensible repeat-purchase trend and an average order value, you can build a rough estimate of what a marketplace buyer is worth over time, even though you can never see that buyer again. The estimate will be a range, not a point, and that is correct. It is still enough to decide whether a category is worth scaling.

This is where cohorts feed directly into estimating customer LTV on marketplaces. The repeat rate gives you the probability of a next order, the order value gives you its size, and the cohort trend tells you whether that probability is improving or decaying. Stack those and you have a lifetime-value band built entirely from signals the platform did not mean to give you. It will be coarser than a D2C model. It will also be the difference between guessing and reasoning.

Cohorts also tell you where to spend retargeting effort

The cohorts that repeat well are the audiences worth chasing back. Even without customer identity, the platforms let you reach lookalikes and prior viewers, and knowing which cohort actually returns tells you which intent is worth paying to re-engage. That feeds straight into retargeting marketplace shoppers when you do not own their data, where the whole game is spending re-engagement budget on the demand most likely to convert again rather than spraying it evenly.

What changed recently

The case for reading retention sideways got stronger over the last year, because the platforms themselves stopped pretending a subscription badge equals loyalty. In February 2026 Zepto quietly shut down Zepto Daily, its loyalty and subscription programme, ahead of its IPO, and Swiggy Instamart’s own chief described the market as so irrationally competitive that customers switch platforms without any real loyalty, per Inc42. The lesson for a brand is direct. A subscribe enrolment is still your cleanest cohort, but enrolment is not retention. Track how many of each month’s subscribers are still active at three and six months, because the platform will happily sign people up and just as happily let them lapse.

The second shift is where the platforms are actually investing their reporting effort, which is the ad side. Quick commerce ad spend on Blinkit, Zepto and Instamart jumped to roughly 4,000 Cr in 2025 and is projected near 6,000 Cr in 2026, and the platforms now hand brands granular shopper signals like basket mix, order timing, locality and purchase frequency to feed that machine, as Inc42 documents. Purchase-frequency reporting is a cohort input the moment you stop reading it as a vanity number and start grouping it by acquisition window. The data exists. Whether it changes a decision is on you.

The third shift is cost. Blinkit and Zepto hiked commissions through 2025 to push toward profitability, which means the bar for a SKU to deserve its slot just went up, per Business Standard. Higher take rates make the repeat curve the deciding variable, not a nice-to-have. A SKU that only ever wins the first order cannot absorb a richer commission. One with a real, rising repeat signal can. The math we used to treat as analysis is now the difference between a profitable line and a subsidised one, which is exactly the discipline behind quick-commerce unit economics after platform fees.

Make it survive contact with leadership

The honest weakness of marketplace cohorts is that they are proxies, and proxies invite an easy dismissal. Someone in the room will say this is not real cohort data, and they will be technically correct and practically unhelpful. Pre-empt it. State the proxy plainly, show the trend over enough months that noise washes out, and tie every cohort to one decision it changed. A cohort view that does not move a budget or a launch is decoration.

Three habits keep these cohorts credible.

Always show the trend, never the single number. One repeat rate is an opinion. Six months of the same cohort metric is evidence.
Name the proxy out loud. Say repeat-purchase rate as a stand-in for retention, not retention. The honesty is what makes leadership trust the rest.
Cut to the SKU and the channel. Blended cohorts hide the variance that is the entire reason to look. A healthy account average can sit on a hero SKU that repeats and a long tail that never does.

Getting this in front of decision-makers without burying them in tabs is its own craft, and it is exactly what we mean by a dashboard leadership will actually read. The cohort table is useless if it lives in a spreadsheet nobody opens.

The short version

Not owning the customer does not mean you cannot see retention. It means you have to read it sideways, from new-to-brand share, subscribe curves, and repeat-purchase trends grouped by how and when demand was acquired. Those proxies will never be as clean as a D2C cohort grid. They are clean enough to tell you which products earn a second order, which promos buy disposable buyers, and which categories deserve more money.

Our Analytics & Reporting work exists to assemble exactly these signals into cohorts a brand can act on, and our Marketplace Performance teams use them to decide where acquisition spend actually compounds. The customer is hidden from you. The curve is not. Build on the curve.