Channel exit: write the trigger before you enter
Most channel exits in India happen about four quarters late, and almost never because the numbers were unclear. The numbers were clear for a year. What was missing was a named person with the authority to say the word, and a date on which they were obliged to say it.
- The person who built the case for entering a channel is usually the person reporting on it, so separate the channel target from the channel profit and loss reporting.
- Contribution after all channel-caused costs tells you the channel is losing money, but only a release test tells you how much cash leaving actually frees.
- A channel you have run badly is fixable and a channel that is structurally wrong for your price point is not, so spend one honest quarter on the fixable list before calling it structural.
- Write the contribution level, the review date and the name of the decision maker in the launch document, while everyone is still optimistic about the channel.
A channel exit is rarely blocked by analysis. It is blocked by the fact that no single person is required to make the call on a given date, so the decision drifts until somebody gets tired. The entry decision usually had a framework behind it, whether that was a channel mix view or a simple margin comparison. The exit decision almost never has one, and it is not the entry decision run backwards.
Whoever launched it is grading their own homework
The person who built the business case for the channel is usually the person reporting on it. That is not dishonesty, it is how the review got scheduled. They picked the metrics at launch, so the deck still leads with GMV growth and listings live, because those were the goals they were handed. Contribution shows up on slide nine, if at all.
Two structural fixes cost nothing. Separate the person who owns the channel target from the person who reads out the channel profit and loss, so finance narrates and the channel owner explains. Then ask the entry question at every review: if this channel did not exist today, what would we spend to start it. Brands that answer honestly often find the answer has been zero for three quarters running.
The second failure mode is quieter. Nobody wants to be the brand that could not make a large marketplace work, or that got pushed off quick commerce. Leaving reads as failure to a board and to a category in a way that never entering does not, so the channel stays open as reputational cover at a real monthly cash cost. Naming that out loud in the room is usually enough to defuse it.
Contribution first, then the release test
Revenue is not the input. The input is contribution after every cost the channel causes: commission, fulfilment or shipping, the ad spend that exists only to defend that channel’s search results, returns and RTO, damages and deductions, storage, and stock parked in the platform’s network that you cannot sell anywhere else this month. Run it at SKU level within the channel before you judge the channel, because two or three items usually carry a long loss-making tail.
Then do the part most brands skip. Ask what actually goes away when you leave, and sort every cost into three buckets rather than two. Cash that stops on the exit date: commission, per-order fulfilment, channel-specific ad spend, platform storage. Cash that stops only if somebody makes a second decision: the cataloguer, the reconciliation analyst who also covers two other channels, the returns grading line, the warehouse space you leased for eighteen months. And cash that never stops, which is the fixed overhead you were allocating to the channel for reporting purposes and nothing more.
The cash that stays is what decides
Put a rupee figure and a release date against every line in the middle bucket, because that is where the forecast dies. A salaried role does not free up for at least a notice cycle, and usually never, because the person gets moved rather than let go. Leased space releases at the end of the lease or at the price of breaking it. Shared software seats release at renewal. A brand that forecasts a saving of several lakh a month and books a third of it has not been misled by anyone. It just never separated the first bucket from the second.
The third bucket has a nastier effect on the reporting. Overhead charged to the exiting channel does not evaporate, it redistributes onto the channels that remain, so your own site and your other marketplaces each look slightly less profitable the month after you go. If nobody warns the team, that reads as the exit having damaged the rest of the business. Rerun the margin comparison across your remaining channels on the post-exit allocation before you leave, so the new numbers land as expected rather than as evidence you made a mistake.
This changes the conclusion more often than people expect. A channel that is contribution-negative but releases very little cash is a management problem, not a portfolio problem. You are not choosing between the channel and nothing. You are choosing between the channel and whatever that same team does next, and if you have not decided what that is, the exit does not pay for itself.
Run the fixable list before you call it structural
Structurally wrong means the platform’s take rate, plus its fulfilment cost, plus the price point its shoppers expect, leaves no contribution at any volume you could realistically reach. Badly run means something else. You never held the buy box. Your listing content was thin and your conversion rate showed it. Your case pack did not suit the platform’s replenishment cycle, so you were always short. You were absorbing deductions you were entitled to claim back.
Give the fixable list one honest quarter with a named owner and a small budget. Fix catalogue quality, claim what you are owed, correct pack configuration, and cut the ad spend that was buying orders below contribution. If the channel is still negative after that with the basics genuinely in place, it is structural, and now the whole room believes it rather than half of it.
Set the date while you are still optimistic
The best week to write the exit trigger is the week you enter, when nobody is defensive and nobody’s credibility is attached to the outcome yet. It is three lines in the launch document: the contribution level the channel must reach, the date it must reach it by, and the name of the person who makes the call. This is the same instinct that makes people write exit terms into supplier agreements, and the reasoning in exit clauses in vendor contracts transfers cleanly to your own channels.
Decisions taken on fatigue get taken in a bad month, usually just after a festive return spike, and they get reversed the next quarter. Decisions taken against a pre-agreed line survive the argument, because the argument already happened. For the underlying case on what the numbers have to show, read when to exit a sales channel. Once the call is made, execution is a separate discipline: leaving a marketplace cleanly covers the sequence, and if the honest answer is a narrower range rather than a full exit, discontinuing a product line is the smaller move.