Deductions Start in Sales and Land in Finance: Fixing the Handoff

Most articles about deduction management describe what finance should do once a deduction arrives: log it, classify it, research it, dispute or accept it, recover what is invalid. That work matters, and we have written about the AR-side lifecycle and its best practices in detail.

This one is about the part that happens earlier, in a different department, usually months before finance sees anything: the commitment that created the deduction in the first place.

The deduction was created in a sales conversation

Walk backwards from an invalid deduction and you almost always arrive at the same place. A distributor short-pays an invoice by ₹1.8 lakh, citing a damage allowance. The AR team searches the agreement, finds no damage clause at that rate, and raises a dispute. The distributor's response is not a denial — it is a name. Your area manager agreed this in March.

Sometimes the area manager remembers, sometimes not. Sometimes the manager has left. What is almost never in question is that something was agreed; what is in question is what, and on what terms.

This is the structural fact that most deduction-management advice steps over. Finance owns the consequence but did not create it. By the time a deduction is deductible, the decision that made it valid or invalid is old, undocumented and contested.

Three commitments that reliably become deductions

The verbal scheme. A distributor is hesitating on a quarter-end load. The area manager offers an extra 1% if they take the full quantity. It works. Nothing is circulated, because the offer was specific to one partner and slightly outside the published scheme. The distributor deducts it the following quarter, correctly by their reckoning, and finance has no scheme to match it to.

The damage assurance. A batch arrives with transit damage. The salesperson, on site and wanting the relationship intact, says the company will take care of it. That sentence sets no rate, no cap and no documentation requirement. Whatever the distributor claims later is, from their side, exactly what was promised.

The price-protection promise. A price revision is coming and the distributor is holding stock. The salesperson assures them they will be protected. Whether that meant full stock at the old landed cost, or only unsold stock at the differential, is precisely the ambiguity that shows up as a deduction.

None of these are dishonest. All three are ordinary commercial judgement made in the field. The failure is that the judgement never became a record.

Why sales does not chase the invalid ones

There is a second, quieter dynamic underneath this, and it is an incentive problem rather than a discipline problem.

Most channel sales roles in Indian distribution are measured on primary sales, or on collection against primary. Neither number moves when a deduction is written off. The salesperson's scorecard is largely indifferent to whether a claim was valid; finance's P&L is not.

So when finance asks sales to push back on a ₹1.8 lakh deduction, they are asking someone to spend relationship capital on an outcome that does not appear in their own results — and which may cost them the next quarter's cooperation. Quietly conceding is the rational choice. Writing off the deduction is the cheapest way to end the conversation.

You will not fix this with process documents. It is fixed, if at all, by putting some version of net realisation after deductions into the sales scorecard — and even then, carefully, because a blunt version punishes salespeople for contractually valid claims they were entirely right to agree to.

Who should own what

The two common ownership models both fail, in opposite directions.

Finance owns everything. AR raises disputes directly with the distributor. Relationships fray, because the person disputing has no relationship to spend and no context for what was agreed. Sales finds out when the distributor complains.

Sales owns everything. Sales decides which deductions to contest. Almost nothing is contested, for the reasons above, and finance discovers the leakage at year-end.

The model that holds is a split along a specific line:

• Finance owns the process and the decision — classification, evidence, the valid/invalid determination, the recovery workflow, and the number that gets reported.
• Sales owns the customer conversation — because they hold the relationship, and because they are frequently the only person who knows what was originally agreed.

We have written separately about how the wider deduction roles divide up. The distinction that matters most here is between deciding and communicating. Conflating them is what produces both failure modes.

Fix the capture, not the dispute

Most improvement effort goes into the dispute process — faster research, better classification, tighter follow-up. That is worth doing, and it is also downstream of the actual problem.

A deduction is easy to validate when there is something to validate it against. It is impossible to validate when the only record is a conversation. So the highest-leverage change is not in AR at all: it is making it easy enough to record a commitment that the field actually does it.

In practice that means:

A scheme is an agreement, not a circular. If a partner-specific offer is made, it needs to exist as a record with a partner, a period, a base and a rate — not as an email that a regional office may or may not have filed.

Allowances get a rate and a cap at the moment they are promised. "We'll take care of it" is not a commitment anyone can settle against. "Up to 2% of the batch value, on evidence, within 30 days" is.

Price protection states its scope. Unsold stock or all stock; differential or full; verified how. Three seconds of specificity at the promise removes months of argument at the deduction.

The record is reachable from a phone. A capture process that requires a laptop and a desk is a capture process the field will skip, and every commitment it fails to catch becomes an unverifiable deduction later.

What this looks like when it works

The change is not that disputes get won more often. It is that far fewer deductions are ambiguous in the first place.

A distributor deducts against a scheme; the scheme exists with dates, slabs and eligible SKUs; the system matches the deduction to it and either confirms the amount or shows exactly where the two disagree. The AR analyst is no longer reconstructing history from memory and email — they are looking at a variance, which is a five-minute conversation instead of a five-week one.

The deductions that remain genuinely contested are then a much smaller and more honest set: real disagreements about real terms, which is what a dispute process is actually good at handling.

Where the software sits, and where it does not

RebateLedger is built around this idea: the agreement is the thing a deduction is checked against. Schemes, allowances and protections are recorded with their terms; inbound deductions are matched against them; what does not match is surfaced with the reason attached rather than left for someone to work out. The valid-versus-invalid determination becomes a comparison rather than an investigation.

What software does not fix is the incentive. If your sales team is measured purely on primary sales, they will keep conceding deductions, and no amount of matching logic will change that. That one is a compensation design decision, and it belongs to whoever owns the sales scorecard — not to finance, and not to a system.