Discounting the Second Time: When the Same Concession Is Taken Twice

The reason this matters is arithmetic, and the arithmetic is worth doing slowly. Here is a worked example we constructed for this article. Every figure is invented so that a reader can check it by hand. The 18% GST rate appears only to keep the totals checkable and is described that way; nothing here is guidance on how anything should be treated, and any question about tax presentation belongs with your own chartered accountant.

The work is priced at ₹2,00,000 taxable. A first discount of 10% is applied by the person selling it: ₹2,00,000 × 0.10 = ₹20,000, so the quoted total becomes ₹2,00,000 − ₹20,000 = ₹1,80,000. The quote of record now says ₹1,80,000, and the ₹20,000 is gone from the record entirely, absorbed silently into the price. Nothing in the quote says a concession was made, because the quote only carries the price.

Now finance receives the invoice request. Finance does not have the quote, or does not look at it, and applies what it calls a standard account adjustment of 5% to the invoice total: ₹1,80,000 × 0.05 = ₹9,000. The invoiced value is ₹1,80,000 − ₹9,000 = ₹1,71,000.

Read as a story, the concession taken was ₹20,000 + ₹9,000 = ₹29,000 against a list price of ₹2,00,000. Read as a percentage of the list price, the second discount alone is 9,000 ÷ 2,00,000 = 4.5%, and the two together are ₹29,000 ÷ ₹2,00,000 = 14.5%. We are not going to tell you whether 14.5% is a lot. It is three discounts and one quote that we invented for this article, and the number describes nothing outside this page.

What is worth noticing is the size relationship. The second discount was five percent of a number that had already been reduced, so it produced less money than it would have against the list price, and it was still taken without anybody knowing it was the second. Nobody in this example cheated. Everybody was reasonable. The total is the problem, and the total is the only thing that was never a record.

This is the part that makes the failure hard to catch. A system that records each step accurately is doing its job, and the failure is not a bug in the system. The failure is that the steps do not know about each other.

The quote is a record of a price that was offered. It is a complete and accurate record. The invoice is a record of what was billed, raised against a project, and it is a complete and accurate record of what finance decided to bill. The reduction is an approved adjustment, recorded with a reason, a person, and a date. Every one of those three records is true.

What does not exist is a fourth record: the total concession given on this project, from list price to collected value, with both decisions in one place. That record is not a byproduct of keeping the other three. It is a separate thing that has to be constructed on purpose, because it is the only view in which the two decisions are adjacent to each other.

The practical consequence is that the question a manager would want to ask, which is how much did we give away on this project, cannot be answered from either record. The quote cannot answer it because the quote does not know about the invoice. The invoice cannot answer it because the invoice does not know the list price, only the discounted price it was raised from. Two records, each correct, and the question falls between them.

A discount that is applied and then disappears into the price is not recorded, it is absorbed. The difference is not semantic. An absorbed discount cannot be totalled, cannot be attributed, and cannot be reviewed, because the number that survives is the only number there is.

For a concession to be countable, it needs at least three things. It needs its own identity, as a reduction rather than an adjusted price, so that it can be added up later. It needs the price it was applied to, so that the base of each discount is known and two discounts can be seen to be sitting on top of each other. And it needs a reason, held by the person who authorised it, that survives the month end.

The reason matters more than it looks. A concession with a written reason can be reviewed later as a policy question: is this reason common enough to be a policy, and if so who set it. A concession without a reason becomes an accumulation, and an accumulation is defended by inertia rather than by argument. Nobody approves a pattern of concessions. It simply happens, one reasonable exception at a time, and then it is the way the business prices.

It is worth being clear about what the platform does and does not do here, because a pricing conversation is exactly where a reader assumes there is a rule engine quietly preventing this.

In NoxOrigin an invoice and a payment are different records, and paid is a projection of allocations rather than a stored flag, so money and billing never collapse into one field. That discipline is what makes the chain auditable. But there is no discounting rule engine that knows your commercial policy, and there is no approval matrix for reductions that we would be able to describe accurately enough to promise. What the platform does is keep the quote and the invoice on one project, so that the two decisions are adjacent in the data even when they were made by different people in different weeks.

There is also no contract editor, no e-signature, and no change-request record. That matters more here than in most articles, because a change-request workflow is the obvious place to catch a revision, and a discount taken at the invoice is a revision of the commercial terms. The honest answer is that a discount taken at the invoice is a new quote raised against the same project, and the comparison between the two is a thing a human does rather than a thing the software enforces. The friction is deliberate; the enforcement is not available, and pretending otherwise would be the more dangerous kind of optimism.

Expenses are Nox-Billings only, and purchasing lives in Commerce. There is no timesheet record and no payroll module, so a discount that was granted in exchange for labour cannot be checked against hours worked anywhere in this platform. Nor does NoxOrigin file GST returns or any other statutory return; the reduction you apply is your decision, and its tax presentation is a question for your own chartered accountant.

The review is a single comparison, run per project rather than per month: quoted value against invoiced value, with every reduction between them counted rather than netted. Run consistently, it takes the double-taking out of the category of invisible things, and it converts a series of small reasonable exceptions into a number somebody has to look at.

The reason a project is the right unit rather than a customer or a month is that a customer-level view hides which project carried the concession, and a monthly view hides which decision started it. The project is where the two decisions belong to the same piece of work, and therefore where their sum means something.

The companion failure modes in this cluster belong in the same sitting. The quotation that never became a scope is the first article, and a double-discount is often its consequence, because an undocumented concession has nothing to be measured against. The estimate nobody revisited is the third, since a stale estimate is a concession that time took rather than a person. And whether a margin figure can be trusted at all when scope and billing live apart is covered in the article on why project profit is wrong when scope and billing live in different systems.