What do you do when you gave investors a wrong number?

You reported ARR of ₹8.4 crore in the last investor update. While preparing for diligence, somebody notices that two implementation fees were counted as recurring, and one customer who gave notice in January is still in the base. The real figure is ₹7.6 crore.

Nobody was careless. The definition was never written down, two people prepared the number in different months, and it drifted. It happens in most companies at least once.

What happens next is entirely within your control, and it is one of the few moments where a founder can materially change how they are perceived in either direction.

Why the correction is cheaper than the alternative

The instinct is to wait. Perhaps the number will catch up. Perhaps it will not come up. Perhaps you can correct it quietly next quarter without drawing attention.

Three reasons that reasoning fails.

The first is that it will be found. Diligence rebuilds every material metric from raw data, and a discrepancy against something you reported is exactly what that process is designed to surface. The question is not whether, it is who says it first.

The second is the difference in what it means. An error you disclose is a process problem: your reporting had a gap, you found it, you fixed it. An error someone else finds is a trust problem, because the immediate question is what else has not been checked. Those two are not close in cost.

The third is compounding. A wrong number gets built on. It goes into the next update, the deck, the model, the board pack. Correcting it after three months means correcting four documents and explaining a trend that was never real.

An error you disclose is about your process. An error they find is about your character. The gap between those is the whole reason to move quickly.

Before you send anything

Two hours of work, and it prevents the much worse outcome of correcting a correction.

1. 01Establish the right number properly. Not approximately. Rebuild it from source data with the definition written down, and have someone else check it.
2. 02Find out how far back it goes. If the definition drifted, the error probably exists in earlier periods too, and you need to know before you say anything about the trend.
3. 03Work out what else it affects. A wrong revenue figure moves growth rates, retention, margins, CAC payback and probably the model. List everything.
4. 04Understand why it happened, specifically enough to explain in a sentence. No written definition, two preparers, a manual step, a formula that broke. The cause is what determines whether your fix is credible.
5. 05Decide what you have changed so it cannot recur.

Do all five before sending. A correction that turns out to be incomplete is considerably worse than the original error, because it removes the one thing you were trying to demonstrate.

How to write it

Its own message, sent to everyone who received the wrong figure. Not buried in the next monthly update, and not delivered verbally to one investor.

Six parts, in this order.

1. 01What was wrong, in the first line. Not a preamble. The number we reported was X; the correct figure is Y.
2. 02Which periods and which documents it affects.
3. 03Why it happened, in one or two sentences, without defensiveness and without excessive detail.
4. 04What else moves as a result. All of it, including anything that improves.
5. 05What you have changed so it does not recur.
6. 06An offer to walk anyone through it.

Short. A page at most. Over-explanation reads as anxiety, and anxiety makes people wonder what else there is.

One thing to avoid: presenting the correction as though the business changed. The business did not change; what you knew about it did. Conflating the two makes it sound like you are managing the message rather than correcting the record.

A worked example

Illustrative, using the situation at the top of this piece.

We reported ARR of ₹8.4 crore in the September update. The correct figure is ₹7.6 crore. Two implementation fees totalling ₹52 lakh were counted as recurring, and one customer worth ₹28 lakh who gave notice in January remained in the base. The same error affects the June and July updates, where ARR was overstated by ₹31 lakh and ₹44 lakh.

Then what else it moves: growth for the quarter falls from 11% to 7%, net revenue retention is a point lower, and the model has been rebuilt from the corrected base.

Then the cause and the fix: ARR had no written definition and was prepared by two different people. The definition is now documented, excludes one-off fees and customers under notice, and one person owns the figure with a second checking it before it goes out.

Then the offer: happy to walk through the rebuild with anyone who wants to see it.

That is seven sentences and it does everything the situation requires. The founder who sends it has demonstrated that they check their own numbers, which is the thing the investor actually wants to know.

What to expect back

Less than founders fear. Most investors have seen this before, and their reaction is usually some version of thanks for flagging it.

What they are assessing is not the error. It is three things: whether you found it or someone else did, whether the fix addresses the cause or only the symptom, and whether you told them promptly.

Expect some questions, and expect them to be about process rather than about the number. What else is prepared the same way. Who checks the pack. Whether the other metrics have written definitions. Those are fair questions and the right answer to all of them is the truth.

A small number will take it badly, and a pattern of corrections will genuinely damage confidence. One correction, handled well, very rarely does.

The situations that need more care

• ·The wrong number was in a fundraising deck or a data room. Correct it to everyone who saw it, including investors who passed, and tell your lawyer before you send it, because representations may be involved.
• ·A round is in progress. Tell the lead immediately and in advance of the written correction. Discovering it mid-diligence is the worst version of this.
• ·The error is in audited accounts or a statutory filing. A different process entirely, and it goes through your auditor and company secretary before it goes anywhere else.
• ·The error benefited you materially in a negotiation that has closed. That is a conversation to have with a lawyer first.
• ·It is not an error but a definition change. Say so plainly, restate the history on the new basis, and show both. A definition change presented as a correction, or the reverse, is worse than either.

When you are not sure it is wrong

Sometimes the position is genuinely ambiguous. A metric could reasonably be calculated two ways, the earlier preparer chose one, and you now prefer the other. Is that an error?

The useful test is whether the figure you reported was defensible at the time on a definition you could state. If it was, you have a definition change rather than an error, and the handling is different: say that you are changing the basis, explain why, and restate the history on the new basis so the trend remains readable.

If it was not defensible, or if nobody can now say which definition was used, treat it as an error. The ambiguity is itself the finding, and saying so is more credible than constructing a justification after the fact.

The situation to avoid is the one where a company quietly adopts whichever definition flatters the current quarter. It is rarely deliberate and it is visible from outside within three quarters, because the trend does not behave the way a real one does.

The four that come up most

Most reporting errors in early-stage companies are one of four, and knowing them is most of the prevention.

• ·ARR including things that are not recurring: implementation fees, one-off services, usage above a committed minimum, or customers under notice. The most common by a wide margin.
• ·Revenue recognised on invoice rather than on delivery, which overstates every month in a growing business and shows up the moment an auditor looks.
• ·Cash reported without netting statutory dues already collected. GST, TDS and the employee share of provident fund sit in the account and none of it is yours.
• ·Headcount and cost figures that exclude contractors, so the cost base and the people number tell different stories.

Each has the same root: a definition nobody wrote down, applied by more than one person over more than one quarter. The fix in every case costs an hour and happens before the error rather than after it.

Preventing the next one

Almost every version of this has the same root cause: a number that nobody owned and nobody had defined.

1. 01Write down the definition of every metric in your investor pack. One paragraph each, with the exclusions named. An hour of work that prevents most of this.
2. 02Give each metric a single owner, and have a second person check the pack before it goes out. Ten minutes, and it catches the large majority of errors.
3. 03Keep the definitions fixed. When one genuinely has to change, restate the history alongside it rather than applying the change from this month forward.
4. 04Reconcile the metrics to the accounts quarterly. Most reporting errors are visible the moment you try to tie a metric back to the ledger.
5. 05Keep the working. If you cannot rebuild last quarter's figure from source, you cannot check it either.

Telling the team

The investor correction gets all the attention and the internal one matters more, because the team is where the next error either happens or does not.

Say what was wrong and what changed, without making an example of whoever prepared it. Almost every reporting error is a process failure wearing a person's name, and treating it as an individual mistake guarantees the next one gets hidden rather than raised.

The thing to establish is that flagging a suspected error is always the right move, even when it turns out to be nothing. A company where people check quietly and say nothing unless certain will find its errors in diligence, because certainty arrives late.

It is worth being specific about the fix in the same message. A definition written down, an owner named, a second pair of eyes before anything goes out. That is what makes the point land as a change rather than a reprimand.

The wider point

Investors are underwriting judgement as much as a business, and judgement is hard to observe directly. What they have instead is a series of small signals, most of which come from how a founder handles things that go wrong.

A correction sent promptly, with the cause named and the fix in place, is one of the clearest positive signals available. It says the numbers are checked, that problems surface early, and that nothing is being managed around.

Which is why the founders who handle this well often end up better regarded than if the error had never happened. That is not a reason to make mistakes. It is a reason not to hide the ones you make.

About Simplify

Simplify is a finance clarity and investment readiness practice working with founders across India, built on six years inside startups. We write about the questions founders bring before a decision, not after it.