Cleaning up your finances before a Series A

Investors rarely walk away from an interesting investment because the numbers are bad but they will walk away if the numbers keep changing.

A business with thin margins and an honest explanation can raise money. A business with good margins that reports a different revenue figure in the deck, the model and the management accounts cannot. The problem is no longer the margin. It is that nothing you say can be relied on.

Confidence is the asset being protected in a raise. Everything below is about protecting it.

The first test is speed. An investor asks for management accounts, a cash flow forecast, a cap table and the financial model. If any of those takes more than forty eight hours to produce, that is a problem to fix before outreach begins, not during it. The delay is the signal. It says the information does not exist in a usable form and is being assembled on request, which raises the question of what it was assembled from.

The second is that the accounting system has to be the source of truth. Reconciled, current, with a chart of accounts that reflects how the business actually makes money rather than how the accountant set it up three years ago. Revenue split by line. Costs split between what is genuinely cost of sale and what is overhead.

If your real numbers live in a spreadsheet and the accounting system is where compliance happens, you have two sets of books. You will be caught in the gap between them.

Revenue recognition is where most early-stage businesses have a genuine problem rather than a tidiness problem.

Cash received is not revenue earned. An annual subscription paid upfront is twelve months of revenue and eleven months of liability on the day it lands. A project invoiced on signature is revenue as the work is delivered, not when the invoice goes out.

Businesses that get this wrong are not necessarily being dishonest. They are reporting on a cash basis while describing themselves in accrual terms, and the two diverge most sharply in exactly the period of fast growth you are trying to raise on.

Investors check this early. Not through an accounting review, but by asking how revenue in the accounts ties back to signed contracts. If the answer requires a caveat, expect the conversation to slow down.

The next area is the one founders find uncomfortable. Related-party items, informal loans, expenses that were never quite personal and never quite business, intercompany charges between entities that were agreed verbally.

These need to be cleaned up and documented before anyone external sees the books. They are almost always innocent. They almost always look worse when discovered by a third party mid-diligence than they would have if disclosed upfront.

Then the model.

A model where revenue is a hardcoded number is not a model. It is a budget with extra steps.

It should be driven by the things that actually move: volume, price, channel mix, conversion, churn. An investor who wants to test a scenario should be able to change one assumption and watch the model respond coherently. If changing a price assumption breaks the file or requires you to update six other tabs by hand, the model will not survive contact with diligence.

The model also has to tie to the accounts. Historic actuals in the model should match what the accounting system says, to the pound. When they do not, the usual cause is that the model was built in isolation and never reconciled back. Every projection built on top of it is now anchored to a number that was never true.

None of this is difficult work. It is unglamorous, it takes six to eight weeks, and it has to happen before the first conversation rather than after it.

A raise moves fast once it starts. There is no quiet period in the middle to fix the foundations.

The founders who raise smoothly are not the ones with the best numbers. They are the ones whose numbers say the same thing every time they are asked.