The value-creating finance function

Every founder knows the headline drivers of valuation: growth, market, team, momentum. Far fewer think of the finance function as a driver in its own right. But it is. Not because clean books make an average company more valuable, but because messy books make a good company look uncertain, and uncertainty dings the price.

An investor cannot pay for future performance that seems implausible. If the link between the realized figures and the forecast model is too weak, the growth story can start to fall apart. A bit of a hockey stick is expected, but if projections imply operational metrics (CAC or employee productivity, say) that are far from historical figures, the questions tend to get tougher and tougher.

Here are the key things your finance function should excel at across four main areas.

Operational finance: build a ledger & process that produces reliable metrics

Everything downstream (your forecast, runway, KPIs, the data room) inherits the quality of your monthly close. So close your books on an ongoing basis, on a fixed calendar, ideally within a few working days of month-end. A close that happens twice a year is a historical record. A close that happens every month is a management tool.

Accrue properly, every month. Without accruals, your monthly numbers are noise: annual subscription costs land in random months, you invoice your customers months after the work was performed, or a supplier invoice arrives long after the cost was incurred. These things cause margin swings and make your burn look erratic, and a jagged series makes it much harder to judge where you are heading.

Get gross margin right. After growth, it is probably the most scrutinized number. Overstating it can be the most expensive mistake you make in due diligence, because multiples are applied to revenue but justified by margin. Hosting, third-party data, payment processing, support and implementation all belong in Cost of Revenue (COR), not Operating Expenses (OPEX). So does AI usage: for AI-native products, inference is a genuine variable cost that scales with customers, and parking it under “software subscriptions” inflates gross margin and hides your unit economics. Investors will find it, and a margin corrected during due diligence is far more expensive than a lower margin disclosed up front.

Track expenses by department. Allocate every cost by department from day one. Without that, calculating CAC becomes archaeology: someone will have to re-read eighteen months of invoices the week before a term sheet, and the number that comes out is shaky at best. With proper allocation, your input for CAC, gross margin and R&D spend is handed to you every month.

Define revenue before someone else does. Never include VAT in revenue or ARR. It happens constantly when ARR is pulled from a billing system or a bank feed rather than the accounting system; it inflates every metric, and it is the quickest way to lose credibility in a first meeting. Know the difference between accounting revenue and ARR, too. Revenue is what you earned, recognized over the period you delivered: it is backward-looking. ARR is what you have contracted to earn, forward from a point in time. They answer different questions, will never tie exactly, and ARR can be defined in more than one way. Write your definitions down, apply them consistently, and be able to bridge between them on request.

Compliance: it adds nothing until it destroys something

Compliance is asymmetric. Being compliant wins you no points. Being non-compliant can end an otherwise promising fundraising or exit process.

Good investors and acquirers check, and they check specifically: filings done on time and in full, VAT returns that reconcile to the ledger, payroll done by the rules. They also check the corporate housekeeping behind them, from the cap table to the warrant program.

What kills deals is rarely the exposure itself, because the amounts are usually modest. It is the signal. Most startups are less than perfect on these things, not out of bad intentions or neglect, but because of a lack of resources, and in moderation that is fine. But at some point the line between “move fast” and “total chaos” is crossed, and that can get very expensive.

The fix is unglamorous: make sure someone takes responsibility for the “boring stuff”, and run diligence on yourself before anyone else does. Assemble the data room three months before you intend to raise, not three weeks.

Strategic finance: build a model you can explain and verify

Operational finance tells you where you have been. Strategic finance turns that into a defensible view of where you are going, and that view is what an investor actually prices.

Model runway from your accounting, not from your bank account. The bank balance flatters you. It ignores the VAT you owe next month, accrued holiday pay, the invoices still sitting in the drawer, and the annual bills scheduled ahead. Liquidity squeezes are rarely a genuine surprise, only an unobserved one, and knowing in advance which levers you would pull, and how many months each one buys you, is what keeps a bad quarter from turning into a bad round.

Wire the forecast directly to your actuals. A forecast built as a standalone spreadsheet is out of date the day it is finished. Build it on the same structure as your monthly reporting, with the same departments and the same revenue definitions as your ledger, so each closed month flows in automatically. Updating a scenario then takes an hour instead of a week, and you see forecast-versus-actual variance every month. Nobody expects your forecast to be right, but they expect you to know why it wasn’t.

Forecast on value drivers, not on line items. A model that grows revenue eight percent a month because last year grew eight percent a month explains nothing and defends nothing. Build from the drivers underneath: leads, conversion rates, average contract value, churn, sales capacity and ramp time, headcount by function. That is what makes a hockey stick plausible. A curve produced by named assumptions can be argued with, adjusted and agreed on; a curve produced by a growth percentage can only be believed or not. It also shows you which levers actually move the needle: what happens if you hire four more salespeople, if payback stretches from fourteen months to twenty, or if churn moves a point.

Define your metrics, and keep them current. ARR, NRR, CAC, payback, gross margin, burn multiple. Write down the definition of each, agree on it internally, and produce it from the same source with the same formula every month. Answering a metric question in the meeting from a number that has been produced identically for eighteen months, rather than pulled out of a spreadsheet the night before, puts you in a completely different position.

Advisory: know your options before you need them

The last area is the judgment layer: the difference between knowing what your numbers are and knowing what they tell you about your strategic options.

Model your valuation; don’t discover it. Before you go out, form a fact-based view of what a realistic valuation could be. Run the numbers in both directions: what a given valuation implies about the multiple you would have to defend, and what your ARR, growth rate and margin suggest against comparable transactions. Build internal scenarios of amount, valuation and dilution. A term sheet below your ask becomes a decision instead of a shock, because you already know what each alternative costs you.

Look at the cap table as a timeline, not a snapshot. This round is probably not the last one, so model ownership forward across the rounds you expect. Check in along the way to see whether your realized performance still points to the valuation you are hoping for at the next one, and set expectations accordingly. Founders who optimize a single round for the highest headline valuation regularly leave themselves with limited room to maneuver later.

Know when to change course. Occasionally the right advice is to wait. Or to dial back and shoot for break-even rather than the next VC round. This is where solid reporting and modeling must be paired with market insight, experience and the guts to pause and ask the really tough questions.

Related reading: Raising at the right valuation, with what Danish software companies were priced at in 2026, and Improving your exit valuation.

Originally published in Pre-Money, Issue 01 (Crispa, 2026), p. 10. Figures are as reported at the date given with each chart; medians describe a market, not any one company, and nothing here is advice on the price of yours.