Most raises that fail do not fail on the business. They fail because the founder pitched a Series A story with seed-stage evidence, or spent four months talking to funds whose mandate never included them, or arrived at diligence with a data room that quietly contradicted the deck.
All three are avoidable, and all three are finance problems before they are fundraising problems.
One question per round
Investors at each stage are underwriting one specific risk. Everything else is secondary, and arguing the secondary thing is how founders waste meetings.
| Round | The question being asked | What actually gets tested |
|---|---|---|
| Pre-seed | Can this team build the thing at all? | Founders, insight, speed of iteration. Often no revenue and that is fine. |
| Seed | Does anybody want it enough to pay? | Early revenue, usage depth, retention of the first cohorts, a repeatable way of finding the next customer. |
| Series A | Does it repeat without the founder in the room? | Unit economics, a sales motion that works twice, net revenue retention, churn by cohort. |
| Series B | Does it scale, and what does scale cost? | Efficiency of growth, payback period, gross margin at volume, a management team below the founders. |
| Series C and beyond | Is there a path to a public-market or strategic outcome? | Market share, durable margin structure, governance and reporting an acquirer or an exchange would accept. |
Raising at the stage you wish you were at. A seed-stage company pitching Series A funds does not get a Series A — it gets a slow no, four months later, and a reputation for being early that follows it into the next round.
Match the round to the evidence you can actually show. Raising a smaller round on honest evidence and clearing the next milestone is almost always faster than reaching for the bigger one.
What each stage buys
Capital is not an end. Each round exists to buy a specific proof, and the amount should be derived from the cost of that proof plus a margin for being wrong.
- Pre-seed buys time to build something people can use. Typically 12–18 months of a very small team. The deliverable is a working product and the first evidence that the insight was real.
- Seed buys the search for repeatability. You are paying to run enough go-to-market experiments to find one that works more than once. The deliverable is a channel you can describe, with a cost attached.
- Series A buys the first real scaling of that channel plus the team to run it. The deliverable is a sales and delivery machine that does not depend on a founder closing every deal.
- Series B buys market position. Hiring ahead of revenue, new segments or geographies, and the infrastructure to support both. The deliverable is defensible share.
Work backwards from the proof. A round sized on "what we think we can raise" rather than "what the next proof costs" tends either to run out three months early or to set a valuation the next round cannot clear.
Product-market fit, measured
Product-market fit is treated as a feeling and assessed as a number. The numbers that persuade are unglamorous.
- Retention curves that flatten. A cohort chart where usage decays and then levels off is the single most persuasive slide in an early deck. A curve heading to zero is not product-market fit however fast new signups are growing.
- Net revenue retention. Above 100% means existing customers grow faster than they leave. At Series A, above 110% is strong for SMB and above 120% is strong for enterprise.
- Concentration. If one customer is 40% of revenue, you have one customer and a pipeline, not a business. Investors will model its loss.
- Pull versus push. Inbound share, organic growth, referral rate. Growth that stops the moment you stop spending is a channel, not fit.
- Depth of use. Weekly active over monthly active, or seats expanded within an account. Logos are easy to win and easy to lose.
Traction that survives a spreadsheet
Every number on a traction slide will be rebuilt from the ledger by an analyst during diligence. If the rebuild does not match, nothing else in the deck is believed.
The usual causes of a mismatch are not dishonesty. They are definitional, and they are fixable in advance.
| Metric | Where it goes wrong | The fix before you pitch |
|---|---|---|
| ARR | Pilots, non-recurring services and one-off implementation fees folded in | Define it once, in writing, and reconcile the total to the revenue account in the ledger. |
| CAC | Founder time, contractor cost and brand spend excluded | Fully loaded, by channel, including salaries of anyone selling. |
| Churn | Logo churn quoted when revenue churn is worse | Report both, by cohort, and explain the gap yourself before they find it. |
| Gross margin | Hosting, support and third-party licence costs sitting in overheads | Put true cost of delivery above the line. A 90% margin that becomes 62% in diligence costs you the round. |
| Pipeline | Unweighted, with stale opportunities still in it | Weight by stage, age out anything untouched for 60 days, and show the conversion history behind the weighting. |
The team test
At pre-seed and seed the team is the investment, and the assessment is less mystical than founders assume. Investors are looking for evidence of a non-obvious insight earned by doing the work, for speed of iteration as a demonstrated habit, and for whether the founders have done the hard conversation — equity split, roles, what happens if one leaves — or deferred it.
Two things get checked quietly and cause real damage when wrong: an unvested or badly split cap table, and a founder whose previous employer may have an intellectual property claim. Both are cheap to fix early and expensive to discover in diligence.
From Series A onward, the question shifts from the founders to the layer beneath them. An investor is underwriting whether this becomes an institution, and a company where every decision still routes through one person does not look like one.
What VCs actually look for
A venture fund needs a small number of investments to return the entire fund. That single structural fact explains most investor behaviour that founders find baffling.
- Can this be very large? Not whether it will be. Whether the ceiling, if everything goes right, is big enough to return the fund. A good business with a modest ceiling is a bad venture investment, and this is not a judgement on the business.
- Why now? What changed — regulation, technology, behaviour, cost — that makes this possible now and not three years ago. An opportunity that has been available for a decade and untaken invites the obvious question.
- Why you? What you know that others building this do not.
- What breaks it? Founders who name their own biggest risk gain credibility. Founders who claim there isn't one lose it immediately.
- Can we own enough? Most funds need a target ownership percentage. If your round size and valuation do not let them reach it, they pass for reasons that have nothing to do with you.
Run the process properly: build a target list filtered by stage, sector and cheque size; get warm introductions where you can; and compress first meetings into a tight window so that momentum is real rather than manufactured. A raise run sequentially over six months reads as a raise nobody wanted.
The model behind the deck
The deck makes a claim. The model is what makes the claim credible, and it is the first thing a serious investor asks for after a good second meeting.
What it has to be:
- Driver-based, not grown by percentage. Revenue should be built from units — customers, seats, price, churn — not from last year plus a growth rate. The question in diligence is always "what has to be true for this number", and only a driver model can answer it.
- Three statements, properly linked. Profit and loss, balance sheet and cash flow, where the balance sheet balances without a plug. A forecast that is a P&L only cannot tell anyone when you run out of money, which is the question being asked.
- Reconciled to actuals. The model's history must tie to the filed accounts and the ledger. A model whose prior year does not match the accounts is discarded.
- Assumptions on one sheet. Every assumption visible and changeable in one place, with nothing hard-coded inside formulas. Investors will change your assumptions. Make it easy and they will respect the model; make it hard and they will rebuild it themselves, less generously.
- Scenarios, with the downside modelled. Base, upside and a genuine downside. A plan with no downside case says you have not thought about one.
- Runway as an output. Months of cash remaining should fall out of the model, not be asserted on a slide.
Model the round itself: the new option pool and whether it is created before or after the money, liquidation preference, anti-dilution, and the founders' position after this round and the next two.
A pool created pre-money is paid for entirely by existing shareholders. Founders routinely agree to this without modelling it and discover the cost at the next round. The clauses that cost more than valuation goes through the mechanics.
Diligence, start to finish
Diligence typically runs four to eight weeks from signed term sheet, in parallel streams. What determines the length is almost never the quality of the business — it is how quickly you can answer.
- Commercial. Customer reference calls, market sizing, competitive position, pipeline verification. Line up referenceable customers before you are asked.
- Financial. Revenue recognition policy, quality of earnings, the metric rebuild described above, working capital, and a reconciliation of management accounts to filed accounts. This is where most delay happens.
- Legal. Corporate records, the cap table and every instrument on it, material contracts, change-of-control provisions, employment agreements, intellectual property assignment.
- Technical. Architecture, security posture, key-person risk in engineering, open-source licence exposure.
- Tax. Filing history, any disputed positions, transfer pricing if you operate across borders, R&D claims and their basis, PAYE and GST compliance.
Three habits shorten it materially. Build the data room before the term sheet, not after. Run your own quality-of-earnings review first so you find the problems rather than the buyer. And disclose known issues early and in writing — a problem you raise is a negotiation, a problem they discover is a re-price. What a Series A data room actually contains sets out the full index.
From term sheet to money
The term sheet is largely non-binding and entirely determinative. Almost nothing in it improves after signing, so the negotiation that matters happens before, not after.
Then: confirmatory diligence, long-form documents (subscription agreement, shareholders' agreement, amended constitution), conditions precedent, board and shareholder approvals, signing, and completion. Money usually lands a few days after completion rather than on it.
Watch the conditions precedent. Anything you must deliver before completion — a key hire, a signed customer contract, a clean-up of the cap table, consents from existing shareholders — is leverage you have already given away. Keep the list short and make sure every item is inside your control.
Where New Zealand founders lose
A few patterns recur locally and are worth naming.
- Raising too late. Starting a raise with four months of runway means negotiating from weakness, and sophisticated investors can see your bank balance in the model you sent them. Start at nine to twelve months.
- Management accounts that are not investor-grade. If the monthly pack is produced six weeks late and restated each quarter, the diligence rebuild will be painful and the price will reflect it.
- Structure decided for a domestic business and never revisited. If offshore capital is the plan, the holding structure, share classes and jurisdiction need to be right before the round, not restructured during it. If you are also expanding offshore, the structuring and transfer pricing questions interact with the raise — our expansion guides for the United States, the United Kingdom, Singapore and Australia cover that ground, including whether to flip the New Zealand company.
- Treating the raise as a project that pauses the business. A founder consumed by fundraising for five months arrives at diligence with a flat quarter to explain. Protect delivery capacity and give the process to someone whose job it is.
The round is won in the two quarters before it opens. By the time you are in the room, you are presenting a position you already hold.
XLCFOWe build the model, the data room and the metric definitions that survive diligence, and sit alongside founders from first meeting to completion. Mandates from under $1m to $200m+.
This is general information about a commercial process, not legal, tax or financial advice, and not a recommendation to raise capital or to accept any particular terms. Round structures, valuations, instruments and timelines vary widely by sector, stage and market conditions. Confirm your own position with qualified legal, tax and financial advisers before acting.
