A valuation is an argument with arithmetic in it.

When the number becomes the problem

• A round where the founder's expectation and the investor's model are far apart.
• A secondary sale, where an early shareholder needs a defensible price.
• An option grant or an employee plan, where the value has to be stated rather than negotiated.
• An acquisition approach that arrives with a number attached.
• A bank, a covenant or an internal disagreement between founders.

Why the argument is usually not about the methods

Both sides rarely disagree about which approach to use. They disagree about the base: which revenue is recurring, which margin is real, what the market will look like in three years. The method is the vehicle, the assumptions are the substance.

Four approaches, four failure modes

• Comparable multiples. Useful when the peer set genuinely shares business model, growth and margin. A sector label is not comparability, and a screen of listed companies says little about a private company at a different stage.
• Discounted cash flow. Useful when the business has predictable cash generation. Fragile where terminal value is most of the answer, which it usually is.
• Precedent transactions. Useful as a sanity check. Includes a control premium and the mood of the year it happened in.
• Asset-based. Useful for holding vehicles and asset-heavy businesses. Misleading for anything whose value is customers and know-how.

The range is the deliverable

A single-point valuation is a negotiating gift: it can only be attacked or accepted. A supported range, with the assumptions that move it, gives you a position and the ability to trade within it.

The four figures that matter

• Fully loaded acquisition cost, including sales time and channel spend, not just media.
• Payback period: how many months until the cash from a customer returns the cost of getting them.
• Contribution margin after the real cost of delivery, support and retention.
• Retention by cohort, because an average churn rate hides whether the early customers stay and the recent ones do not.

Why the ratio is not the argument

A lifetime-value to acquisition-cost ratio of three sounds healthy until the cohort curve flattens at month four and the payback assumption was fourteen. The ratio summarises; the cohort data decides.

Where a business serves several segments, blended economics are close to meaningless. The question is which segment is profitable and whether it can absorb the growth case.

Predictable weak points

• Peers selected to support the number rather than to test it.
• A single scenario, presented as the plan rather than the middle of three.
• Blended unit economics across a mixed customer base.
• Working capital and capital expenditure left outside the model, which is where the cash actually goes.
• No view of the dilution the next round creates, and therefore of the value that matters to the founder.

Each of these is answerable. The mistake is discovering it in the room.

What a buyer is actually buying

Rarely the current profit. Usually a position, a route to a market, a capability, or a customer base that would cost more to build. That reading is what a defensible premium has to be argued from.

The document set

• Valuation model with an assumptions register: what was accepted, what was rejected, and why.
• Sensitivity table showing which three variables move the answer most.
• Valuation memo, written to be read by the other side without your commentary.
• Negotiation brief: the target range, the walk-away point, and the argument for each anchor.
• Prepared responses to the objections most likely to be raised, including the ones about the model itself.

What we need from you

Management accounts, cohort level customer data where it exists, the current model, and any offer or term sheet already on the table.

Build, test, then argue

We rebuild the base rather than accept the existing model, because the disagreements usually start there.

Then the comparable set and the method choice, argued rather than assumed.

Then the memo, tested in a rehearsal where we take the other side.

Throughout, the range moves only when an assumption moves, and the register records it.

What we are careful about

Overstating precision. A model with two decimal places and one scenario is less persuasive than a range with three, and the reader knows it.

Where the work stops

It is not a statutory or fairness opinion, which has a defined legal purpose and a different author.

It does not determine what a particular buyer will pay for strategic reasons specific to them.

It does not structure the deal: liquidation preference, ratchets, earn-outs and drag rights are counsel's work and the term sheet's language.

It does not grow the business that justifies the number: growth strategy.

Related work

Where the number has to be defended in front of an investor, it sits inside a wider case: investment business plan.

Where this work leads

One document answers one reader, but the question behind it rarely stays inside a single direction.

• Growth Strategy

The level above

Documents solve a defined file. Where the decision itself is open, the work moves up a level, into a direction, and eventually into growth strategy.