Economic Reasoning: How to Build a Financial Model Investors Believe
A financial model investors believe is not a forecast that turns out right. It is an argument they can test. It starts with unit economics, builds revenue and cost from a small set of drivers, ties every key assumption to a source, shows real scenarios, and matches the deck to the dollar. Investors know your numbers will miss. They want to see whether you understand why the business makes money.
The evidence on misses is blunt. In the 2016 working-paper version of a survey of nearly 900 venture capitalists, later published in the Journal of Financial Economics, investors said fewer than 30% of their portfolio companies meet their projections, and 31% of early-stage investors said they do not forecast company cash flows at all. So why build a model? Because it is where an investor checks your economic reasoning: what drives revenue, what each customer costs, and how much capital it takes to reach the next milestone.
This is the method behind the economic reasoning work in GETVISION ASIA's consulting practice, and the second stage of our investment readiness guide.
Key takeaways
- Build the model from drivers, not targets, so every output traces back to an input someone can question.
- Get unit economics right first, because a business that loses money on each customer only loses it faster at scale.
- Keep an assumptions sheet with sources, so an investor can challenge one input instead of distrusting the whole model.
- Show real scenarios that change specific drivers, not one plan with a flat percentage haircut.
- The model, the deck and the data room must show the same numbers, because inconsistency costs more credibility than a weak metric.
What investors actually do with your model
Investors rarely value an early-stage company with a discounted cash flow (DCF). The published version of the same survey found that the most common tool is the cash-on-cash multiple, followed by internal rate of return (IRR), with few investors relying on DCF. That tells you what the model is for.
Investors use it to answer three questions:
- Can this company return the multiple we need? They look at the revenue and margin profile in year four or five and ask what a buyer might pay for it. Our guide to valuation in emerging markets explains why the comparable set matters.
- How much capital does it take to reach the next milestone? The model sizes the round.
- Does the founder understand the business? The structure of the model shows this faster than any slide.
None of these needs the hockey stick to be right. All of them need the logic to be visible.
Start with unit economics
Before you forecast the company, define the unit: a customer, an order, a store, a hectare, a clinic. Then show what one unit earns and costs. If the unit does not work, no amount of growth will fix the model.
| Metric | In plain words | Why investors check it |
|---|---|---|
| Contribution margin | Revenue per unit minus the variable cost to serve it | Whether growth creates or destroys cash |
| Customer acquisition cost (CAC) | Sales and marketing spend divided by new customers | What growth costs |
| CAC payback | Months of contribution needed to recover CAC | How much cash growth ties up |
| Lifetime value | Contribution per period times expected lifetime | What a customer is worth |
| Burn multiple | Net burn divided by net new annual recurring revenue | How efficiently cash becomes growth |
For subscription businesses, the SaaS Metrics 2.0 guide on the For Entrepreneurs blog says the best SaaS companies have a lifetime value above three times CAC, and that profitability looks anemic when recovering CAC takes longer than 12 months. It also calls these guidelines, not laws. Marketplaces, hardware and agribusiness need their own versions, but the logic is the same: show what one unit earns before you show what a thousand units earn.
Build revenue and cost from drivers
A driver-based model calculates every output from a short list of inputs. Change an input and the whole model updates. That is what lets an investor test it.
The driver tree
Revenue is customers times price times retention. Customers come from leads times conversion. Leads come from channel spend divided by cost per lead, or from sales headcount times productivity. Costs follow the same logic: a hiring plan by role and start date, variable cost per unit, and fixed costs.
Typical revenue drivers by model:
- B2B software: sales reps, ramp time, quota attainment, contract value, churn.
- Marketplace: active buyers, order frequency, order value, take rate.
- Consumer subscription: paid acquisition, conversion, price, monthly retention.
- Agribusiness: hectares or suppliers, yield, price, share sold through the company.
- Fintech lending: loan volume, average ticket, yield, losses, cost of funds.
Do not forget cash. Receivable days, inventory, capital expenditure, tax and currency movements decide when a profitable company runs out of money.
The assumptions sheet
Put every key input on one sheet, with a source and a confidence level. An investor who disagrees with one assumption can change it and see the effect, instead of dismissing the model.
| Assumption | Source | Confidence |
|---|---|---|
| Monthly churn | Last 12 months of cohort data | High |
| Trial to paid conversion | Own funnel data, last two quarters | Medium |
| Price increase next year | Signed contracts with escalation clauses | High |
| Salary costs by role | Current offers and local salary data | Medium |
| Exchange rate path | Conservative path, stated explicitly | Low |
The emerging-market inputs
In emerging markets, three inputs deserve their own rows. Currency: model revenue and costs in the currencies they are actually earned and paid in, and report in dollars. Inflation: price increases and salary increases rarely move together. Tax regime: special regimes such as technology park residency carry conditions, so model what happens if they lapse. Founders who explain how they price country and currency risk answer the question before it is asked.
Scenarios, not haircuts
A downside case that is the base case minus 20% tells an investor nothing. Real scenarios change specific drivers and show what the founder would do in response.
- Base case: the plan, with drivers at their sourced values.
- Downside: slower sales ramp, higher churn, a weaker currency, and the cost cuts you would make.
- Upside: faster conversion or a larger contract segment, and the extra hiring it would need.
Show months of runway in each case. Carta's data put the median gap between seed and Series A at roughly two years in 2025, so a downside case that runs out of cash in 14 months tells the investor you will be raising under pressure. The downside should still reach a fundable milestone, or show exactly what you would cut to get there.
A good downside case is a plan, not a smaller number.
Consistency with the deck and data room
Every number in the deck should trace to a cell in the model. Historical figures should reconcile to management accounts and bank statements. Metrics should have one definition, written down once. The pitch deck investors read in three minutes is a summary of the model, not a separate story.
In the models we review, the most common failure is not a bad assumption. It is a deck that shows one revenue figure while the model shows another. That gap does more damage than any weak metric.
The five questions investors ask of a model
Prepare a crisp answer to each before the first meeting:
- What has to be true for the base case to happen?
- Which three assumptions move the outcome most?
- How much cash does it take to reach the next milestone, and what is the milestone?
- How did last year's actuals compare with last year's forecast?
- What happens in the downside, and what would you cut?
The fourth question catches many founders off guard. Keep last year's forecast, and be ready to explain each gap between it and the actuals. An honest variance explanation builds more trust than a forecast that happened to land.
Common model mistakes
Most weak models fail in the same handful of ways. Each is easy to fix before an investor sees it.
| Mistake | Why it hurts | The fix |
|---|---|---|
| Top-down market share | No mechanism behind growth | Build bottom-up from channels |
| Revenue without headcount | Growth with no one to deliver it | Tie revenue to sales capacity |
| Unexplained year-three jump | Looks like a target, not a forecast | Name the driver that changes |
| No working capital | Cash runs out before profit | Model receivable and inventory days |
| One currency only | Hides currency risk | Model each currency separately |
| Numbers typed into formulas | Impossible to test | Keep all inputs on one sheet |
| Monthly detail for five years | False precision | Monthly for 24 months, then annual |
The same working paper found that venture investors who do forecast typically look three to four years out. Build the detail where they will look, and keep the later years simple.
Frequently Asked Questions
It is a model in which revenue, costs and cash are calculated from a small set of operating inputs, such as customers, conversion, price and retention. Changing an input updates every output, which lets investors test the logic.
Three to five years is standard, with monthly detail for the first 24 months and annual figures after that. Investors focus on the next milestone and the shape of the business at exit, not on precision in year five.
They usually start with growth drivers, churn or retention, customer acquisition cost and gross margin, because these decide whether growth creates cash. Sourced assumptions survive this review far better than round numbers.
Yes, but a simpler one focused on unit economics, the hiring plan and the cash needed to reach the first milestones. The purpose is to show how the business will make money, not to forecast revenue precisely.
Model revenue and costs in the currencies they occur in, convert at a stated exchange rate path, and include a weaker-currency downside case. Report in dollars so foreign investors can compare.
Read next
Valuation in Emerging Markets: Why Your Comparable Set Is Wrong
Investment Readiness: The Complete Guide for Founders in Emerging Markets
SAFE, Convertible Note or Priced Round: Structuring an Early-Stage Raise
The Investor-Grade Data Room for Emerging-Market StartupsRaiseTalks turns readiness into a record: an investor-grade data room, a readiness score across Team, Product, Market and Deal, and a view of which documents investors open. Build your data room on RaiseTalks.