Valuing a private company is a problem finance never fully solved, so we built the tool that could.
The capital markets’ own discipline for pricing risk, turned on private businesses and built into software: each one valued on its own risk and benchmarked against real peers.
Location / capital-markets photographyValuing a private company is one of the problems finance has never fully solved, and the proof is hiding in plain sight. When two Stanford researchers took 135 of the world’s most valuable venture-backed companies and valued them properly, accounting for the preferential terms behind each headline number, the reported valuations came out on average 48% too high, and almost half of the companies were not really worth the billion dollars that had made them unicorns. These were not obscure businesses valued by amateurs. They were the most scrutinised private companies in the world, priced by the most sophisticated investors in it, and the numbers were still wrong by half. If valuation breaks there, it breaks everywhere below.
It breaks because of how private valuation is actually done. The common methods price a business by analogy: a comparable, meaning whatever a roughly similar company last raised or sold at; a venture-capital rule of thumb worked back from the return an investor needs; a headline post-money figure that multiplies the latest share price across every share outstanding, as though the newest preferred stock and the founder’s common stock were worth the same. None of these look at the individual business long enough to price its individual risk. Two companies with identical revenue can be nothing alike once you weigh their margins, their growth and how long they have survived, yet a comparable treats them as interchangeable. The founder walks in with a number they cannot defend, and the investor commits real money against a figure that was never a measure of the risk they are taking on.
- Surveys & questionnaires
- Desk research
- Financial modelling
- Forecasting & projection
- Scenario & sensitivity analysis
- Framework development
- Investment risk analysis
- Profitability analysis
- Liquidity analysis
- Credit & solvency analysis
- Growth & momentum analysis
- Trend analysis
- Benchmarking
- Pre-revenue / early-stage valuation
- Going-concern valuation
- Listed / market valuation
- Asset & intangible valuation
- Solution & mechanism design
- Investor alignment
- Synthesis & report writing
- Tool & model building
- Knowledge-product & toolkit creation
Judged against real peers. Not a rule of thumb.
The gap was never missing data. It was a missing method.
Underneath the methods sits the real fault line: the private and public halves of the capital markets never learned to value companies the same way. In listed markets there is genuine discipline, a company valued by projecting what it can earn and pricing those earnings against its risk and its cost of capital, on machinery built from decades of market data. That machinery does not reach a private company: it leans on a traded share price to read how risky a company is, and a private business has none; it assumes a diversified investor who cares only about market-wide risk, when the founder with everything riding on one business, or the investor taking a concentrated illiquid stake, carries the specific risk of that one company as most of what they bear.
The result is two systems that do not reconcile, and the seam pulls apart whenever a company crosses from one to the other: valuations reset downward, in the down rounds that now run at close to a fifth of all financings and in the markdowns that catch up as companies meet public-market discipline. More data was never going to close this gap, because the gap is not missing information. It is a missing method: a way to price a single private company’s own risk that still reconciles with how the disciplined markets value everything else. Pricing risk in markets that do not price it well is exactly the kind of problem we are hired to solve for others. iValuation is what happened when we turned the same lens on a problem of our own.
How we pulled it together
We calibrated on real companies, failures included.
A benchmark is only as good as what stands behind it, and most valuation datasets are quietly broken: they hold the survivors, the companies that made it, because those are the ones still around to measure. Real economies are mostly small businesses, and most of those struggle or fail. We assembled and cleaned the financials of well over a hundred thousand real companies across every major industry and deliberately kept the failures in, so the benchmark knows what unhealthy numbers look like as well as healthy ones.
We built the full financial model before valuing anything.
A valuation is only as sound as the projections beneath it, and the quick methods have none. So the engine does the accountant’s work before the valuer’s: it turns a set of straightforward inputs into a complete, coherent set of financial statements, an income statement, a balance sheet and a cash flow, and values the business off that model rather than off a multiple applied to a single number. The projections stand as a deliverable in their own right.
We corrected comparables rather than trusting them raw.
The market values a private company by comparables and stops there, treating every company in a sector as interchangeable. We didn’t throw comparables out; they carry real market information. We overlaid risk on top of them. Each business’s own risk, read from its own financials, adjusts the comparable up or down, so two companies that look alike but differ underneath no longer receive the same number. Because that risk is priced the way the disciplined markets price it, the valuation holds up as a company matures.
We built it with no fudge factors, so the valuation is genuinely arm’s-length.
The temptation in any valuation model is a tuning constant that quietly nudges the answers to where someone wants them. We refused it. Every input is either a measurement taken from the business’s own numbers or a figure drawn from real data; none of it is a dial set by hand. That is what makes the output a genuinely independent, arm’s-length valuation, the same whoever runs it and impossible to steer, which is what a valuation must be to carry weight with an investor, an acquirer or a regulator.
What we delivered
iValuation needed several kinds of expertise that rarely sit together: a valuation methodology rigorous enough to answer to a capital market, a financial-analysis engine to sit beneath it, the institutional risk-modelling to price a company’s risk properly, and the technology to turn all of it into secure, scalable software. No one person holds all four, so the work was owned by a small senior team, each responsible for the part they knew best.
Thierry Clarke led, carrying the valuation methodology and the capital-markets discipline behind it, drawn from years of the market-building work InvestorConnected does for development banks. Jeremy Hunt built the projection and financial-analysis engine, the accountant’s half of the tool: a Chartered Accountant whose command of financial reporting, modelling and data analytics shaped how the engine turns raw inputs into a coherent set of financials. Robert Scott brought the institutional risk-modelling, a CFA charterholder whose career pricing and stress-testing risk for the most demanding institutional investors gave the model the part that judges how solid or fragile a business really is. Yashar Soltanzadeh led the technology, a senior technology and cybersecurity specialist whose expertise in secure, cloud-based enterprise platforms turned the model into software built to run at scale.
Beneath the four, the build drew on many hands over the life of the product; the senior team owned the method, the engineering and their integrity.
Thierry Clarke
Jeremy Hunt
Robert Scott
Yashar Soltanzadeh
