Time and cash
DPI, TVPI and IRR
The three numbers used to describe venture performance, what each one hides, and why only one of them is money you can spend.
Venture performance is quoted in three metrics, and the differences between them explain most of the confusion when investors compare notes. Two of them measure paper value and one measures cash.
DPI — distributions to paid-in capital — is money actually returned divided by money actually invested. It is the only one of the three that is real. A DPI of 1.0 means you have your capital back.
TVPI — total value to paid-in — adds the estimated value of what you still hold. IRR is the annualised rate of return, which incorporates timing and is exquisitely sensitive to it. Both are useful and neither is money.
What each one measures
DPI is cash out over cash in. It is unambiguous, it cannot be flattered, and in the early years of a portfolio it is close to zero — which is why it is the least quoted of the three.
TVPI is DPI plus the current estimated value of remaining holdings, over cash in. It is the number most often quoted because it is the largest, and it depends entirely on how the remaining holdings are valued. Since private companies are marked at their last financing price, TVPI is a statement about what other investors recently paid rather than about what anything is worth.
IRR annualises the return using the timing of cash flows. It is the right metric for comparing investments with different durations, and it is easily distorted: an early small distribution can produce a spectacular IRR that says nothing about the eventual total.
| Metric | Calculation | Result | What it means |
|---|---|---|---|
| DPI | £60,000 ÷ £150,000 | 0.40× | You have 40% of your money back |
| TVPI | (£60,000 + £240,000) ÷ £150,000 | 2.00× | Paper value, dependent on marks |
| IRR | annualised over six years | varies with timing | Sensitive to when cash moved |
| Spendable today | £60,000 | — | The only real number |
Illustrative. £150,000 invested, £60,000 distributed from two exits, remaining holdings marked at £240,000.
Why marks are unreliable
A private company is valued at the price of its most recent financing. That price was negotiated between the company and one investor, for preferred shares carrying rights your ordinary shares may not have, at a moment that may be two years ago.
It therefore says nothing about what a buyer would pay today, and it ignores the liquidation preference stack entirely. A position marked at £240,000 in a company with a preference stack larger than any plausible exit is worth considerably less than the mark, and the mark will not say so.
The practical rule: treat TVPI as a rough indicator of whether things are going in the right direction, and treat DPI as the truth. An investor quoting a strong TVPI in year four has told you what other people recently paid for shares in their companies.
What to track for yourself
Keep it simple: total invested, total distributed, and a list of surviving positions with their last round price and the date of it. That is enough to know where you stand and it takes an hour a year.
Adding the date of each mark is the detail that matters most. A portfolio whose marks are all three years old is being described by a market that no longer exists, and knowing that is more useful than the marks themselves.
In short
What to take away
- DPI is the only number that is money. Everything else is an estimate.
- TVPI depends entirely on marks, and marks are last-round prices for shares with different rights.
- IRR is distorted by timing — a small early distribution can produce a flattering rate.
- Track the date of every mark. Stale marks describe a market that has moved on.
From the decoder