It is crucial for startup founders to accurately estimate their venture’s cost of capital.

The cost of capital is featured prominently in the formula that’s used to calculate the lifetime value of a customer:

\[ LTV = \dfrac{m}{1 + i - r} - COCA \]


  • \(m\) is the cash flow generated by a customer contract per period
  • \(r\) is the customer retention rate
  • \(i\) is the startup cost of capital
  • \(COCA\) is the cost of customer acquisition

A startup must only pursue customers in a market segment where it can eventually generate a positive LTV.

The cost of capital in the LTV formula has the same order of magnitude as the retention rate. Therefore, it is amazing that so many founders obsess over the retention rate and choose to entirely ignore their cost of capital.

The cost of capital is also an important factor in estimating the present value of a startup:

\[ NPV = \sum_{k=0}^{T} \dfrac{C_k}{(1 + i)^k} \]


  • \(C_k\) is the cash inflow-outflow during a single period \(k\)
  • \(T\) is the number of periods
  • \(i\) is the startup cost of capital

What is the present value of your company? An answer to this question determines the share of the company's equity that you will be giving up in the next round of funding.

How does one estimate the cost of capital? Going by the book, one could use the capital asset pricing model (CAPM):

\[ ER_i = R_f + \beta_i(ER_m - R_f)\]


  • \(ER_i\) is the expected return of the investment
  • \(R_f\) is the risk-free interest rate
  • \(\beta_i\) is the beta of the investment
  • \((ER_m - R_f)\) is the market risk premium

The expected return produced by the model is the investment’s cost of capital.

To estimate the value of \(\beta_i\) in the formula, one could look for a public company that is “similar” to the new venture and use its unlevered beta as an estimate. When building a social network, for instance, one could use Facebook's unlevered beta, which happens to be 1.2. Assuming the expected market return \(ER_m\) of 10% and the risk-free interest rate \(R_f\) of effectively zero, the CAPM model produces 12% as the investment's expected return.

Using this number as the startup cost of capital would be a mistake, however.

Intuitively, investing in a startup is much riskier than investing in an established public company, such as Facebook. Startup investors usually want to get much more than 12% in return for bearing the risk. How much more?

The answer depends on the type of investors.

Consider a $100M venture fund, for instance. The fund must generate a return of at least 12%. Otherwise, the fund’s limited partners (LPs) would be better off investing in Facebook. LPs must also be compensated for the illiquidity risk associated with tying up their money in the fund for 10 years or longer.

Let’s assume the final number to be 15%. This is the venture fund’s hurdle rate. In 10 years, the fund must return at least \(1.15^{10} = 4\)x of the original $100M investment, that is $400M.

Let’s also assume that the fund invests in 10 startups, $10M per startup, for the average of 25% of the company value in non-participating preferred shares. To produce the desired return, each startup must exit at $400M/0.25/10 = $160M or 16x of the original $10M.

In reality, the majority of the fund's investments will fail or exit at low multiples, and the desired return will be generated by only one or two companies. With two exits, each company must be valued at 80x of the original investment. VC’s use 100x as the rule of thumb.

As a consequence, every VC investment must potentially be able to return 100x in 10 years. This yields the cost of venture capital of \(100^{1/10} – 1 = 0.58 \) or 58%.

58% is the 10-year geometric mean. The true cost of capital is higher at an earlier stage, some say as high as 80%, and lower at a later stage in the startup lifecycle.

58% is a big number with important implications. For instance, if you are raising $2M from a venture fund by selling a 20% stake in your company, you must convince the VC that your company can be valued at  $2M\(  \cdot 1.58^2/0.2 \) = $25M in two years.

This number can also turn a promising customer segment into a segment that should rather be avoided. Just remember that for every dollar that you spend on customer acquisition today you will need to return at least $2.5 in two years.