How many investments make an ideal search fund portfolio?

As an investor managing a fund of search funds, I spend a lot of time reflecting on portfolio construction. A question that I think hard and often about, is portfolio sizing. How many SF acquisitions should we invest in if we want to maximize our chances of beating the 30%IRR benchmark from the Stanford and IESE studies?

Portfolio construction requires weighing the benefits of diversification against the investor’s ability to effectively support entrepreneurs and help them find, acquire and operate great businesses. The amount of capital available to invest, investor bandwidth (including board capacity), the opportunity set (having access to quality deal flow), fund mandate restrictions or even tax implications, are also limiting factors that often drive portfolio sizing decisions.

Let’s take the example of a fund of SFs like Istria. Fund size inevitably plays an important role in determining the number of portfolio investments. It is almost impossible to run a concentrated portfolio of 10-15 SF investments when you have over €20m in assets, given the small average size of equity tickets, at least in Europe. Running a portfolio of 50 companies is also challenging for a small fund with an investment team of 3-4 people, due to lack of sufficient manpower to support the portfolio companies. Having access to top entrepreneurs and deals is also key, since we know that the top quartile SFs drive the lion’s share of the asset class returns. Finally, fund mandate restrictions may also apply, such as having a maximum allocation for a given country or geographic region.

But leaving aside these limiting factors, from a modern portfolio theory approach, what is the optimal number of investments in a SF portfolio to maximize returns? The answer to this question will depend on the investor’s view about probability distributions and the distribution of SF returns. 

Probability distributions and private equity (VC and non-VC) returns

For a user-friendly explanation on the math behind probability distributions and the most important concepts applied to investing, I would suggest to read here or here (most of the relevant studies and articles belong to the VC space).

It seems to be widely accepted that early-stage VC returns follow a power law distribution, with many strike outs and a few home runs driving most of the returns (see, for example, here and here). On average, 7 out of 10 investments in a typical VC fund will lose money, 2 will make enough to cover for the losses and the remaining one should provide for all the returns of the fund. According to different academic studies, 4% of VC deals return more than 10x capital and 0,4% of deals return more than 50x capital. As a result, an optimal VC portfolio strategy should aim to assemble 500 investments, with 100 investments being the minimum. If we believe that VC returns are subject to strong power law behavior (what is known in the jargon as having an alpha < 2), then an investor would increase its expected return by investing in almost every possible deal.  

There are fewer studies on the analysis of returns distribution in traditional (non-VC) private equity, partly due to the difficulty of obtaining reliable data on the individualized gross returns of PE portfolio companies. The standard assumption in most studies (see here) is that PE returns are also rightward-skewed (although not as much as VC), forming a lognormal, a Pareto-type or some other related form of distribution. 

Are search fund returns normally distributed? Making an educated guess

The SF asset class is still too young to benefit from the quantification that has revolutionized modern finance. Despite the thorough repository of returns data gathered by Stanford and IESE, the sample size is still too small to draw any statistically relevant conclusions on SF returns distribution.

In the absence of empirical evidence, our job as investors is still to make our best educated guess based on the information that we have available and the inferences we can draw from other areas of PE. This is what I believe:

  • The Stanford and IESE data suggest that SF returns do not follow a normal distribution, but some type of right-ward skewed distribution, with the top quartile of SFs driving the lion’s share of the asset class returns.
  • Although SF investing is not an extreme outlier business like VC (the median SF investment in the Stanford and IESE data generated positive returns), a few outliers have earned a significant part of the SF asset class returns (Asurion being the best-known example), which could be indicative of some power law behavior. 
  • It is impossible to determine, with the existing data, which type of right-ward skewed distribution would fit better the SF asset class returns (lognormal, Pareto, power law…). We have seen that a few home runs have generated exceptional returns, but power law distributions are often hard to differentiate from other skewed distributions (according to the studies, you need at least 1,000 data points to confirm a power law distribution).
  • My intuition is that, even if the returns of individual SF investments do not follow a power law distribution, it is still possible that overall asset class returns exhibit some power law behavior, due to some outperforming investments having much longer holding periods that end up making the distribution’s tails heavier. In this regard, we see that some of the most significant successes in the US had holding periods over 10 years (e.g., AsurionServiceSourceAlta Colleges, MedMart).
  • The data suggest (yet again, I doubt there is statistical evidence to support it) that there is a positive correlation between returns in the top quartile (and a negative correlation with those in the bottom quartile) and those companies that meet certain economic investment criteria – company growth in an industry with favorable tailwinds, strong free cash flow generation (asset light business models often with negative working capital) and a high degree of recurring revenues.
  • My intuition is also that investors may experience different distributions of expected returns when investing in so-called “growth deals” (e.g. niche SaaS software businesses in industries growing at 20% CAGR, often acquired at 6-8x EBITDA although more and more measured in multiples of ARR), compared to more traditional “free cash flow deals” (stable businesses growing at 5-10%, acquired at 4-5x EBITDA).

Some key takeaways for portfolio sizing and portfolio construction

Based on my educated guesses described above, here are some key takeaways for portfolio construction:

  • Since I am highly convinced that SFs do not follow a normal distribution, I believe that running a concentrated portfolio of 15-20 SF investments does not make sense (from a portfolio theory perspective). I don’t believe investors have the ability to pick winners, neither do I believe that an investor’s increased dedication to a smaller portfolio of companies will end up yielding better returns than a more diversified approach.
  • Since we cannot know (today) whether SF returns follow a power law or some tamer form of right-ward skewed distribution, I believe that the optimal portfolio size should be anywhere between 35 to 100 investments. A smart question for a SF investor to think about would be: what is the chance of getting a 10x investment or a 50x investment in my portfolio? 
  • The diseconomies of scale associated to SF investing makes it hard for most investors to run such a diversified portfolio of investments. Unless you are a passive investor (which most SF investors are not), diversification brings significant operational challenges, mostly due to manpower resource constraints.
  • I believe that, for a small fund like Istria, running a portfolio of at least 30-35 investments increases your chances of investing in some of the top performers, while staying reasonable in terms of number of companies given the size of the team. I also believe that your chances of hitting a home run also increase if your holding period is flexible and can be extended up to 10 years, allowing you to ride the winners for a longer period than the regular 6-7 year hold. 
  • At Istria, we see other additional advantages in running a bigger portfolio size. It increases our profile as an investor, which helps improve our deal flow. Evaluating and investing in more companies also gives us more data to build our pattern recognition capabilities (improving our investment process), and it helps expanding our network.
  • It is important for SF investors to avoid adverse selection and build up a brand/reputation that allows you to have access to the top 25% of the deals. It also makes sense to run a barbell strategy, having a combination of “free cash flow deals” and “growth deals”, and always focusing on high quality teams and SF acquisitions that meet the economic investment criteria.
  • An individual SF investor should consider whether he or she would be better off investing through a fund of SFs. Although direct investing in SFs is fun and rewarding, unless you are a professional serial SF investor, building a diversified portfolio of direct investments is hard and requires a high level of dedication and human resources (and definitely not the same skill-set you need for fund allocation and fund manager selection). From a pure financial returns perspective, the average individual investor might be better off investing through a fund of SFs, even after accounting for all expenses and fees.


As with other asset classes, the right SF portfolio size is ultimately a judgment call based on risk appetite/risk tolerance, opportunity set and some notion of investor bandwidth. From a portfolio approach, and in the absence of empirical data, I believe that running a highly diversified SF portfolio seems to be a better option, since it increases the chances of capturing some of the outliers falling to the right of the bell shaped curve.

This article first appeared on 02.06.2021 on SF Thoughts, at

Leave a Reply

Your email address will not be published. Required fields are marked *