Rise of the Algorithms

6 min readJul 7, 2024

Private Markets Indexing is Coming.

Rebel Fund is an excellent Micro VC. This week, founder Jared Heyman wrote about how it uses data algorithms to select Y Combinator companies in which it invests.

He states:

the algorithm achieved a net portfolio IRR of 44–59% in the more mature YC batches from 2018 and earlier, and a 27–34% IRR in the less mature 2019 & 2020 vintages (in green)³. These are insane levels of performance, dramatically better than investing randomly in YC startups (in red), which itself is much better than an S&P 500 index average (in light dashed gray). Perhaps AI is coming for my job next!⁴

The entire essay is a fantastic read for anyone considering seed-stage investing.

He explains the many features his fund uses to differentiate one seed-stage company from another and shows which features are better predictors than others. He also shows that the results are far better than a random selection would create.

Venture data is critical to predictive algorithms. Jared has good proprietary data.

Also this week, BlackRock announced the planned acquisition of data provider Preqin for $3.2 billion. In discussing the purpose of the acquisition, Larry Fink, BlackRock CEO, talked about the trend toward private market indexing.

BlackRock said its proposed acquisition of privately held London-based alternative markets data provider Preqin could lead to the indexing of private markets, while assuring that BlackRock’s clients will also have the ability to access Preqin as a stand-alone service once the transaction closes. Fink added: “We anticipate indexes and data will be important future drivers of the democratization of all alternatives. And this acquisition (of Preqin) is the unlock.”

Given my day job as a founder at SignalRank, I was particularly drawn to this sentence:

Fink further indicated in the analyst call that this acquisition is “about driving evolution and growth in the private markets by measuring them, understanding their drivers of performance and making them more investable.” Fink also pointed out how the public markets were transformed by data, benchmarks and risk analytics. “They made markets more accessible from the developed to the emerging from stocks to bonds,” he said. “They’ve made investment performance and drivers of return much more transparent … They’ve grown durable, high-growth revenue pools that are adjacent to asset management and they have generated enormous value for clients and shareholders. Our aim is to do all of that in the far less mature data, analytics and index business for all the private markets.”

Of course, an index has to own equities in the companies in the index. Otherwise, there is nothing to “own.” So, data is only one part of the equation.

There are three others: using the data to create algorithms to select index qualifiers, accessing funding rounds in those companies, and having the capital to write checks into them. After several years of doing the first three and a year of writing checks, all are challenging.

These challenges change depending on the stage of entry. Take data and algorithms, for example. Early-stage selections will be data and algorithm-enhanced; later, they can be data-driven, and at the very late stage, selections can be data only. The earlier an algorithm can be relied on to select winning companies, the better.

Selecting good companies at seed, like Rebel Fund, is super hard. Series A is also challenging. That is why early-stage investing always needs great humans to select companies.

Selecting companies at a later stage, Series C and beyond, is easier and requires much less data, as there are far fewer companies, most of which are already de-risked. However, an index built there will not perform as well as an excellent earlier-stage index due to much of the growth already achieved.

So, the words private markets and index sound compelling, but there are many approaches, and there will be many indexes in the future. The role of data in selection will be variable. The execution is non-trivial.

There are tipping point moments in history.

The NYSE listing of the Destiny Tech 100 Index (Ticker DXYZ) marks the first catalyst. It trades at 3–4x its Net Asset Value. It combines common and preferred stock in a few later-stage companies. It has no algorithm, but it is an index approach. Investors seem to like it.

Blackrock’s acquisition of Preqin and the associated discussion is a second tipping point. It now seems that an index approach to private companies in public markets is inevitable.

At SignalRank we use Preqin’s data. We also use data from Crunchbase and have specific company data from over 150 Micro VCs we partner with, including Rebel Fund. Our scoring algorithms are mature after three versions in two years. And in a strict backtest, performance would have been outstanding. Our approach to indexing is precise. We are the market leader in building an index of companies at the Series B stage. Significant gains are still available, but risk is already reduced.

Our language differentiates “index qualifiers” from index members. The difference is we own equity in the index members. Our equity acts as a share in the index. Each shareholder owns a piece of each company. When compared to the S&P 500 and the NASDAQ, the backtest indicates the potential desirability of such an index:

The chart shows an assumed $1 investment in the S&P 500, NASDAQ, and SignalRank Index from 2014 to 2019. The results show the multiple of invested capital five years after each investment. The average SignalRank predicted return is on the red line, the median is on the black, the S&P 500 is Cyan, and the NASDAQ is green. Of course, a backtest cannot be assumed to represent future performance. And I am not giving you investment advice. Since May 2023, SignalRank has invested in 15 Series B Rounds. Only two funds have invested in more of the top-scoring B Rounds. The SignalRank Index is now a real thing. Our goal is 30 Series B Rounds annually, making the Index the most prolific B Round investor globally.

The data, access to algorithms that appear to predict outcomes far beyond the market norm, partnerships that bring access to companies that score highly, and the capital to write checks are the four requirements of successfully creating a private markets index. Since we started SignalRank, there has never been a week when I felt the world was moving to a place where the power of what I work on is recognized, and the emergence of an index for private market investments is inevitable. This week, I do. It’s a nice feeling, very motivating.

Hat Tip to this week’s creators: @jaredheyman, @PeterJ_Walker, @cartaInc, @Om, @davemcclure, @vntrcapital, @mgsiegler, @sarahfielding_, @elevenlabsio, @fredwilson, @markgurman, @e_howcroft, @UtkarshShetti, @steph_palazzolo, @runwayml, @EricHSchwartz, Cloudflare

Rise of the Algorithms

Private Markets Indexing is Coming.

Contents

Written by Keith Teare

No responses yet