FJ Labs’ Liquid Crypto Data Due Diligence

How FJ Labs uses data to make capital allocation decisions in the liquid crypto market.

As discussed in a previous post, FJ Labs has allocated $30m of the core fund to a liquid crypto strategy. Our investment process is comprised of both an evaluation phase and a data due diligence phase. The below post outlines how we approach the data aggregation and discovery.


After we have discussed an asset in our investment committee, we move it from the “Research” phase to the “Data Due Diligence” phase of our process. Before creating our dashboards for each vertical, we first define which key metric categories matter if we want to track growth and competitive position within a given vertical.

To do this, we regress a robust set of independent variables against the change in asset price for every asset in a given vertical. These independent variables are shown below, and comprise both macro and project specific factors.

*Note – the below analysis provides only a high-level example of our strategy. In practice we use a different, far more robust, set of independent variables for each vertical.

Macro Data Set

Crypto Metric Data Set

The output of these regression models provide insight into which fundamental metrics are most significant for each crypto vertical we invest in. These variables then inform which data we track for each of our crypto vertical dashboards.

Real-Time Data Insights

In the crypto space, most code is open source and everything that happens on the blockchain is transparent (barring zk and other technical obfuscation). This means that, with sophisticated data science capabilities, we can aggregate the most relevant metrics for each vertical we invest in, and track these metrics over time to better understand how use of different blockchains and decentralized applications evolve over time.

For each vertical, we create a suite of data dashboards, which provide a comprehensive view of activity for that vertical, and a fundamental source of truth upon which we can base our investment and trade management decisions.

Below we will provide an example of how we use these dashboards to make data-driven decisions around capital allocation and investment sizing for the “Layer 2” vertical.

*Note – The below data dashboard case study focuses on the Layer-2 vertical for Ethereum. After completing this dashboard build, we realized the data insights would be valuable to the broader Ethereum community. We applied for and received an Ethereum Foundation grant to build out the interface and open-source the dashboard. The resulting interface can be found at Enjoy!

A Layer 2 Case Study

With the proliferation of Layer 2 execution environments, decentralized applications that were formerly built on the Ethereum base chain have migrated to Layer 2 execution environments for faster settlement times and lower fees.

One type of Layer 2 execution environment is “Optimistic Rollups”, the leaders of which are Arbitrum and Optimism.

The first step in our analysis involves aggregating and cleaning data across Ethereum, Arbitrum, and Optimism to create a high-level view of how these different ecosystems are growing or contracting relative to one another.




Deriving Insights and Making Capital Allocation Decisions

Using the insights from these dashboards, we can determine which of these layer 2 ecosystems is growing fastest across the metrics that we think matter for this vertical. Once we make an investment decision, we can also track these investments and compare their fundamental metrics to other assets in the same vertical. Through this process we can stay exposed to the projects with the most promising metrics in each vertical we invest in.

Expanding Our Analysis Across All Ecosystems

Using the same process shown above for Layer 2s, we have built a set of high-level key metric views and fine-grained analysis dashboards for all of the verticals we track across the crypto space (currently 19 verticals).

As the space grows, and activity gravitates toward different ecosystems and execution environments, we believe we will be able to make more rigorous, data-driven capital allocation decisions in advance of price movements across multiple verticals.

Masterclass with Everything Marketplaces

I had a fun group chat led by Mike Williams on Everything Marketplaces.

We covered:

  • My previous operator experience & key learnings from OLX.
  • An overview of FJ Labs as a leading early-stage venture fund that’s backed over 1,000+ startups.
  • What we look for in marketplaces that we invest in.
  • Marketplace metrics & benchmarks for the Seed stage when fundraising today.
  • Trends with the rise of B2B marketplaces.
  • The biggest mistakes that early-stage marketplace founders often make.
  • Fundraising tips.
  • Much more!

0:06 Intro
1:19 Fabrice’s background
4:09 Fabrice’s experience scaling OLX & key learnings
8:48 An overview of FJ Labs as a leading early stage venture fund
10:42 The metrics & benchmarks for Pre-Seed & Seed stage marketplaces
13:31 Trends with the rise of B2B marketplaces & marketplace design changing
17:51 How investors evaluate B2B marketplaces at the earliest stages
19:54 The geographies that FJ Labs invests in & opportunities for marketplaces in emerging markets
21:43 Mistakes that early stage marketplace founders often make
24:28 Fundraising tips for early stage marketplace founders
27:57 Group Q&A on using supply to help drive marketplace demand
30:46 Group Q&A on B2B marketplaces starting out with a narrow vertical focus vs. being more horizontal
33:24 Group Q&A on opportunities for marketplaces to leverage AI
38:35 Group Q&A on metrics & benchmarks that investors use the evaluate marketplaces at the Series A stage
41:09 Parting marketplace advice

FJ Labs’ Liquid Crypto Evaluation Process

How FJ Labs uses a VC lens to make capital allocation decisions in the liquid crypto market.

As discussed in a previous post, FJ Labs has allocated $30m of the core fund to a liquid crypto strategy. Throughout this post we will elaborate on these unique elements and discuss how we adapt a traditional VC investment approach to allocate to this emerging asset class.

The Process

The goal of our process is to take in the total investable landscape of crypto assets and output the best possible diversified portfolio of high quality, early-stage crypto projects and protocols. Once we make an investment, we monitor these projects real-time through a suite of proprietary dashboards and make changes to our portfolio based on persistent growth or contraction in theses underlying metrics.

This consists of three distinct phases: 

  • Research (purple)
  • Data Due Diligence (blue)
  • Investment (green)


Currently there are 22,932 unique crypto assets, 99% of which are (or should be) entirely worthless. To reduce this total investable landscape, we first apply a coarse filter that truncates the list based on quantitative metrics such as market cap, fully diluted value, token issuance, trading volume, and token distribution. We then qualitatively backfill this list with compelling early-stage tokens that might have been removed by our coarse quantitative filter.

Vertical Segmentation

We then segment the token landscape into distinct verticals, around which we have an internal thesis. For example, if we are interested in exposure to DeFi DEXs, we will aggregate projects such as Uniswap, Curve, Sushiswap, Orca, Pancakeswap, and Balancer, amongst others. This delineation accomplishes a few goals:

  1. It lets us look at a vertical in aggregate, and select the most compelling projects in each vertical.
  2. It provides the competitive landscape against which we can track growth of specific projects, intra-vertical.
  3. It provides the basis of our regression analysis and data dashboarding (more on this later).


At this point, we have a rough idea of which projects warrant a deep dive for each unique vertical, and we proceed with thorough diligence. For each token, we conduct a 10-20 page writeup that follows a formulaic template to comprehensively asses the project. To date we have done 97 10-20 page token writeups.

Investment Committee

After we conduct all writeups on a certain vertical, we organize them and pitch the most compelling projects to the FJ Labs Token Investment Committee. Since we assess all projects for a certain vertical together, we can more easily compare the merits of each. After debating the team, product, traction, value accrual mechanism, and token economics of these projects, we move toward a final decision:

  1. Should we invest in this vertical?
  2. If yes, which project(s) in this vertical are most compelling?
  3. What are the unknowns and what further research (if any) must be done?
  4. What are the key metrics we should focus on to track the evolving dynamics for each vertical?
  5. What is a reasonable fair price for this asset (from both top down and bottom up)?
  6. Where should we be a buyer?
  7. Where should we be a seller?

In the next post, we will outline how we add a layer of data insights to our evaluation process to both stress-test our investment decision and to manage our liquid portfolio.

The Value of Ignorance

 “They did not know it was impossible so they did it.” Mark Twain’s quote resonates with me. As much as we encourage founders to validate their startup idea in an extremely rigorous way, you do not need to validate everything ahead of time as long as you believe you will figure out how to deal with whatever challenge comes your way.


Aucland was my first venture backed startup. It was an eBay of France and Southern Europe. Prior to launching I did a fundamental analysis of eBay’s S-1 to validate product market fit. I felt that as long as the idea was proven in the US, it could be adapted to work in Europe as humans all essentially want the same thing: to communicate, be entertained and have a sense of purpose. My analysis of the S-1 gave me confidence that the business was viable. In sharp contrast to almost all the other businesses I researched, eBay had over 60% gross margin and was already profitable. This was enough for me to quit my job at McKinsey, sell my apartment, and embark on my first large scale entrepreneurial journey in July 1998 at the age of 23.

The list of the things I did not know was endless:

  • I had never raised money from VCs, did not know any VCs, or to approach them. Knowing what I know now, I realize how naïve my approach was of just sending them cold emails with a massive 80-page vertical business plan attached rather than being introduced by mutual connections and attaching a deck. As I was getting nowhere, I ended up just launching, executing, getting some visibility, and ultimately VCs reached out to me.
  • I did not realize that all the infrastructure required to launch as startup was lacking and that we essentially needed to create our own data center to host our servers. In a way it was a massive distraction from building the product of the startup I was meant to be building, but as a hardware nerd, I loved putting together the extremely powerful servers we used to host Aucland.
  • I had never recruited anyone before and made every hiring mistake possible. Not finding a front-end developer I liked, I even taught myself HTML and CSS to code the front end in a way I found aesthetically pleasing.
  • I did not expect the extent of the legal challenges we would face:
    • The French had given a monopoly on auctions to government licensed auctioneers or “commissaires priseurs” and got sued for violating their monopoly. The case went all the way to the European Union which ultimately overturned their monopoly.
    • I had not my mandatory military service because I was in the US for college and McKinsey. I had to fight off their attempt to arrest me on national interest reasons as I was employing over 100 people doing $10M per month in sales which was deemed more valuable to France than essentially cleaning toilets for 12 months for 100 euros per month (the military does not value you and makes you do extraordinarily menial tasks).
    • Incorporating in France was significantly slower and more bureaucratic than expected.
    • Letting go of wrong hires proved insanely complex as well.


Despite the Internet bubble bursting, I wanted to remain an entrepreneur and was willing to compromise on what I was building to continue my entrepreneurial journey. Venture capital had essentially disappeared from the ecosystem so I needed an idea that was capital efficient and could reach profitability rapidly. The only tech idea that seemed to be profitable at the time in Europe and Asia was selling ringtones for the Nokia phones of yore. The US market was years behind the European market. It was fragmented between countless operators using different technologies (CDMA, TMDA, iDen, GSM). Text messaging was not standard on any carrier. Even if you had a text messaging subscription you could not text between carriers. However, I had no doubt the US would follow the same path as the rest of the world and that eventually carrier billing and cross-carrier messaging would be possible. I did not know how long it would take, but that’s a risk I was willing to take. I spoke to the company best positioned to launch a competing service as they had carrier connections and had just raised $18 million, Upoc, but upon ascertaining they were not interested, I launched Zingy.

Once again, the list of things I did not know was endless. I had never listened to music before and did not even know what music we should be attempting to license. However, it was not hard to find cooler people to deal with the issue. The more fundamental problem was that the licensing mechanism for songs was unknown. In France you could go to one agency, la SACEM and get a blanket license for all songs. The US had a similar agency for mechanical rights called The Harry Fox Agency, which kept telling me they would eventually be able to license ringtones to me, but never managed to secure those rights. Worse no database existed to tell us who owned the rights to what songs. It took us years to figure out which singers worked with which song writers who worked with which lawyers and were represented by which publisher. It was an extraordinarily arduous detective task that ironically was accelerated when we made mistakes and would be sued for the statutory $250,000 per download penalty. I recall with glee the shock on the lawyer’s face when I would meet with them after they were suing me for billions expressing excitement at the fact we were finally talking and could license the song properly. We settled all the suits and ended up being the only ones fully licensed when the time came which turned out to be a massive barrier to entry, but it took years to get there.

Likewise, with no open content delivery networks we had no way of delivering our ringtones to the cell phone operators for the first few years of our operations. We literally hacked into the networks of GSM & TDMA operators in the early days to deliver our content. It took us nearly two years to get our first carrier deal – attending every trade show, establishing our presence and credibility, essentially bribing MSN to work with us until the fateful day we got a cold inbound call from Motorola which then led to a deal with Nextel and ultimately to all carriers.


After I left Zingy, I decided to return to my first love: marketplaces. After being rebuffed in my attempts to buy and/or run Craigslist, I decided to build a better, pre-moderated and mobile friendly, version of Craigslist for the rest of the world: OLX.

It was obvious that the opportunity was very large, and I had a good understanding of how to build liquidity in marketplaces. However, that knowledge was dwarfed by the scale of the things I did not know. When we launched, we did not appreciate the importance of SEO. We had raised $10 million and were able to scale traffic quickly through paid marketing. However, it’s only after we acquired another classified site, MundoAnuncio, whose entire traffic was SEO generated, that we realized the potential for free traffic that classified sites had as every ad could be indexed. When we made the acquisition, none of our content was indexed. Upon realizing the potential of SEO, we created a dedicated SEO product and tech team and essentially fully re-architected the site to be SEO friendly. While ultimately, we focused on user experience and branding, SEO alone brought us to over 100 million unique visitors per month which created the foundation for what was to come.

The other we were unaware of was the presence of a deadly competitor. Being based in New York, I had a US centric approach to the world and focused on Craigslist which was not a threat as they did not ambition to conquer the world. I had also analyzed all the markets we were going after: Brazil, Portugal, India, Pakistan, Eastern Europe, and the local competition was not relevant. However, I was unaware of the existence of a Norwegian publicly traded competitor that owned extremely successful classified sites in Northern Europe called Schibsted (now also called Adevinta). A few years in as I became aware of their existence, I assumed we would live in peaceful co-existence as they owned Western Europe and the Nordics while we owned emerging markets. Sadly, egged on by Telenor with which they created a joint venture, they attacked us in our two core markets first, Brazil and Portugal, starting a multiyear war which saw us collectively spend $500 million in marketing until we finally merged in our favor. This war sent me into the arms of Naspers as I needed the capital to fund the battle. They were great in their supportiveness and willingness to be aggressive, but it also sadly led me to lose control over my baby. 


If there is a common theme in all three stories is that there were many more unknowns than knowns when launching these three startups. That’s ok. Once validating that the opportunity is large and attractive, if you feel compelled to do something, just do it! Trust you will figure things out along the way. As the quote, usually misattributed to Goethe, goes: “Whatever you dream you can do, begin it. Boldness has genius, power, and magic in it!”