The Power of Platform

Firm Headcount Research Methodology

For this research we focused our dataset on firms that identify primarily as venture capital firms in the United States and Canada with at least one fund of a 2000 to 2019 vintage of $10M USD. This data set provided us with 850 firms that we were able to acquire data for.

For the 850 firms, we worked with a data collection team to identify every employee who ever worked at the firm based on information available on LinkedIn and identified 76,072 individual employment records across those firms (after removing accounts identified as ‘bots’ or ‘spam accounts’). Employment records, based on our data set, are any one employment period on an individual’s LinkedIn profile. For example, if one individual held two unique job titles at Firm A and one job title at Firm B, and both Firm A and Firm B are included in our data set, that individual would have three total employment records in our dataset. We then used each individual’s job title to categorize their role into one of 9 job categories within venture capital (Investor, Technology & Software Engineering, Operations, Research & Data Science, Finance, Tax & Accounting, Investor Relations, Legal & Compliance, Platform, Other). For Platform roles, we further categorized job titles, when possible, into 6 sub categories, or specializations, (Talent, Marketing, ESG, Business Development, Community, Investment Operations). The categories were selected based on membership data available from the VC Platform Global Community as well as industry research and insights. These categories were assigned across 17,973 unique job titles. The assignment was done manually by the study’s lead researchers who intimately knows the venture capital profession. As part of this assignment, many of the 76,072 individuals were removed from the data set that were categorized as administrative professionals, non-permanent employees (eg. Fellows, Interns, Advisors, Externs, Entrepreneurs in Residence, Contractors, Consultants) or contained erroneous data based on the job titles and other information provided on each LinkedIn profile. The removal of administrative professionals and other similar roles was done to limit the reports research scope to the Core Team members of a firm. We also removed data on all individuals in the dataset who were not based in the United States or Canada based on their LinkedIn profile. Additional data sanitization was done to remove any duplicate records or overlapping records, for example if one individual stated to hold two full-time roles during the same time period on their LinkedIn profile. This resulted in a final data set of 29,400 individual employment records.

Based on this data, we calculated the total headcount that each of the 850 venture firms in our sample had in each year from 2000 – 2022. When an individual was identified as joining or leaving a firm mid-year, we calculated them as a fractional employee accounting for the percentage of the year they were in that role based on their LinkedIn profile. For example, if an individual worked at a firm from July 2015 – February 2017, they’d be calculated as: 2015 = 0.50, 2016 = 1.00, 2017 = 0.17. We then used the 9 job title categories to calculate what the headcount percentage breakdown was for each job title category in each year for each firm. These percentages were used as key input calculations across much of the analysis in the study.

Firm Headcount Assets Under Management Classification

After determining each firm’s annual headcounts based on the research method above, firms were further classified by assets under management per year. The AUM calculation was made primarily with data from Preqin of all fund data from 1990 to 2022. For each firm, we calculated their annual AUM by summing the total amount of capital raised by that firm across all funds (inclusive of non-core funds) in each year from 2000 – 2022 accounting for funds raised 10 or less years from each specific year. 10 years was used as the typical fund lifespan; while a trend across the industry is for funds to take longer than 10 years to fully distribute realized returns, 10 years is typically the maximum time horizon where funds collect management fees, and therefore the length determined to use when correlating fund size with firm headcount (a management fee expense). Since we calculated all headcount on a full-year basis, we did the same for firm AUM and did not account for funds raised midway through the year.

The classification used for firm AUM is slightly different than fund AUM (explained below), as it takes into account multiple funds and, therefore the average firm AUM is much greater than the average fund size. The analysis in the report includes a few combined AUM groupings, but the primary classifications that were utilized to group firms were: Less than $250M USD, $250M – $500M USD, $500M – $1Bn USD, $1.0Bn – $2.5Bn USD and More than $2.5Bn USD.

Fund Returns Research Methodology

For our research we used a collection of data sources from Pitchbook and Preqin who provide fund returns data (including Net IRR and TVPI, also referred to as Net Multiple) based on public data sources, self-reported fund data and independent research. We focused our dataset on firms that identify primarily as venture capital firms in the United States and Canada with at least one fund of a 2000 to 2019 vintage of $10M USD or more.

From the returns data available in Pitchbook and Preqin, we manually sanitized the data to focus exclusively on “Core Funds” (removing secondary and non-core funds from our data set such as Entrepreneur Funds, Opportunity Funds, Global Funds, etc.) and only included funds that were greater than $10M. We then ensured that for any data included the fund returns data (Net IRR and TVPI / Net Multiple) were reported later than 1/1/2020, according to our data sources. The data provided by these data sources include a mix of realized and unrealized returns based on the reporting date and fund vintage. This left us with a data set of 754 unique funds across 308 firms to include in our data consideration (out of the 850 firms that we attempted to identify data for). 500 of the fund performance metrics were sourced from Preqin’s data and 254 were sourced from Pitchbook. When we received fund returns data from both data providers, we used the data coverage and reporting date to select the data source with the most recent and comprehensive data to use.

All analysis in this report was calculated using a Pooled Returns calculation method, which calculated the returns for any given group of funds combining the total fund sizes and fund returns of that group on a pooled basis. This method was used to avoid the potential impact that any single outlier fund could have on the research and factored in the portfolio approach that most institutional investors take when investing across the asset class. In other words, the data we are presenting should display a historical return profile for an LP that invested in a number of different funds within a specific research grouping.

Additional analysis was conducted in preparing this report on the data based on percentiles rankings of individual funds, however that data was not included in the final published version. The data findings based on percentile of individual funds was very consistent with the trends presented in this report based on pooled returns. Viewers interested in reviewing the percentile-based analysis can request access to it by contacting the report lead researchers.

Some funds in our dataset have IRR data but do not report TVPI, and others have TVPI data and do not report IRR. We did our best to get the most complete data set across our different data sources, however when only one return method was available, we only included that fund’s returns analysis by that metric. Because of this, sample sizes may differ for IRR and TVPI calculations throughout the report.

Platform Category Research Methodology

For each of the 754 funds we calculated a “Platform % Score”. The calculation was done using the Firm Headcount Research Methodology described above. For each fund we used the firm’s average Platform headcount percentage for the fund’s Typical Investing Period (which we define as the 4-year period of the fund’s vintage year and the three years to follow). For example, if a fund was a 2010 Vintage fund and the firm had 5.0% Platform headcount in 2010, 6.0% Platform headcount in 2011, 10.0% Platform headcount in 2012 and 15.0% Platform headcount in 2013, the Platform % Score for that specific fund would be 9.0%. That percentage represents the proportion of the entire core team over a cumulative four-year basis for the fund that was dedicated to Platform.

That percentage was then used to categorize funds based on Platform Category. The three categories were defined as: No Platform, Moderate Platform and Significant Platform. The framework was designed looking at historic trends in Platform concentration at funds and how headcount percentage aligns with a firm’s Platform approach. Based on that research, it was decided that Significant Platform would be used to categorize firms greater than or equal to 10% of their core team’s headcount focused on platform and Moderate Platform would be used for firms that have less than 10% of their core team’s headcount focused on platform, but more than 0.001%. Funds with less than 0.001% of their headcount dedicated to platform in a four-year time were categorized as having No Platform.

Fund Performance Classification Research Methodology

When analyzing the data to determine trends across the report, funds were classified by numerous attributes and characteristics. In the final report, three methods of classification were included to display the analyzed findings: Vintage Year, Fund Size and Platform Focus.

Vintage  Year Classification

Each fund’s vintage year was provided by our data provider and is based on the year of first investment. If the year of first investment is unknown, the year of the final close is typically used as the vintage year. However, if a firm publicly declares via press release or a notice on their website a fund to be of a particular vintage different than either of the first conditions, the firm’s classification takes precedence.

Across the dataset, fund returns data varies significantly by vintage year, which made accurate classification of this data very important. The primary input that led to how vintage years were grouped was the pooled returns across the entire dataset (and broader research across the industry at large) to align grouping with consistent industry trends. These cyclical impacts were generally a result of macro-economic and industry wide trends that go beyond the scope of this research.

Based on that, we decided to classify the data into primary and secondary vintage year groups. The primary classification was by the 10-year periods of 2000 – 2009 and 2010 – 2019. The secondary classification was by the 5-year periods of 2000 – 2004, 2005 – 2009, 2010 – 2014 and 2015 – 2019. While there was some variety in pooled fund performance on an individual year basis within each primary and secondary classification, this grouping provided the most normalized view of the data compared across consistent peer sets.

Funds with vintage years of 2020 and greater were not included in the data set, as they are still actively within their Typical Investing Period (4-years). Additionally, those funds typically are earlier in their fund performance cycle and most of their fund performance data is based on primarily unrealized returns which may or may not matriculate. While unrealized returns are included in our analysis, limiting the vintage years included to 2000 – 2019 optimized for the most accurate view on realized and potential returns.

Fund Size Classification

Fund size data was provided by our data providers, and all funds included had at least $10M USD in capital. All fund sizes were converted and classified in USD by our data providers, when needed.

Based on the data, we elected to classify funds into three primary groups according to their fund size: Less than $250M USD, $250M to $500M USD, and Greater than $500M USD. This classification was made generally based on the industry trends and sample sizes available for the different groupings throughout the report.

Platform Focus Classification

This report uses 6 primary Platform specializations used to classify each individual’s role based on their Platform focus (Marketing, Community, Investment Operations, Talent, Business Development, ESG). Those six categories are each grouped based on their primary Platform Focus. Marketing, Community and Investment Operations are all considered primarily Pre-Investment Platform Focus areas. Talent, Business Development and ESG are all considered primarily Post-Investment Platform Focus areas.

Individual funds were then assessed based on the breakdown of their Platform headcount on a fund-vintage year basis to determine the fund’s primary Platform focus during that investing period. The analysis was done based on total cumulative employee counts. The Primary Platform Focus Categories were assigned based on this logic: