Investing at betaworks

from the betaworks 2015 shareholder book

Why we seed invest

At betaworks, we make seed investments alongside our primary business of building companies at the studio in New York. When investing, we look for great companies that are addressing the same sorts of issues that face our studio companies, and we particularly like to invest in companies that can benefit from being connected to our network of companies and people.

At the end of 2014 we realized that it had been six and a half years since we started investing, so we decided to calibrate the performance of our investment activity from inception — separate and distinct from the value of things we are building or have built at the studio. Private technology investing tends to be more opaque than other investment categories which we think is unhealthy so I thought it would be interesting to share publicly how we are doing.

The post was written for the betaworks 2015 shareholder book — thus the wonderful illustrations by Henry McCausland. Over the coming two weeks we will publish on medium a few other essays from the book. The book itself should be out next week.

Focus and competence:

Investment wise, the stage we are focused on at betaworks is seed. Our investments reflect what we are building at the studio and our thesis in regards to investing in and around a network of loosely coupled companies creates value for the companies in that network. We try to stay within what Warren Buffett calls the “circle of competence.” The investments we made in 2014 fell into four areas:

  1. Social publishing, sharing, and discovery: Medium, Quibb, Product Hunt, Yo, and few others that aren’t yet announced.
  2. Creation and distribution: Gimlet, Medium, Beep and one other that isn’t public yet.
  3. Internet Glue: URX, Estimote and IF(ttt).
  4. Data driven tools: CrowdTangle, Hachiko, Rivet & Cuff, and Bowery.

Performance …

Since the fall of 2007 betaworks has invested in 106 companies (this count does not include companies built at the studio). Over the period we averaged 13 investments per year — over the last four years the average is 17. The target per year is 20 — the actual number depends on the quality of companies we saw in any given year and our perspective on the relative cost of investing in equity external to betaworks compared to building things at the studio (when valuations are high, we tend to build more stuff ourselves).

On average we invested $123k in each company, for a total of $13.1M since betaworks started seed investing. When we invest in a company at betaworks we allocate funds directly from our balance sheet, and we look and feel like an angel investor. We typically are first money in, we don’t take board seats and rarely do follow on investments since we do not hold reserves for our investments. Our setup is not optimal from a pure investment standpoint since leaning into winners is a cardinal rule of investing. It’s important to understand this difference since it affects the shape of our return curve.

Who we invest in …

We see a lot of companies at betaworks. What they have in common is they are great entrepreneurs who are building in the areas that overlap with our focus. Last year we met with 372 companies, over one per day and we did 14 investments, for a total of $1.6M. In 2013 we deployed $2.1M, and the average was $132k. Betaworks, the studio and our culture, are product focused, and our investments are too. We like to see and use products and betas — almost all the companies in which we invest in have working prototypes that we could evaluate and this past year about a third of them had a product in market when we invested. Location wise, almost half our active portfolio companies are in New York. A quarter are in San Francisco, but less than 5% in Silicon Valley. The remainder are in the rest of the US, London, Paris, Berlin, Poland and Israel.

Priced vs. convertible notes:

Over the last six years there has been an industry shift towards convertible notes as opposed to priced rounds, and our investments follow this trend. The exception to this is 2008 when given the financial collapse we insisted on priced rounds as an express investment criterion, and more than 90% of our investments in that year were priced. In contrast, over the past few years more than 50% of the investments we have made are initially structured as debt in the form of convertible notes.

Results

Of the $13.1M we have invested over the six-year period, we have $72.4M of realized and unrealized returns, split roughly evenly between realized and unrealized. As expected, returns are concentrated — but less concentrated than we would have thought. The top three investments make up around half of the returns, the top five 70% and the top ten make up 85% of the returns. The shape of our return curve is colored by the fact that we don’t hold reserves. If we had invested in subsequent follow on rounds, our performance would have been more concentrated at the head of the curve, as is the case for VC’s. Also, there are occasions where betaworks receives additional, advisory equity from a company for specific help, for example with product development. This analysis excludes that additional equity. Had we included this additional equity in investment returns, the total returns number would be well in excess of $100M and Twitter would have been our single most successful holding.

TVP what?

When analyzing investment returns by year, we measured the TVPI (Total Value to Paid In), which is the cash-on-cash return of an investment (it includes unrealized returns, where we use the most recent financing or external valuation). One of the reasons this metric paints a more accurate picture than, say, IRR (Internal Rate of Return) is that we will sometimes see a talent acquisition or “acquihire” of a company in a short period of time after a financing round. This can create a high IRR as a function of the short time period, even with a low exit value. Looking at the TVPI paints a picture that feels more accurate. Venture funds combine all the returns for a single “vintage” (i.e., the year a fund was incorporated). Combining our returns results in a TVPI of 5.5×. To compare that to a benchmark, Cambridge Associates U.S. Venture Capital Index has funds with a 2007 vintage returning an average TVPI of 1.73. Cambridge Associates caveats their results with: “analysis and comparison of partnership returns to benchmark statistics may be irrelevant.” We agree and take this only as a directional indicator that we’re doing a pretty good job of investing.

We are already well into the first quarter of 2015 — we have an amazing set of companies growing up and growing out of the studio, matched with a set of new companies under development and a whole network of companies we have been fortunate enough to invest in over the past six plus years. We do a lot of building at betaworks, literally and metaphorically. We view building and investing as coupled and related and we don’t think we could be as successful doing one, if we didn't do the other.

Illustrations by: Henry McCausland.