Tearing Down The Teardown

CB Insights published a “teardown” of USV’s investment strategy last week. It’s a pretty solid piece of work considering they did not have access to any of our internal data.

The got some stuff wrong, however.

1) We did not participate in Zynga’s “megaround” in Feb 2011.

2) We only have six main investment vehicles;

USV 2004 – $125mm

USV 2008 – $160mm

USV Opportunity Fund – $125mm

USV 2012 – $200mm

USV 2014 – $175mm

USV Opportunity 2014 – $175mm

3) We have invested in a lot more than four YC companies. I think the number is closer to ten. I think YC is by far the investor we follow the most, particularly in recent years.

But, as I said, they did a pretty good job considering they don’t have access to our internal data. And I do not believe our limited partners are sharing our reports with them.

What this shows is that the venture capital business is becoming more transparent because so much of our investment activity, and the activity of our peers, is being tracked as it happens. When we raise a fund, that is reported (we must disclose that fact due to securities regulations). When we make an investment, that is generally announced by the portfolio company. And there are reporting requirements for that too. It is possible to do a stealth financing, but it’s not easy. When we make a follow-on investment, that is often announced and/or disclosed. And most exits are disclosed.

What is harder to figure out is what our ownership levels are in our portfolio companies. If you knew the amount we invested and the valuation of the round, you could figure that out. But that would be very hard to do accurately and consistently. I don’t think anyone is going to be able to do a teardown of a VC fund and its returns any time soon.

But even so, it’s impressive what CB Insights and others are doing to track and measure and report on VCs and their investment activities. Entrepreneurs should be able to get some third party assessment of the quality and performance of the VCs they might work with. That’s totally possible now and I think that’s a good thing.

#VC & Technology

Comments (Archived):

  1. William Mougayar

    If there were no reporting requirements, would you keep all that info private? Is there a competitive edge in keeping more of that info private than in releasing it?

    1. Anand Sanwal

      William – good question. Let me offer our perspective.Even if an investor wants to keep it all private, it’s hard (impossible) as there is so many sources of this data and lots of “data exhaust” on the web we use to track it down. Whether a news publication, Twitter, a tech blog, a company’s website, a portfolio page on an investors’ website, etc, there is usually a breadcrumb we can find and follow.Of course, many of these sources are unstructured and messy and so the technology to crawl, classify and extract the salient structured data is one we continue to refine, but the transparency and visibility is only going to increase. And in big ways.

      1. Dave

        Very interesting piece. Probably easier to figure out VC returns on an investment than other asset classes since VC is a home run driven business. In home runs typically the returns leak out pretty quickly or are announced via an IPO or sale. Everything else is an ok return or a loss that is less relevant to the performance picture. Different than PE and way different than hedge funds.

        1. Anand Sanwal

          Dave – yes. Very true. While you cannot know the exact returns, you can estimate prior performance using some fairly good heuristics.If you assume VCs target a certain level of ownership and you know when they entered a company, you can quickly tell who the best VCs are (top decile) and everyone else.On a related note, because it’s such a home run driven biz as you suggest, the returns follow a power law as well. If interested, more on that here – https://www.cbinsights.com/…Thanks for reading and for the comment.

    2. LE

      In business, in general, the idea is to keep things private. If you aren’t experienced enough, or don’t know enough to detail exactly the benefits of being open, it definitely makes more sense to keep your mouth shut. If you’ve got a specific reason to be open then that’s fine.In my particular business there are people that brag about things in order to get industry press and people who don’t. One person in particular wouldn’t agree to a big deal because CNN didn’t want disclosure on the purchase price. The person in question wanted disclosure because a) he’s got a huge ego and likes being stroked at industry shows b) he holds an industry show and almost certainly see’s it as a way to gain publicity and get more show tickets sold c) He likes to see his name next to the word “king”. Which is funny because it’s the peons that care about that label not anyone “important”.Countering that of course is a) the risk he runs by requiring disclosure (for the above reasons) b) becoming a legal target (long discussion but he does become a target) c) potentially getting a reputation (well deserved) as a guy who isn’t reasonable d) killing a deal by holding out for a term that is mainly driven by his ego and not by $$.Now maybe he has carefully considered all this and maybe he hasn’t. But it’s his seat of the pants feel that allows him to make a decision that is good for him. Each case is different. In retrospect you can always rationalize something that has happened as being positive unless it is clearly a mistake “helped me sell more show tickets” after all nobody is privy to the down side.One thing that I ran into early on was that (by comparison) many hedge funds are private and some literally don’t even have web sites. They aren’t interested in the world knowing their business. If you have something good you generally don’t want to put people on the scent of what you are doing. It gives them ideas and invites competition and envy and in some cases can make the government look in your direction and that could cause many problems.

      1. Matt Zagaja

        I tend to believe secrecy is over-valued. Knowing how much money Fred makes from each round/exit might be interesting but it certainly won’t help me get there. Maybe there are narrow cases such as not wanting a competitor to bid up ad space that you are both competing for but the nature of that kind of activity gives you only a short window of advantage before you are forced to reveal your hand (when the ad goes live) and then the competitor can follow. I do not know if business suffers from the same issue as politics, but it is often the case that something that is proven effective is ignored by incumbents because they are either not interested in it or decline to believe the data.In government where many things are open, the resources for investigation and enforcement tend to be constrained so even if the open data suggests something is wrong, it will not be chased unless there is media attention.

        1. LE

          Each case is different but another way to look at things is that the more people that know about something the more competition there is and the less likely it is that existing competitors can stand out. [1] And ultimately one of those competitors gets a foothold and becomes the big fish. This is not a reason to not be open (because as mentioned it needs to be weighed against the downside) but you have to realize that secrecy is not over-valued as you are saying. It’s strategic given the exact particulars of what you are trying to protect.Maybe there are narrow cases such as not wanting a competitor to bid up ad space that you are both competing for but the nature of that kind of activity gives you only a short window of advantage before you are forced to reveal your hand (when the ad goes live) and then the competitor can follow.It’s called adding fuel to the flames. With ad space in particular a local company may place an ad in the local shopper on a repeated business. When talking to competitors he may acknowledge even that “he gets some business from the ad” so as not to make the competitor think he is hiding something (you know when they talk at some trade show or to a common salesman). Otoh he would be stupid to brag somewhere about how great the ads are. That is almost certainly going to give the competitor more of a reason to place their own ad. Then the person viewing ads has more people to call for a quote. So it’s a balance. People know you are doing something but you have to be careful not to appear to successful lest they will potentially go down the same path.As far as “will find out anyway” that to me is like the “lock on the door won’t prevent a determined thief from breaking in”. You don’t have to give things to people on a silver platter otoh if they work for it it’s fair game.

        2. PhilipSugar

          You are transparent to the point it is valuable to you. Do you think Fred or Brad would pos the deals they are working on that they haven’t closed???? Hell no. So don’t be naive. Is it worth it to be transparent on how you structure deals. Yes.Politics is VERY different than business. As an elected official you work for me. That is a fact that is lost on almost all politicians. You work for me. Not the other way around. You are open because I demand you to be, it is your responsibility, it is your oath.In business, that is your choice. It is a competition. You do what you do for competitive advantage. I am VERY strict about not having employees give competitors information. I have had somebody say well they can get it anyway, well then make them work for it.

      2. PhilipSugar

        I agree with you.

    3. Vineeth Kariappa

      Think the question should be, ” Why dint CBinsights ask Fred to edit/ read/ verify the article before they published? ” He just gave everything!

    4. fredwilson

      There is some info we would not releaseBut we certainly believe in sharing the news on new funds we raise, new investments we make, and exits

  2. JimHirshfield

    “Best pound for pound VC”

    1. fredwilson

      Yeah. We are flyweights!

  3. pointsnfigures

    At least it gives transparency to potential LP’s, and if you are good helps your brand if you fundraise the next time. I see too many articles saying VC is dead. I’d prefer to see rational articles on why VC failed to get above average returns from certain vintages. Too easy for LP’s to shrug VC off as a dead asset class.I like what the Angel Capital Association does too. For example, in 2012 there were 66k companies angel funded with a median pre-money of 2.5M and a median raise of 600k. It’s a good benchmark, but not absolute. At least it forces both the entrepreneur and investor to talk about why they value the company more or less.

    1. Anand Sanwal

      Great point. For LPs who are assessing new funds who don’t have a track record yet, LPs use this type of data to understand those investors and the quality of their syndicate networks, historical selection aptitude, etc. These have all been shown to be drivers of fund performance and so LPs can augment the narrative with some hard data.

    2. LE

      I see too many articles saying VC is dead.Speaking of dead it’s Interesting how the VC biz isn’t organized enough (with a good pr flack industry organization) to fight that meme. To see what is possible you just have to look at the gun lobby. I mean you can make numbers say anything. And VC is numbers.Along those lines even with the recent spate of airline accidents there was mention on the nightly news recently and several places countering the “it’s getting more dangerous” meme with carefully placed rebuttal data (which no doubt came from some industry source) to make everything seem fine and still very safe. This almost certainly wasn’t dug up by reporters and editors it was given to them by PR flacks.

  4. Anand Sanwal

    Thanks Fred. Happy to answer any questions folks have (co-founder, CB Insights) – although about to jump on a flight in Austin.On the point of transparency which Fred brings up, it’s something the private markets in general are missing. On assessing VC effectiveness, we’re seeing a lot of institutional buyers use the data esp since in VC, past performance IS a good indicator of future performance (performance persistence as its known) and so tracking the smart money is useful. These include:- LPs who want to track existing GP commitments and more often that who want a read on a potential GP they’re evaluating. VC performance data is incredibly opaque, poorly reported and time-lagged. As a result, today, folks cherry pick data to make themselves “top quartile” and then complement that with a fantastic narrative. That is becoming tougher.- Corp dev/corporate strategy folks who are looking to understand disruptive trends and the companies and investors behind them. They’re esp keen on tracking the “smart money” as a signal for where the world is going. Again, they know that not all investment is smart money so if they’re looking to see where the world is going, they often will use the bets of top VCs to make that assessment (see here as an example assessing fin tech’s future based on where USV, Sequoia, etc are investing – http://cbi.vc/WUmqFV)- Fellow investors who are trying to understand their peers, syndicate partners, follow-on investors.

    1. LE

      Wow you now have another quote to add to your “Customer Love” page.https://www.cbinsights.com/…Here it is:CB Insights published a “teardown” of USV’s investment strategy last week. It’s a pretty solid piece of work considering they did not have access to any of our internal dataFred Wilson, Partner Union Square Ventures.

    2. Kasi Viswanathan Agilandam

      If you are in Austin you should try and meet with JLM … one of The best from AVC.

      1. Anand Sanwal

        Next time – just got back to NYC from Austin. Thx for the heads up.

  5. aweissman

    I did some looking in the spring and found that, of the active USV companies, at least 12% were YC companies

    1. Richard

      So let’s assume the question is what is the probability that a Y combinator company is funded by USV given that it was funded at some level.This seems like a it could be an application for BAYES rule.So picture two boxes, one box is of USV funded Y combinator companies at some Funding level + other USV companies and the other box is of non USV funded Y combinator companies at some funding level + other non YC companies funded at this funding level.So let’s say we pick some random YC company that was funded at this funding level.What’s the probability that it is a USV portfolio company ?The probability of picking a Y combinator company from either box is random, that is it is .5If the proportion of Y combinator companies at in each box is also known is known.So the probability that box 1 (USV) was selected given that the company is Y combinator funded at some level is what we want to know.This seems like something BAYES formula can solve.P(usv/YC) = .5(P(usv YC/ total usv)/ (5(P(usv YC/total usv) + .5 (nonusv YC/ total non usv))

    2. LE

      It should be interesting to see what happens to that number now that Michael is running the family business and “the Don” has decided to get his brain back.http://valleywag.gawker.com

  6. jason wright

    a bit noseywhat’s in it for cb?

    1. Anand Sanwal

      We sell data

      1. jason wright

        is your article that Fred links to just a summary of the data?

      2. Kasi Viswanathan Agilandam

        ha…ha… thanx for admitting it openly there are some who sell without owners knowledge :-).

  7. Brandon Burns

    “And most exits are disclosed.”What exits are not required to have details disclosed? A sale to a private company?

    1. pointsnfigures

      yup. also size. if it’s not material, it’s not disclosed (and GAAP, FASB and AICPA have rules around that)

      1. Brandon Burns


    2. Anand Sanwal

      Both public and private cos don’t have to disclose acquisitions. For public cos in particular, if the target company is below some materiality threshold, they need not be disclosed.Google and Apple do these small acquisitions and acqui-hires frequently and will just say “we did 30 acquisitions last quarter and spent X”. But they don’t disclose all co names.

      1. Brandon Burns


      2. LE

        I would think you could mine linkedin data and uncover things like that. Identify people working at startups and monitor changes in employment appearing on linkedin. Or monitor their blogs. No doubt that type of thing would be mentioned by someone somewhere.Another thing you could monitor is changes in whois data. In fact that would typically be an early signal prior to any other disclosure.

        1. Anand Sanwal

          The whois data is an interesting idea.LinkedIn is a treasure trove of data, but they are understandably protective of it. So unless you do dodgy things to get to it (violate their ToS), it’s not really possible.But we’re constantly looking for new breadcrumb trails to follow. There is more and more data exhaust out there so would expect that less and less will be “hide-able” over time.BTW, thanks for reading and for the comment.

    3. fredwilson


  8. jason wright

    why would one want to do stealth financing?

    1. fredwilson

      They don’t want to tip their hand to competitors and potential competitors

  9. AlexHammer

    Fred, you are a highly transparent individual, which I think has cemented your industry leader status and I’m sure helps with elite deal flow, etc.

  10. stevewfindlay

    You can get some VC-fund level performance returns from publicly available data. These aren’t frequent, but in the US a good example is CALPERS:https://www.calpers.ca.gov/…This has performance data on many VC funds, but only those that CALPERS has invested in. For example, through their investment in California Emerging Ventures fund of fund, you can see Upfront Ventures II (2000) fund achieved 2.4x.As noted elsewhere, there is a significant lag to this data. Usually 6-9 months+.In the UK, it is possible to get similar data from freedom of information requests (https://www.gov.uk/make-a-f… as many public bodies (e.g. public employee pension funds) have invested in UK-based VC funds. As well as VC funds elsewhere.If you are looking for *fund level* performance; it is the best to get the data from a “top-down” source. It is almost impossible to determine this from the “bottom-up” using individual portfolio company level data.For example, the following may have a big impact on the overall fund performance to LPs, without ever being publicly available: – GP fees: the management share to cover costs – Hurdle: the minimum return a GP must achieve, before carry is paid. And whether there is a “catch-up” (this can have a big impact on returns) – Abort costs: fees associated with deals tat don’t complete but need to be paid for by the fund – Draw-down timings: LP capital can sometimes be used and “recycled” within 6 months without impacting the denominator in the cash multiple of IRR calcs – Etc…!

  11. Adam Steiner

    I’d expect even that harder to get information (valuation investment amounts) will become more open and accurate with time…start pulling portfolio COIs for a good approximation of the round size and valuation, track that against the amounts typically invested by a particular VC for a particular round, etc. It’ll give a ballpark….and someone will automate the process soon enough (if they haven’t already)

  12. Conrad Ross Schulman

    Can’t it be verticalized? I can build an alogorithm..

  13. ShanaC

    Do you feel the transparency has changed the way you do business

    1. fredwilson

      We’ve adopted transparency as an offensive weapon

      1. Steven Kane


      2. Kasi Viswanathan Agilandam

        it is said that “Gandhi used non-violence very violently” !! 🙂

  14. leigh

    i love deconstructing people’s strategies based on publicly avail. data. It’s one of my fav. things to do as a marketer. It’s amazing what you can find out by a random interview here, an annual report over there and a social feed on top of that which has some paid media behind it.

  15. Archer Hobson

    What’s stopping a VC firm from from releasing this type of information? Fear of competition from other firms? My feeling is that if you feel it is better for the industry this way and that it feels inevitable, why not take the lead and offer up something that other firms will not.

  16. Kasi Viswanathan Agilandam

    That is almost a-billion-$ investment fund in a decade….spread over decade makes sense…BUT investing in a single company!!!…What you think of investing 1-billion-$ in a single company … I am surprised at the following news… investing $1-B in an Indian Company…http://techcrunch.com/2014/

  17. Tom Labus

    Do VC’s have to disclose returns to their investors?

    1. fredwilson

      yes, absolutely. we do that quarterly in a comprehensive report

  18. awaldstein

    If anyone Fred understands the power of thoughtful ‘transparency as a weapon’ it is certainly you.One of the things I respect most about you actually.

    1. fredwilson

      thanks arnold