Josh Kopelman tweeted about my weekend project today, which is still in alpha (at best.) Until yesterday, the only follower was Dave McClure. But he follows so many accounts that I figured he just auto-followed anyone who tweeted '500 Startups.' I was going to wait until I got back from next week's vacation before mentioning it to anyone, but the twitter-feed is open, so... it's been released.
The website is the API: I wrote a script to look at the portfolio pages of VC websites every night and tweet and post new companies that seem to have been added. It's pretty useful in a way, but has some severe limitations.
The biggest is that scraping is inherently fragile. And I'm going on vacation next week and leaving the computer it's running on at home, running it. If it starts to spew garbage on Monday, well... sorry. I'll fix it when I get back.
It is reporting on differences. So when Stickybits changed to Turntable.fm, it showed that First Round added Turntable. When Tremor Media changed its name to Tremor Video, it got reported, etc.
VCs don't always update their portfolio pages in real-time. This is no substitute for Techcrunch (or Crunchbase, even), it's just faster and easier to scan. I've been adding the VCs in drips and drabs over the past few weeks, so there are certainly additions that got missed because of the timing. The web site is pretty cool, it lets you filter by name and date and sort. That will be more useful over time as more deals get added.
The list of VCs looked at is here. If I missed yours, let me know and I'll add it to the to-do list. Unless your site is in flash (cough, Norwest) or has no portfolio company names (ff Ventures, among several others) or doesn't allow bots (yes, oddly enough, there is a site that checks the user agent and sends my script a page with no real content; the robots.txt--as with all of the pages I look at--allows, but the server doesn't. Why?)
Enjoy.
Thursday, June 23, 2011
Wednesday, June 8, 2011
The new adtech is disruptive, and that's a good thing
I was at Luma Partners' Digital Media Summit today. Great event, saw lots of familiar faces. I watched the adtech panel. Of the people I didn't get a chance to invest in, these are four of the smartest: Brian O'Kelley, Joe Apprendi, Michael Barrett, and Mike Leo.
Mike Leo said something that struck me as wrong. He said, roughly, "30% of the media spend is getting spent on the pipes"--by which I think he meant the other panelists' companies--"and that's eating into the creative and the content, where it should be spent."
This is exactly the wrong way to look at it.
The entire process between the maker of a product or service and the user of that product or service--what we call marketing--is friction. This includes the pipes, the ads themselves, and even the content created to wrap the ads. Friction, all of it.
I agree that we should reduce the friction, make things more efficient. But if we give credit to Wanamaker's 50% waste in advertising spend, then the 30% that's now being spent on the "pipes" is a 40% improvement. A 40% improvement over five years is pretty spectacular. But disruptive technologies do that.
Leo is looking at the wrong place--he's complaining about the one area of the marketing process that has actually shown efficiency gains, while giving a pass to creative and content, the areas that have fallen behind.
Mike Leo said something that struck me as wrong. He said, roughly, "30% of the media spend is getting spent on the pipes"--by which I think he meant the other panelists' companies--"and that's eating into the creative and the content, where it should be spent."
This is exactly the wrong way to look at it.
The entire process between the maker of a product or service and the user of that product or service--what we call marketing--is friction. This includes the pipes, the ads themselves, and even the content created to wrap the ads. Friction, all of it.
I agree that we should reduce the friction, make things more efficient. But if we give credit to Wanamaker's 50% waste in advertising spend, then the 30% that's now being spent on the "pipes" is a 40% improvement. A 40% improvement over five years is pretty spectacular. But disruptive technologies do that.
Leo is looking at the wrong place--he's complaining about the one area of the marketing process that has actually shown efficiency gains, while giving a pass to creative and content, the areas that have fallen behind.
Tuesday, June 7, 2011
Valuation for investors
Twice in the past week I have had pre-product entrepreneurs tell me that they were raising seed rounds at an approximately $10mm pre-money. In both cases I had to pass, despite the merit of the management teams. Both companies told me that they have other early-stage investors ready to fill their rounds, and I'm glad. I am generally of the opinion that if an entrepreneur can get a better valuation while still getting value-add investors, then they should.
Plenty has been written recently by venture capitalists about venture capital to help entrepreneurs. Not much has been written to help newer venture capitalists. I think, in a way, this is because VCs don't care so much if those new to the industry succeed or not. If a new angel loses his shirt, well, one less competitor for me down the line.
I don't believe that, though. I think we need more investors. But smart ones, investors who make some money on their investments and so feel confident reinvesting it in a new round of entrepreneurs. Investors, like everyone else, get smarter and more helpful the more experience they have. Smarter investors is better for the ecosystem.
So, valuation.
Valuation in venture capital is tough. The amount of uncertainty between investment and exit is immense. But that doesn't mean that you shouldn't try to pay the right price. Valuing a startup correctly means estimating risk, not contemplating the unknowable. The idea that was briefly tossed around that valuation doesn't matter because a startup either goes big or dies is wrong. Ludicrous, in fact. There's a range of outcomes for every fund. The cliche that out of each ten investments, two are winners, three are failures and five go sideways shows this. The two winners determine whether the fund is an overall winner or loser, but how sideways the five go determines whether it's a good return or a great return. What happens to the second tier of investments matters. And, for the math challenged, even if startup returns were binary, when you invest in many of them you get a binomial distribution, so expected value matters.
I look at valuation this way. For every company,
This is the post-money valuation.
An example. Company X is an amazing data-driven adtech company. It's going to disrupt some existing companies and if it does it should sell for $400mm in five years. I think, given the risks, that there's a one in ten chance they will succeed.
But I know they will need to raise a Series A to commercialize the product once it's ready, and a Series B to ramp sales once they have product-market fit, and a Series C to expand the product line. The Series A will be 33% of the company, the Series B will be 25%, and the Series C 20%. My stake will be diluted down to 40% of my original ownership.
So my post-money expected value of the company is $400mm * 10% * 40% = $16mm. I would be looking for a post-money of $5mm. If the company is raising $1mm in the seed round, the pre-money valuation would be $4mm.
You can see the difficulty in the $10mm pre-money. If the company is raising, say, $2mm at a $10mm pre, then the expected exit value would have to be $12mm/(10% * 40%) * 3 = $900mm.
Billion dollar exits are the sine qua non of the venture business. But they are rare. Rarer than you think.
I made a list off the top of my head of some 125 business-to-business advertising exits. I may be missing some obvious ones, but there were only a handful of $500mm plus exits in the last ten years, even fewer billion dollar ones (M&A exits, I didn't count IPOs, so I probably undercounted by one or two.) Here are the $500 million and up exits I have.
Of the 125 exits, five were more than a billion, seven were between $500 million and a billion, 20 were between $200 million and $500 million, and 15 were between $100 million and $200 million. The rest were sub-$100 million. Remember, these were all exits--companies that didn't make it weren't counted. There's also a bias in the list because I am more aware of the large exits; I would be surprised if I missed too many billion dollar exits but I am sure I missed many $10mm exits. Also note that only a couple of the billion dollar exits here were as straightforward as my model: aQuantive was built through acquisition (and thus had substantially more dilution), Doubleclick had gone through several owners including the public markets, etc.
In fact, all else being equal (a priori, that is) billion dollar exits returned less overall than $500mm-$1bn exits, because there were fewer of them. Exits between $200mm and $500mm probably returned slightly more than $500mm-$1bn exits (also because there were more of them). The $100mm-$200mm range and less than $100mm range each return less than the $200mm-$500mm range**. Here's my estimate of what each of these ranges returned.
The sweet spot in adtech, the "average" expected exit value, seems to be around $400mm.
Every industry is different, you need to know yours. Make a list of exits over the last ten years, all exits not just the good ones. Then try and figure out how many companies were funded in your industry. This will inform your expected exit values in the success case as well as help you decide what percentage of funded firms get to an exit. Conditions change all the time, of course, but looking at the last ten years will probably keep you reasonably conservative.
-----
* I asked an engineer friend of mine how comfortable he is being in the buildings he helped design. "Pretty comfortable", he said. "You never worry?" I asked. "Look", he said, "There's a lot of math and experience behind choosing exactly how much steel and concrete the building needs to bear its load. I do the work carefully, run the calculations twice and make absolutely sure the answer I am getting is the right one. Then I multiply by three." [Edit: For those for whom anecdote is not analysis, I'll point out that a 25% IRR compounded annually for five years is 3x].
** The list I made is up here. Click on the headers to sort. I think all of the #N/As are sub $50mm exits except the two bolded ones, which I think are ~$200mm exits. [Edit: Wow, the sorting on the linked table was all screwy. Sorry. Fixed it.]
Plenty has been written recently by venture capitalists about venture capital to help entrepreneurs. Not much has been written to help newer venture capitalists. I think, in a way, this is because VCs don't care so much if those new to the industry succeed or not. If a new angel loses his shirt, well, one less competitor for me down the line.
I don't believe that, though. I think we need more investors. But smart ones, investors who make some money on their investments and so feel confident reinvesting it in a new round of entrepreneurs. Investors, like everyone else, get smarter and more helpful the more experience they have. Smarter investors is better for the ecosystem.
So, valuation.
Valuation in venture capital is tough. The amount of uncertainty between investment and exit is immense. But that doesn't mean that you shouldn't try to pay the right price. Valuing a startup correctly means estimating risk, not contemplating the unknowable. The idea that was briefly tossed around that valuation doesn't matter because a startup either goes big or dies is wrong. Ludicrous, in fact. There's a range of outcomes for every fund. The cliche that out of each ten investments, two are winners, three are failures and five go sideways shows this. The two winners determine whether the fund is an overall winner or loser, but how sideways the five go determines whether it's a good return or a great return. What happens to the second tier of investments matters. And, for the math challenged, even if startup returns were binary, when you invest in many of them you get a binomial distribution, so expected value matters.
I look at valuation this way. For every company,
- I think about what the expected exit would be if the entrepreneurs were right: if they are right about the problem, about what their customers want, about their ability to execute, right about everything.
- I figure out how much dilution I expect before the exit.
- I decide how likely it is they are right and multiply by the expected exit to get an expected value.
- I divide by three, because I'd like my investments as a whole to return 3x*.
This is the post-money valuation.
An example. Company X is an amazing data-driven adtech company. It's going to disrupt some existing companies and if it does it should sell for $400mm in five years. I think, given the risks, that there's a one in ten chance they will succeed.
But I know they will need to raise a Series A to commercialize the product once it's ready, and a Series B to ramp sales once they have product-market fit, and a Series C to expand the product line. The Series A will be 33% of the company, the Series B will be 25%, and the Series C 20%. My stake will be diluted down to 40% of my original ownership.
So my post-money expected value of the company is $400mm * 10% * 40% = $16mm. I would be looking for a post-money of $5mm. If the company is raising $1mm in the seed round, the pre-money valuation would be $4mm.
You can see the difficulty in the $10mm pre-money. If the company is raising, say, $2mm at a $10mm pre, then the expected exit value would have to be $12mm/(10% * 40%) * 3 = $900mm.
Billion dollar exits are the sine qua non of the venture business. But they are rare. Rarer than you think.
I made a list off the top of my head of some 125 business-to-business advertising exits. I may be missing some obvious ones, but there were only a handful of $500mm plus exits in the last ten years, even fewer billion dollar ones (M&A exits, I didn't count IPOs, so I probably undercounted by one or two.) Here are the $500 million and up exits I have.
| Company | Acquiror | Price ($mm) | Date |
| aQuantive | Microsoft | $5,900 | May-07 |
| Doubleclick | $3,100 | Apr-07 | |
| Omniture | Adobe | $1,800 | Sep-09 |
| Overture | Yahoo! | $1,630 | Jul-03 |
| Digitas | Publicis | $1,300 | Dec-06 |
| NetRatings | Nielsen | $817 | Feb-07 |
| AdMob | $750 | Nov-09 | |
| Right Media | Yahoo! | $680 | Jul-07 |
| Lending Tree | IAC | $675 | Aug-03 |
| 24/7 Real Media | WPP | $650 | May-07 |
| Rosetta | Publicis | $575 | May-11 |
| Razorfish | Publicis | $530 | Aug-09 |
Of the 125 exits, five were more than a billion, seven were between $500 million and a billion, 20 were between $200 million and $500 million, and 15 were between $100 million and $200 million. The rest were sub-$100 million. Remember, these were all exits--companies that didn't make it weren't counted. There's also a bias in the list because I am more aware of the large exits; I would be surprised if I missed too many billion dollar exits but I am sure I missed many $10mm exits. Also note that only a couple of the billion dollar exits here were as straightforward as my model: aQuantive was built through acquisition (and thus had substantially more dilution), Doubleclick had gone through several owners including the public markets, etc.
In fact, all else being equal (a priori, that is) billion dollar exits returned less overall than $500mm-$1bn exits, because there were fewer of them. Exits between $200mm and $500mm probably returned slightly more than $500mm-$1bn exits (also because there were more of them). The $100mm-$200mm range and less than $100mm range each return less than the $200mm-$500mm range**. Here's my estimate of what each of these ranges returned.
| Exit Range | Companies | Total Est. Value |
| $1bn + | 2 | $3,000 |
| $500mm - $1bn | 7 | $5,000 |
| $200mm - $500mm | 15 | $6,000 |
| $100mm - $200mm | 25 | $4,000 |
| $50mm - $100mm | 50 | $3,750 |
| < $50mm | 100 | $2,500 |
The sweet spot in adtech, the "average" expected exit value, seems to be around $400mm.
Every industry is different, you need to know yours. Make a list of exits over the last ten years, all exits not just the good ones. Then try and figure out how many companies were funded in your industry. This will inform your expected exit values in the success case as well as help you decide what percentage of funded firms get to an exit. Conditions change all the time, of course, but looking at the last ten years will probably keep you reasonably conservative.
-----
* I asked an engineer friend of mine how comfortable he is being in the buildings he helped design. "Pretty comfortable", he said. "You never worry?" I asked. "Look", he said, "There's a lot of math and experience behind choosing exactly how much steel and concrete the building needs to bear its load. I do the work carefully, run the calculations twice and make absolutely sure the answer I am getting is the right one. Then I multiply by three." [Edit: For those for whom anecdote is not analysis, I'll point out that a 25% IRR compounded annually for five years is 3x].
** The list I made is up here. Click on the headers to sort. I think all of the #N/As are sub $50mm exits except the two bolded ones, which I think are ~$200mm exits. [Edit: Wow, the sorting on the linked table was all screwy. Sorry. Fixed it.]
Monday, June 6, 2011
On failing
Don't confront me with my failures, I have not forgotten them.
- These Days, J. Browne
Six years ago. I was working on a startup with a bunch of friends, a big idea and one that got me out of bed every morning, excited to go to work. We had raised money from a Name and some amazing venture investors. We were going to be the next Google. My other gig, a venture fund, had just had the exit that put it over the top: one of the portfolio companies had gone public. Me and my two partners--one active and one silent--had made the fund back and then some. I was going to make more than living wage from the eight years of work that portfolio represented. At home, we had finished furnishing the house, our third child was on the way, and my roses were finally getting some traction.
I was out with a good friend recently. He told me an old college buddy of ours was having some trouble with his career, with his marriage. I'll call him, I said, I should talk to him. No, my friend said, he doesn't want to talk to you right now. He thinks you wouldn't understand. Everything's always worked for you.
Five years ago. Our Name investor had realized that if we stopped trying to grow the business, it would immediately begin making money. This was not what we wanted, we wanted to be the next Google, not a tiny mortgage lead generator. But he wasn't a venture guy, he wasn't an investor at all, he was a buyer of assets. He scared away the outside money, starved the business of cash, and forced us to accept a buyout. I had been pushed out of the company I had helped start. In the entrepreneurial community, there aren't many marks blacker than that.
Meanwhile, my active partner in the venture fund decided that the silent partner should not be paid their share of the fund, contrary to what our contract seemed to my non-lawyer eyes to say, and certainly contrary to what fairness dictated. When I refused to go along with that, he sued me. He also sued the silent partner. The silent partner fought back, and the whole kit-and-kaboodle got locked down. The vast majority of my assets were frozen, being held hostage to some ridiculous legal squabble for some indeterminate period of time.
Not entirely coincidentally, my marriage fell apart at that exact moment.
I had failed, utterly. I had no job, my net liquid assets were approximately zero, my reputation was in tatters, and the one thing in my life I thought was permanent was over. The struts that supported my sense of self all got knocked out pretty much simultaneously. Everything I had, everything I hoped for. Five years ago, today.
In what way are you the same person today that you were twenty years ago? There is no atom of matter in you that was there twenty years ago, there is no piece of you that is the same as it was. The pieces of you, the functioning of you, the pattern of you are all different. Why are you you? What is you? What is it that continues? When the Buddha told us to meditate on death this is the question he wanted us to ask ourselves: what where you before you were born, what are you after you die? In what way are you still you even twenty minutes from now?
I'll tell you an answer, although me telling you no more gives you wisdom than reading a cookbook gives you nourishment: we are a span of links in a chain of causality. My existence now causes my existence a minute from now, and that a minute from then. There is no me except my self causing my next self every tick of the clock. In consequence, I am subject to the Markov property: how I arrived at the current me is irrelevant, every option available to me is embodied in who I am now, not in how I got here.
I'm no saint and failure did a number on me. Someone told me what intense stress would do to my psyche, the stages of irrationality I would go through. Frustratingly, the knowledge of them did not allow me to avoid them. But failure taught me some humility, and being humble forced me to abandon the personas I had built and my arrogance of idealism. And I finally learned that except in how it hobbled me, my failure made no difference--all that mattered was what I did next.
Five years after everything fell apart I have the best portfolio of startup investments in one of the hottest spaces in tech. People I respect and admire recommend me to entrepreneurs. I am allowed to be productive. And I spent Memorial Day with my loved ones and was happy and relaxed and even looking forward to getting back to work after a long weekend. I appreciate it, all of it, every time. I haven't forgotten I had nothing. I take nothing for granted. But I don't now and never did believe that the past determines the future. The future is determined only by the choices you have now--the ones you can find a way to allow yourself--and what you do with them.
I don't think saying that everything's always worked for me is wholly accurate. But it may be that it is entirely true.