To me this is essentially admitting that Andreesen Horowitz is loosing the ability to identify impactful startups. That is fine, its harder to do at scale, but nothing is different now than 5 years ago. We have lots of people doing experiments, some hit early success, some pivot. Once you get product market fit, you raise that B/C/IPO on the back of the growth you've been able to afford thanks to your raises and revenue.
Just as Berkshire Hathaway had to change their investment strategy when they grew very large, it is fine for AH to do the same, but I don't believe for a second that things are hugely different in 2013.
Maybe you're right, but that's not what Scott was saying and that's not what I think.
In my own words: consumer startups more and more have this very interesting "lightning in a bottle" effect where sometimes they take off like crazy and sometimes they just don't. I give full credit to the teams that figure out how to get the flywheel spun up, but it is also important to realize just how many highly capable founders are hard at work trying to get traction who don't. There are a lot of really excellent founders pursuing consumer ideas that just never work -- that's why companies like Yahoo and Google and others can do so many acquihires. So, if we have the theoretical ability to invest in a given category -- remembering that we can only make one primary venture investment per category -- in either the A or B round, we find it often makes sense to let other firms fund the A rounds before anything is proven and wait to see the early signs of lightning and then step in hard at the B. The end markets are so large for the winners that the investment returns in the B can still be outstanding, and we can still offer a lot of useful help to the companies at the B stage such as talent sourcing.
In contrast, enterprise startups are much more (take your pick) tractable, execution centric, brute force, predictable (as startups go). If you back a killer founder with a great engineering team, with a great idea, into a big market, the odds are high that magic will happen -- a very interesting product will get built, early customers will adopt, and value will be created. In other words, the link between founder/team competence and success is more direct. One thing that helps a lot is that whether the product will be adopted by customers or not is far less of a mystery -- you can simply go talk to the likely customers ahead of time and they will give you a very good indication. That plays well to our market development program where 1,200 big company management teams are coming through our office every year -- we ask them what they think about new ideas and they tell us. So here, backing the A round when possible makes more sense.
None of this is religion -- we still do plenty of consumer A's and enterprise B's. We just think it's useful to talk about these things in public so that entrepreneurs know before they come see us how we are thinking about things -- it optimizes their chances of getting to the right outcome with us (whatever that is).
My working theory is that consumer ideas are both art and engineering. The art component either takes in the mind of the consumer or it doesn't. Like movies or music, sort of. Lots of great engineering teams can't get the art to work with high probability, and even the great consumer internet artists swing and miss as often as they hit (eg odeo vs twitter). And if the art doesn't take, the engineering won't save you.
Whereas the enterprise ideas are mostly engineering - or rather the art lies in really understanding the customer and the domain, which is relatively straightforward for the great enterprise entrepreneurs, because they are already so deep in it and you can easily go talk to the customers one at a time and learn what you need to know to predict success or failure with pretty high confidence. Not easier but different.
These are all gross over generalizations of course. I think of this as a framework for thinking - one of many - not a literal description of the truth in all cases. One lens.
I love this quote about art and engineering, music and film. Spot on. So as a startup founder, filmmaker, musician - here's my question: how are you ever gonna get these blockbusters/chart toppers without taking a chance. The biggest film in the U.S. right now (Gravity) is by a Mexican filmmaker who made arty films like Y Tu Mama Tambien and Children of Men - not hollywood stuff. Really breakout, worldchanging, consumer, Apple-style stuff is made by weirdos, punks, poets and freaks. With all the capital A16 has to deploy - why not create a fund for some of the crazy ones?
Andreesen Horowitz is loosing the
ability to identify impactful startups
Pet peeve: it's losing, not "loosing".
I don't know if anyone can properly guess if a consumer startup is going to get traction until it does, especially in the mobile space. That said, they did manage to get a 312x return on their seed investment in Instagram (which they declined to invest in a second time as explained here: http://bhorowitz.com/2012/04/22/instagram/)
A-H offers seed investments to YC companies in order to get a foot in the door with the small fraction who will succeed. Offering blanket seed investments confirms that they really don't have any idea at first either. There's nothing wrong with that, it's just the nature of the beast. (http://venturebeat.com/2011/10/14/andreessen-horowitz-to-giv...)
I don't know if anyone can properly guess if a consumer startup is going to get traction until it does
A classic example is the drug Viagra, one of the most profitable consumer technology investments in recent history. Its (most profitable) use today was originall reported as a "side-effect" in early clinical trials.
Losing the ability to identify or losing the desire?
When you get down to it, if you have a $50m fund, you might want to invest in ~10 A-rounds at ~$5m each. If you have a $1.5bn fund however, do you really want to invest in 300 A-rounds?
Nope. The best way to spend your time is looking at the bigger, later investments. You probably still only want to have roughly the same number of total investments to oversee. So if an A-round company is gonna take the same amount of attention as a B-round or later, you shouldn't waste even a second of your time considering them.
It's nothing to do with ability to identify, just correctly prioritising the time the VC's have available to them.
In our case that isn't actually how we think about any of this.
First, when we (or any venture firm) makes an A-round investment, we typically reserve another 2-3x of the A-round investment size for participation in future follow-on rounds for that company. So a $5M Series A shows up on our books more like a $20M commitment. The other $15M isn't necessarily always deployed, of course, but we also double down even more strongly in certain cases (either out of opportunity or sometimes necessity) so it balances out.
Second, it's not either/or -- we do venture rounds as small as $3-5M and we do growth rounds as high as $100M. Each fund has a blend of both.
In practice, we don't pay a lot of attention to any of this. When we get a great A round opportunity, we take it. Same with B rounds, and same with later-stage growth rounds.
Interesting to know, thanks. My example was probably a bit too hastily made-up.
So I guess in summary, it's not so much about investor time and attention but mostly about managing the risks of the different sectors.
Consumer startups are inherently less predictable and therefore you invest only when a company appears to have found its market and be at least partly on the way to success. Whether that be A round or later is irrelevant.
Whereas the enterprise startups are more predictable, in that if you find a great founding team with a good idea, they are more likely to make a success of it in that sector. Therefore you can cast a wider net, earlier in the lifetimes of the companies, than you would for consumer startups.
Or to summarise the summary: A good, predictable bet is better than an unpredictable one. Or maybe, never look a gift horse in the mouth?
Generally yes. Distilled down, quality of team seems to be a better predictor of success in enterprise than consumer, in that we see many more great consumer teams struggling to get traction than we do great enterprise teams struggling to get traction (on a proportional basis). There are many other factors for both but the "lightning in a bottle" element for consumer startups is real and very frustrating for teams that end up not being able to capture it.
I would just take out "only" from your comment -- there are exceptions everywhere. When we talk about patterns like these, they really are only patterns -- the truth is always in the specific details.