VCs Are Not Your Friends
steveblank.com
steveblank.com
Imagine Steve had been in a somewhat different situation: a friend of his was running a company that was a potential strategic partner, and after an initial investigation decided that it didn't make sense. Would that mean they weren't friends? Of course not: it just means that the business relationship didn't work out in this case. Real friendships survive that.
More clearly noted here, when he writes about not confusing a VC's good behavior, as their personal trait.
... they add tremendous value to your startup (recruiting, strategy, coaching, connections, etc.) they are not doing it out of the goodness of their hearts.
... At the end of the day VC’s have to provide their limited partners with great returns or they aren’t going to be able to raise another fund. If you succeed so do they. Great VC’s do everything they can to make you successful. But just like your bank, credit card company, mortgage holder, etc. they are not confused where their long term loyalty lies.
I don't see it that way. if person X is my friend, that's true whether or not they're acting as a VC in a situation. VC's who I am friends may well come to the same conclusion as a VC's I don't know but our interactions are likely to be different. Friendship doesn't have to get in the way with a business relationship; conversely, difficult situations in a business in a relationship don't have to interfere with friendship.
A couple of my angels are invite-to-the-wedding friends, but I still expect them to make rational decisions when it comes to their pro rata in later rounds, and I don't automatically expect them to invest in whatever I decide to do in the future.
"Don’t confuse being friendly with your ___’s with ___’s as your friend"
If that signal truly meant there would be no follow on round, why would you expect the CEO to not throw in the towel as well? I've neve been in that situation - anyone out there who's been there care to comment on what it's like to face the world after your lead investors bail?
Throwing in the towel when the going gets tough seems like a spineless move.
Expectations are different, sure. There are a lot of good things that can come from it in spite of that.
It's the people involved that make it crap or awesome.
src: http://www.woopidoo.com/business_quotes/authors/john-rockefe...
Edit: not a criticism of the quote itself. These things wouldn't get misattributed if they weren't worth repeating. But they wouldn't get repeated if they weren't misattributed, either; it's a kind of cultural appeal to authority.
Google Books is great for verifying this sort of quote: http://books.google.com/books?id=IHnZAAAAMAAJ&dq=A%20fri...
The recent blanket offer by Yuri, I think, adds credence to this line of thinking: money is getting cheaper for software startups as demand for it falls. (An alternative theory, which I am also sympathetic to, is that we are at the tail end of a reflationary credit bubble, set to burst Real Soon Now.)
But, regardless, why sleep with the enemy if you don't have to? You can build a POC on heroku or EC2 for peanuts...
However, "the cloud" passes on very little of the economies of scale involved on to the consumer, so while it's great to start in "the cloud" you want to be careful not to lock yourself in. It's a great place to start, but once it makes sense to start buying your own hardware, you want to be in a position where you can do so.
2. a 32GiB ram 8 core, 4 disk server is around $2000 worth of parts. renting something like that is going to cost you around 1/3rd to 1/4th that /every month/
3. power costs need to be factored in. you can get a full rack with 15a power (remember, only use 3/4 of that) which can handle maybe 8 or 9 of the aforementioned servers for around $400/month
Of course, there are a bunch of other costs that come with owning your own hardware, and certainly, on the small end of things, the cloud saves you a lot of money if you don't have a hardware guy on staff.
My point is just that you have to be careful, because as you scale up, the "cloud" prices don't scale down very much at all. We can argue all day about when reach the point where owning hardware starts to make sense for the stuff you leave on all the time, but it's pretty clear that at some point you do reach that point.
Moving off ec2 on to your own hardware, usually, isn't that difficult. Linux is Linux. But moving off something like appengine on to my own hardware? that starts to get... difficult.
The equation of cloud vs not cloud is much more complicated than pricing. In fact, I'd argue price is the least impactful component, especially since it's getting cheaper and cheaper.
The most important questions are around personnel, risk, time, and so on. The cloud dramatically changes these and my view each day that goes by it makes more sense to run your business on AWS than it does to rack your own servers.
for some parts of some businesses, you are right. if compute time is a very small portion of your budget, who cares if you pay a little more? I mean, tools are a vanishingly small portion of my budget, but they are really important to the continued functioning of my company, so I'm willing to pay a very hefty (percentage wise) premium for nice tools. It's only a few bucks extra, even if it is a 50%-80% premium over the cheap stuff, so who cares?
My understanding was that netlfix moved it's cpu to the cloud, for the reasons you stated. If processing is a small portion of your budget, paying more for hosting of your CPU, and having the excellent flexibility of the cloud might make a lot of sense.
My further understanding was that netflix used something else for actual streaming distribution of the movies; amazon wanted too much money for that, and unlike CPU time, bandwidth /is/ a significant cost for Netflix.
The parallel in my business is that probably the largest monthly cost of mine is new hardware (followed by power for my existing hardware) Unlike netflix, my margins on each bit of hardware I get are pretty thin. An 80% increase in the cost of my hardware would kill me outright. I put a lot of effort into keeping my hardware (and my power) costs down.
(to be clear, moving to amazon ec2 would increase my costs by quite a bit more than 50%-80% - in most cases, ec2 would charge me more wholesale than my customers pay me, retail, for guests.)
and I don't know. Maybe if I started actually selling hardware (which I've considered... I'm pretty good at choosing/assembling, and can do it quite a bit cheaper than Dell's retail prices; the big hangup would be sales.) and I started spending thousands of dollars a month on tools, I might want to spend more effort getting the cost of tools down.
That's the idea behind using standard parts. A screwdriver is a screwdriver; I can easily switch brands. The same, really, could be said for "unix box on demand" services like ec2. It's not that difficult to move from one provider to another, or on to your own hardware, if that makes sense down the road.
On the other hand, with the "proprietary programming environment" systems like appengine, it may be somewhat more difficult to move off of the system. (the open-source runtimes may have progressed further since last I looked, but last I looked it was pretty much run it on google, or re-write your code.)
No, those are business relationships too. I know that in real-estate deals when the development entity needs to raise more money from the partners, those who cannot do so will find themselves "unfriended." In fact, it is not uncommon for raising additional funds to be used as a tool for increasing returns for remaining partners in that environment. I suspect that if VC's never do so, it is because of contractual obligations not a the kindness of their hearts.
VCs have a job to do, just like everyone else. Sometimes they need to make hard decisions.
It's important for every business professional (entrepreneur or anybody else) to understand the basics of other people's goals and incentives.