Every company needs to be able to generate a constant stream of innovations if they expect to generate above average profits.

The trick is creating an organizational structure that generates that constant stream of innovation.   Google’s got their 20% of each engineers’ time.   LinkedIn has taken a different approach.   They have created a Light Engineering Development team.   The L.E.D. team builds apps really quickly using Ruby on Rails.   And, they have managed to perfect the process of turning those 1-week turn-around apps into huge, highly scaled applications.

Dennis Howlett over at ZD/Net has just written an article about how LinkedIn used this process to build a Rails app that scaled to 1 Billion page views a month.

We’ve also released a video over at Joyent:

Oliver Marks has launched a new blog on ZD/Net called Collaboration 2.0 . It looks like it is definitely worth adding to your subscription list.

A friend of mine sent out an email today saying he was upset that the Fed has stepped in to bail out Bear. My friend said he would have preferred to see the market find a natural bottom. My reaction:

“A natural bottom”???

Are you crazy?!?

Think about a real crash, a real war and the real humanitarian hell it produces.

I was born in Zimbabwe. Look what has happened to my country.

There is only one natural bottom, at it is total and utter depravity.

This is too serious a situation to play with theories about a supposedly self-regulating market. With both the dot com bubble and now the housing bubble, it is clear that the markets need some regulation.

If Bear had gone under tonight, Monday would have seen huge defaults in all the Money Market accounts of people like you, who have some portion of their cash in accounts that eventually lent some of their Billions to Bear. Just one witless moron on CNN saying “They said it was safe, but they done gone and lost every penny of the $55 thousand I had in my money market account” and there would be a massive bank run on money market accounts.

Years ago now, I used to sit next to the funding desk at Wells Fargo. I was part of a team that traded derivatives.   They were balancing the bank’s check book.   I listened as, day after day, Wells either lent or borrowed Billions over night. It would go in waves. We would dip into negative territory for 6 months, then pop back up to lending out for 2 months, then back to borrowing for 1, etc.

A default by Bear could easily have started a collapse of money markets. This would have led to a resulting collapse in over night lending, and suddenly a random 50% of the banks in the US would have a lot of trouble funding themselves tomorrow.

That’s what a “natural bottom” looks like. 50% of the US banks collapsing over night.

The Fed had to step in. The reality is that markets are not perfect, and they need regulating and assistance in times of trouble.

Paul Krugman at the NY Times has a nice little article on the inevitable bail-out that is beginning with tonight’s bail-out of Bear. The only debate now is how much to make the idiots who got us into this mess pay.

It’s either that or, we might as well all move to Zimbabwe. No employment opportunities, collapsed education, health care, hyperinflation, not much to eat. But, it’s actually a beautiful place. It’s a lot like Northern California in the summer time.

You’re a developer in Columbia, Brazil, South Africa or Turkey. You know a little PHP. How do you quickly become a world class entrepreneur? How do you reach audiences of millions? You have access to a computer, but you have you have no capital. What would you do?

At Joyent, we have partnered with Mentez to help app developers in emerging markets . Mentez starts with competitions to generate successful applications. It’s American Idol done in Columbia or Mexico and focused on good code instead of terrible singing. Mentez helps with everything from technical support to help monetizing the businesses.

How do you monetize an application in Brazil? Mentez has already succeeded by taking successful applications to corporate sponsors. For example, they have taken one game application that gained reasonable traction and sold a sponsorship of the app to a large well known brand. As part of the Mentez agreement, the developer received 50% of all revenue generated. The revenue was much higher than the low value CPMs the developer could have received by going through a standard Facebook ad network.

Bottom line, the way that apps get developed is changing, and it is benefiting developers in Emerging Markets.

• Joyent gives them free infrastructure
• Mentez gives them training and business development assistance
• Facebook gives these developers access to a massive and powerful marketing and distribution tool

You have read the white papers, seen the blogs, attended the conferences. And you have bought into the hype.

Enterprise 2.0 tools are going to make your organization an innovation machine! I actually believe that’s true. Or, at least, I think they can go a long way to getting you there.

So, now what do you do? How do you get there from here. You can fight with your IT department. Get buy in from the whole world and his dog. Literally buy a web server that is installed in a rack somewhere within your company’s data center. You can set up Apache. Set up MySQL. You can install the long dreamed about Wiki.

It’ll take a while.

Or, you can go the SaaS route and try to get your company security people to “Get it” about using a product from small web company. You data will be on their servers. If your data is financially related, or perhaps includes medical information, you will have to reassure yourself that they have both the systems security to protect the data and the, more importantly, the internal processes to protect your data.

There is another…. way

Or, you can go over to a virtual appliance directory like the one at VMware and find a virtual appliance that includes your SocialText Wiki. You download it, find a Cloud Computing Vendor, and install it. The virtual appliance includes everything you need. Web server, database.

Virtual Appliances + Cloud Computing = Hybrid SaaS

You get all the security and control of running an application within your own environment, but you get the flexibility to fire it up and try it with just a few individuals or a small department without having to go through the political BS of getting IT to buy off on every single element of your Enteprise 2.0 vision.

The Enterprise 2.0 Conference

Today, Steve Wylie, who is putting together the
Enterprise 2.0 Boston 2008 conference asked if Cloud Computing should be considered as part of the Enterprise 2.0 Conference.

I think it should. Offerings like Joyent’s Cloud Computer are already being used by small businesses, large companies and everything in between as an infrastructure to develop and deploy new solutions. These solutions often begin as experimental applications that emerge into major mission critical components of successful innovative new business offers.

Enterprise 2.0 is emergence software, and Cloud Computing is a great way of enabling that.

I used to be a fixed income derivatives trader. I have a Masters in Economics, with a specialization in Econometrics and Financial Derivatives Pricing. I work in technology now because, while in grad-school, I learnt how to code FORTRAN-77. If you wanted to be a quant, and you wanted to price exotic financial derivatives, you had to solve partial differential equations. The only way to do that was to use numerical methods that involved backward induction. And unless you wanted to spend weeks doing it by hand, you were forced to code up a pricer leveraging FORTRAN-77, numerical recipes and code from my brilliant thesis adviser, Robbie Jones.

I got bored with trading at Wells Fargo, moved over to building derivatives trading systems, and since then, have spent time at technology companies and at Ernst & Young. While at EY, I worked in the Financial Services Advisory practice. I have seen and audited the risk management systems at a half dozen of the largest financial institutions in the world.

The WSJ isn’t explaining much

I have just finished reading How to Lose $7.2 Billion: A Trader’s Tale in the Wall Street Journal by David Gauthier-Villars and Carrick Mollenkamp.

Is it just me, or has the WSJ already sunk in quality since the Rupert Murdoch purchase?

The journalists don’t explain much.   Was Jérôme Kerviel working on a desk that clearly allowed proprietary positions?   It isn’t clear.   Was it a pass-through desk dynamically hedging customer accommodations?   It isn’t clear from the article.

His other job consisted of betting on whether European stock markets would rise or fall. The roughly 20 traders on the Delta One desk were supposed to offset each bet that a stock index would rise with another bet in the opposite direction in order to keep risk at minimum levels. The difference between the parallel bets would generate either a profit or a loss.

That is a seriously terrible explanation.   Any trader worth his or her salt can take huge risks even with hedges in place.

There are 50 different ways to bet on an index.   What was this guy trading?   What kind of hedges were supposed to be in place?   What were the risk limits?   Were there notional limits?   Delta limits?   Vega?   Theta?

There is no mention of any of this in the article.   In other words, for people who really know this business, and honestly, it really isn’t that hard to figure out, the article doesn’t explain anything.

And for people who don’t know much about financial derivatives, the article gives a completely false sense that the truth has been explained.   It hasn’t.

SocGen was up by 500 Million Euros… but didn’t notice

The article does say that this trader had SocGen up by 500 million euros as one point, but that the bank didn’t notice.

I have one question only: “How?”

Derivatives trades may be complex bets, but they do result in real money flowing back and forth.   That real money comes out of real bank accounts.   Eventually, the CFO has to notice.   Something like

“Holy Crap!, we have 500 Million more Euros than we thought we would”

And, when your bets start to get into the Billions of Euros, if you are betting exchange traded futures, real margin calls start to happen.   If you are betting OTC derivatives, other banks, with half way decent internal controls, start calling you up and asking for more collateral.

The SocGen CFO and the head of Treasury should have noticed.

Fake Emails Fooled the Mid Office?   Come On!

Beyond that, the fact that any trader could produce fake emails to calm a back office is… well BS.  Here’s how it works at every major bank I’ve ever seen:

1. Trader does a trade with Bank X and enters it into the trade capture system
2. A back office person directly contacts their counterpart at Bank X and confirms the trade.   The variation is that the back office person gets a daily report from Bank X (or Exchange Y).
3. The second back office person sends a legal confirmation to Bank X, which is then sent back.

This double confirmation process makes it impossible to fake trades.   The process above is how it works for Over The Counter (OTC) trades.   It essentially follows the same pattern, but is much more automated for exchange traded derivatives.

And What About the Market Risk Management?

The article talks about all the fancy quants working at SocGen.   Without a nose for profit and a drive for blood, all the math in the world ain’t worth a damn.

But, quant geeks should be able to get you a half way decent risk management report.   And if not, you can buy one from BlackRock or some similar shop.

A stupidly simplistic report would show trades netted out and thus allow fake trades to mask the impact of real positions.   But no one with half a brain uses such simplistic reports.    You split the trades apart.  You find out were the notionals are out of whack.

- Look here, Bob is betting $10M on March DAX contract, and $50B on the June one.   Hmmmm

A $7.2B loss.   This doesn’t look like a rouge.   This looks like one poor patsy being forced to act the scapegoat for gang of bumbling arrogant fools.   Fire the CEO.   Fire the CFO.   Fire the entire trading management team.   Morons!

Here’s a simple question. I have a Apple Airport Extreme. It generally works brilliantly. I love it.

I want to open up my WiFi. Donate it to my neighbours. To anyone parked in front of my house. Anyone with an iPhone. Apple says that it can support up to 50 users. I am guessing few people would actually use it very much. And I am hoping everyone else gets the same idea.

There is a little bit of a problem. I know that people can run programs like aircrack-ng and hack into your network in a way that would let them see every credit card you enter, all your bank passwords, etc.

So, here’s my request to Apple . Please make a new version of the Aiport Extreme that has two networks on it. One that is open and one that is closed. Ideally, Apple could set it up so that the Extreme would give priority to the closed network traffic. Thus, no one would ever have to worry that donating their excess bandwidth would harm their own performance.

Why should Apple do this. Two reasons. First it would encourage the use of iPhones, because more free wifi makes iPhone users happier. AT&T’s dead slow network certainly doesn’t cause extreme joy. Second, it would give Apple users an excuse to show off their wonderful Apple technology is a socially cool way. No more would Apple users just show up with cooler laptops. Now they would show up with gifts in hand. An Apple Extreme owner would be a good neighbour.

I’d happily put this little badge in my window:

Back in October of last year, I looked at what inTrade was predicting for the results of the upcoming 2008 US elections.

There has been a lot of press and a lot of drama, but the result have not changed much.   In a winner take all prediction, here is what inTrade is  predicting now:

If you take the contracts with the highest price in each area, here is what the market is predicting today (Jan 18, 2007)

• Hillary Clinton is going to be the next President of the United States
• Someone from the Democratic field (not any of the Presidential Candidates) will be Vice President
• The Clinton / ??? Democratic ticket will defeat the McCain / Huckabee Republican ticket
• The Democrats will control the US Senate after the 2008 elections
• The Democrats will control the US House of Representatives after the 2008 elections

There have only been two changes.   First, the market is predicting that Obama is no longer going to be the Democratic VP candidate.   Second, the market predicting that McCain is going to lead a loosing Republican ticket, instead of Guliani.

The market behaved very strangely, and very poorly when predicting the New Hampshire outcome for the Democratic field.

The Clinton and Obama contract swung wildly in value, indicating a lack of “wisdom”.

inTrade is also currently predicting a recession in 2008, but only with a 70% chance.   My guess is that we are currently in one, so that might be easy money.

Oh, and to be open about any bias I have, I support Obama.

I love Highrise . For a small fast growing start-up, the workflow is wonderful.

The problem with most CRM systems, including Sugar and Salesforce.com is that they are not designed to help you get work done. They are simply designed help gather and organize data for upper management.

Here is a workflow that I do ever day, and something that is really painful in Sugar.

• Someone emails our support address. They clearly want to be a customer.
• I have to create an opportunity in Sugar - but I can’t because opportunities require accounts
• So, I create an account
• Then I create the opportunity
• Then I email the client using my regular email client
• Then I copy the email into Sugar. I could have used Sugar’s email client, but it is DOG SLOW.
• Then I create a follow-on task

This takes 20 minutes of data entry. In Highrise, it would be 30 seconds to cc the special email address.

It gets worse when I want to quickly review the status of each of our top accounts.   When you open an account in Sugar, you then have to click on every action item and click on every note just to see the current status.   In Highrise, you see that all right away.

Why aren’t we using Highrise now?   We’d like to extend it in some ways, customize it in others and integrate it with many aspects of our operations.   For our purposes, an open source solution would work better.

Sugar is open source, but it’s work flow sucks.   I would hate to have to build a clone of Highrise from scratch…

It would certainly be worth it to us to pay for a license to source code.   We’d pay more than it costs to on the site.

• What’s the second worst thing that can happen to a VC?
  They invest in something that fails.
• What’s the worst thing that can happen to a VC?
  They fail to invest in something that succeeds.

In the world of Web applications, every day the VC’s biggest fear gets more and more real.

It is a real eye opener for me to see how fast revenues at Joyent are growing. We’re private, so I won’t give you exact numbers. But it is exponential. Think hockey stick. And this is real revenue in the door, not vague proxies such as number of users. Revenue is growing because prices are good and people are getting real value from using our Accelerators to run their web applications.

How does this relate to the VC’s biggest fear?

There are two reasons.

• First, Joyent is itself a bootstrapped company. Imagine that. Cash flow positive, growing like hell and not a dime of VC money.

• Second, Joyent’s Cloud Computing offering means that more and more web companies will be able to bootstrap themselves into a successful revenue stream. If more companies can bootstrap, why do they need VCs?

One of my favorite web sites is Alarm:clock . Everyday they have stories about companies raising $600+K here and €15M there. Alarm:clock offers a great insight into real pulse of the technology start-ups. It’s also global in focus.

What floors me about the list of fundings, however, is the sheer number of companies that didn’t need to be funded. At least didn’t necessarily need funding yet. Right now, I do not have data to back up this theory. However, if I had data, it would look something like this:

There is some real interesting data that starts to point towards this. Check out this figure from a report by Cindy Moore from SVB Analytics.

If you eye-ball this, it looks likely that companies that were funded before Jan-95 were more likely to IPO. As I understand it, before the Netscape IPO, most VCs wanted to see that companies they were investing had a real product, real revenues and real growth opportunities. Things got bent out of shape when, post 1995, VCs started to simply invest in business plans and “ideas”.

Cindy Moore goes on to say “In Q2 2007, almost half of mergers and acquisitions with disclosed values produced total consideration for the sellers less than the total venture investment”.

Taken together, this implies that bubble and post bubble companies are less likely to return positive investments for VCs on average that pre-bubble companies.

My theory, as high-lighted by the fake data in the table above, is that companies are better off trying to bootstrap for as long as possible before they ever talk with VCs. In fact, it would seem reasonable to me to suggest that the only time you should talk to VCs is when they hunt you down and beg to invest.

I recommend reading Cindy Moore’s report. It’s called SVB Analytics Research Series Volume 2

BTW, I called Cindy Moore yesterday and asked her if she would consider producing a report on how long companies should boot strap before they talk with VCs. She said she would think about it. Please encourage her. Getting actual numbers for the Waiting for Success table listed above would be useful for start-ups and VCs alike.

This isn’t a real fear for VCs - It’s Good News

OK. I admit that I was doing a bit of baiting with that headline. The truth is that VCs are better off when more start-ups can boot strap themselves to a highly profitable business. Then VC capital is only used for the one thing that it is guaranteed to be good at: scale. VC capital can buy you scale, it can’t buy you a market.

Vote for Joyent on The Crunchies

Considering that Joyent is both a successful bootstrapped company, and that it has helped other companies successfully bootstrap, I’d like to request your consideration of Joyent when nominating The Best Bootstrapped Start-up on the Crunchies, 2007 .