“Well,” said Peter. “On the other side of that door is the rest of our professional lives.”
The three were standing once more outside the door to Vincent’s suite. The fifteen slides were firmly in hand, their best shirts were on, and they had about five gallons of coffee running through each of them. In other words, they were as ready as they would ever be. Peter gave one dramatic sigh. “Well, no one can accuse us of taking the easy road.” And then he knocked on the door.
The three of them sat down at a large conference table that had miraculously appeared in the middle of the living room. All the other furniture seemed to have mysteriously disappeared. In its place was a group of very serious VCs. Paul leaned over to Peter as the VCs all slowly sat down at the table, cups of steaming fresh coffee in front of each of them. “Um, Peter? How come they aren’t smiling any more?”
Peter leaned back. “They are in deal mode. We no longer are cute little entrepreneurs. If this goes well, a large chunk of their investor’s money is about to walk out the door with us.”
Peter had just pointed out something that many people didn’t know. The VCs’ money didn’t actually belong to them. It mostly belonged to large investors such as counties, states and retirement funds. These “long-term investors” measured the success of their investements in years and decades, not months. The VCs were one of the tools that they used to spread around their money and diversify the risk. But the VCs were considered “high-yield” opportunities, meaning the investors wanted big returns for the money they plunked down into a VC fund. Something in the order of 30% return per year. This was in comparison to a risk-free government bond that would probably get you about 4%.
This caused mass consternation and many sleepless nights for the VCs. First off, they had to invest most of the money that was sitting in their fund in companies. The longer a pile of money sat uninvested, the harder it was to get that big whopping return. That meant that when a VC firm closed a new fund, the following action represented a stampede out of the stockades. The VCs shot out of their reserved slots on Sand Hill Road and blasted off in their expensive imported cars to find companies to invest in. Quickly.
Once the money was invested, the second problem arose. How to cash out of that company at the afore-mentioned obscene return? There were only two ways to do this: go public or sell the company. The latter was nice, but usually meant a lower return. It also meant that the VC had to distribute the gains to the partners in the fund, along with the resulting associated capital gains tax chaos. This was interesting, but still taking money permanently off the table and sticking the VC with a return they didn't want. There was however, the former solution, and it was reduced to one almighty concept for the VCs. So holy, so revered a notion that it was reduced even more until it became just one letter in VC-speak.
The letter was “x.”
The “x” was usually seen in the company of a number, and was spoken in hushed tones. “10x” for example. This meant that the VC has received a ten times return on the money they had invested. And the only way to get that kind of return was to get the company public. 10x was good, but 20x was better. And it went up from there, until you arrived at what was known simply as “a home run.” A big huge grand slam in the bottom of the ninth in the seventh game of the World Series. One home run could make an entire fund for a VC.
This brings us to another hallowed phrase for VCs: “IPO.” This stood for Initial Public Offering, the act of taking a privately held company public on a stock exchange such as the New York Stock Exchange, or its technology-loaded sibling, the NASDAQ. Now anyone invested in a company could sell their shares on the open market and realize their massive gains. This was good for the VCs. This meant they could get their home run and celebrate at Il Fornaio in Palo Alto.
But there was one fly in the ointment. The return was measured over time. That meant the VCs had to get the company public quickly so the annualized return didn’t crater and create a giant sucking sound in the middle of their fund. A company that takes twice as long to go public yields half the return over time. The target for VC-invested companies to go public was three to five years. This was the ticking sound in the belly of the beast. The IPO Clock. And every VC was aware of it every waking moment of every day.
Peter knew all about the IPO Clock. And as a result, the last slide of his presentation was the most important. It outlined the financial goals for the company that would get them quickly to that wonderful finish line. It was frankly the only slide that the VCs cared about right now. But Peter had to get there first.
At this point the VCs had all positioned themselves at the table. Paul and Todd got up and began circling the table in opposite directions, giving each person at the table a copy of the presentation. The one good thing about a city that never sleeps was that it was easy to find somewhere to get copies in the middle of the night. They made sure everyone had a copy, and then returned to their seats.
Peter cleared his throat once and then plunged in. “Okay, can I ask you all to go to the first page.” There was much rustling as pages were flipped over. “This is a diagram we call “The Life of A Packet.” We decided to start at the lowest level and then work our way up. In essence we’ll start with the technology and then work our way up to the market. The VCs were all staring at the diagram intently. Unlike Pat Barker’s firm, this room could not be fooled by technological hand waving. Many of the senior partners had started out as engineers and programmers. This called for Paul. “I’m going to ask Paul Cromer here to take over and explain some of the things we intend to work on.”
Paul did not look very comfortable, but at least he had time to brace himself for this moment. “As you can see from the diagram, what we are trying to address is the very glue that holds together networks, the information that passes through it. These structures are true whether you are talking about internal networks, on-lines services or the Internet. And what we are really concerned about is information that is shared.”
“In other words, distributed shared states.” said one of the partners. So far they were staying up with Paul.
“Yes, that’s what we are talking about,” continued Paul. “As you can see, in order for a piece of information to be shared, there’s actually quite a lot of network thrashing that has to happen. Most of it concerns who owns the thing, who they want to give the right to see it, who gets to edit it, stuff like that. And how many times you have to go back and forth across a network to keep all of this information synchronized between a large number of people. Right now there are a number of solutions to these problems, but all of them are network-inefficient and slowly get worse as more and more people come on to the network. What we want to create is a system of distributed objects that operate efficiently across large client-server networks.”
“That’s fairly ambitious.” said one of the partners. “Most of the work around this area is theoretical at best or wildly impractical. No one has been able to figure it out even in some of the largest labs in the country. Even if you do figure it out, how are you going to get it adopted in the marketplace? You are talking about everyone in essence adopting your datagram. I don’t think even IBM or Microsoft could pull that off.”
Peter held up his hand. “Actually, Paul is just going to give you an overview of the technology, and then Todd and I will lead you through the business slides and how we plan to enter the market. If we could hold these kind of questions till we get to the business section, that would be great.”
The VC nodded his head and fell into silence. Peter nodded at Paul to continue. Paul started spinning a pencil that was on the desk between his fingers. “As we thought about this problem, we started to realize that maybe there was another way to tackle this problem. And what we came upon was the idea of Phantoms. If I could ask you to go to the next page, I will explain to you what we mean by Phantoms.”
Peter was watching this part with rapt attention. If the VCs bought into this part of the technology, and the Phantoms passed the “sniff test,” then the following business discussion would fall neatly into place. If they resisted, it was going to be a battle all the way to the finish line. As Paul slowly went through the Phantom slide, Peter watched one VC after another. Many were scribbling furiously in the margins, which was good. This meant that they were probably going to wait with their questions until later. Some were listening, but obviously just waiting until the actual business slides arrived on the scene. Peter worked his way down the table until he arrived at Vincent. He was the most important person at the table. For all his tendencies to pithiness and sound bites, Vincent was strong on the technology side. If anyone was going to launch an all-out attack on their basic premises, it was going to be him.
Peter watched Vincent carefully. His eyes were half-way shut, and he almost looked like he was going into a trance. He then started, ever so slowly, to lean into the table. As Paul reached the end of each sentence, Vincent would ratchet one notch further. It looked to Peter like a cobra slowly getting ready to strike. He crossed his fingers under the table.
Suddenly Vincent broke into a sarcastic grin and waved his hands around aimlessly at the end of the table. “Wait, wait a second. I have to stop you for a minute here. What network protocol are you planning to use, TCP or UDP?”
Peter fired back the answer immediately. “We’re going to be running TCP. Otherwise we are going to be running into too many problems with firewall security. If we can’t run into corporations we are dead before we even start.”
Vincent smiled some more. “But then you are still going to be taking a hit at the packet level from TCP. You have to choose one of two evils. Either you put up with the latency of retransmits, or you risk throwing information away from dropped packets.”
Paul knew this one. “The great thing about the Phantom architecture is it allows us to suppress the TCP retrans requests without a degradation of security, ownership or rights. The most important information to us is not in the payload, but the header. It’s a different way of stitching together the information over the network. Since the Phantoms are shared, and not the underlying data, that means we don’t have to check it. Phantoms check the header, not the payload. See, we’ve taken the most redundant data and made it the priority. And since that is the case, our underlying server architecture can clock the data through faster.”
Vincent leaned back in his chair and stared at the ceiling for what seemed to Peter to be an eternity. He placed his fingers into a steeple, and then rhythmically tapped them back and forth. The room had become deathly silent. Everyone felt that this was a momentous occasion. Either the leading light of the firm would buy this, or the meeting was going to turn into a blood bath. Finally he spoke.
“This is really elegant guys.”
A couple of typos in the paragraph in which you introduce “x”....namely five-ten times? Etc.