WiredJournal

Just read the David Pogue’s review of the revamped .mac service - rebranded as MobileMe. The service essentially keeps all your e-mail, calendars, address books and Web bookmarks on all of your computers — Macs, Windows PCs, iPhones and iPod Touches — synchronized in real time. Synchronization through the cloud isn’t new… I remember playing with a service (fusion sync or something like that) back in 2000 that attempted this…

Of course, as with many technical innovations (VOIP - Skype, Search - Google), it sometimes takes a few market iterations for it to stick, and timing is always a key ingredient. Apple’s real innovation has always been delivering awesome interfaces to technology and I’m sure a lot of people will flock to MobileMe for the connivence of real-time, over the air synchronization.

CloudFire can also be grouped in the same category - as a cloud service for media. But coming from a P2P background, we tend to view the role of the desktop differently. In a service like MobileMe, the cloud is a centralized server farm that serves as the master copy of all your data. You are essentially leasing storage space from Apple and while 20GB might be enough for contacts, bookmarks, calenders (don’t know about e-mail), it will definitely fall short for photos, music, and especially video. You can always add more storage but 20GB @ $49/yr isn’t exactly cheap given that our hard drives and iPods are 160GB.

We don’t view centralized server farms as master storage but rather as a cache. Trying to store everything in the cloud is not only monumental (due to size) but also quite wasteful given that we typically access only our most recent content on a regular basis. Our approach is to proactively determine what’s recent and relevant, cache that, and let the rest be delivered from the user’s desktop with on-demand caching of older content. The cache is dynamic and ages, ranks, and recycles it’s content automatically. This way a 1GB cache could be as effective as 10GB of raw storage, at 1/10th the cost.

In a world where hard-drives and pipes are constantly getting fatter, we strongly believe the focus needs to shift from the center to the edges.

Structuring ownership among founders can be a tough thing to do - especially when you put together a team with different core competencies at different stages in the company evolution.

In technology startups, it is common for technical founders to feel that they put in disproportionately more work in the early stages than their business development counterparts. However, it is important to stay true to the word “partner” and irrespective of core competency, view all founders on the same footing. There has to be trust among founders and an understanding that everyone is highly motivated to see the startup succeed and will push themselves to the limit to get there - in their own different ways. So if you were starting a company tomorrow with 2 techies and 1 business person, you’d split the company 3 ways - 33% each. Done.

It, however, gets a little trickier if you don’t have all your partners on day one. You have to deal with some or all of:

• The initial founders accumulate more sweat equity than latter partners.
• The initial founders should get some credit for starting the company out of nothing - idea, vision, etc.
• You may not know how many partners you’ll find/need in the future, or even how big to make the pie.

We’ve had to go through this reality (I’m sure we’re not alone) and have used 2 models in the past that worked for us - a time based model and a people based model.

Time based model
This model is particularly suited for very early stage companies when a small group of founders come together to start something and enter the world of uncertainty. Uncertainty for how long development, funding, cash flow, etc. might take. The only constant here is time. An early stage startup can be very trying on first time founders and many will bail out. Vesting is one way to deal with this but it still leaves a mess of little stock grants that you have to clean up later.

The approach we chose was to defer any stock grants to a point in the future when a significant financial milestone was reached, such as a funding event or cash-flow positive event. This is similar, in concept, to waiting for a series A to value an angel investment.

We also believed in an all or nothing rule. If a founder leaves before the milestone event, he/she gets nothing. Harsh, yes, but perseverance and determination are some of the most important things tested in a startup. At the end of the milestone, each founder’s weighted contribution is calculated which is the percentage of founder equity they receive. This weighting accounts for latter partners but we also added additional points (not a lot) for being the first/initial founders and for pre-$ commitment.

Here’s what a simple cap table might look like for a startup that was started with 2 founders, added a third partner 6 months later, a fourth partner 9 months later and reached profitability after 12 months.

People based model
The time based model assumes that all partners are created equal but values contributions based on time. Sometimes this isn’t the case. For example a startup of techies might tinker around for a year, eventually hire a seasoned CEO, who immediately closes several big accounts. In this instance, the CEO gets a smaller valuation even though he/she was instrumental to the startup’s success.

The people based model uses the same milestone event and all or nothing rules but doesn’t use time as the basis for valuing contribution but rather dilutes partners on an ongoing basis with no weighting.

Here’s the same example using a people based model:

Month 1: 2 founders start company
Partner 1: 1 point = 50%
Partner 2: 1 point = 50%

Month 6: Add another partner
Partner 1: 1 + 1 = 2 points = 40%
Partner 2: 1 + 1 = 2 point = 40%
Partner 2: 1 point = 20%

Month 9: Add another partner
Partner 1: 2 + 1 = 3 points = 33%
Partner 2: 2 + 1 = 3 points = 33%
Partner 3: 1 + 1 = 2 points = 22%
Partner 4: 1 = 1 point = 11%

What’s interesting to note is that while the people based model does treat latter partners more favorably than the time based model, in the big scheme of things the two results are still pretty close…It’s far more important to come up with an open and fair model that everyone buys into than gripe over a few percentage points. The sooner the better…then get to work…

After my recent visit to the bay area, I reread Paul Graham’s most recent essay on “Cities and Ambition”. In it he discusses the messages that various cities subconsciously send, which in turn attracts like minded people. For instance, New York tells you to be richer, Cambridge tells you to be smarter, and Silicon Valley tells you to be more powerful (have more influence to change the world).

I’ve been wondering what message does Austin send… I recounted my top 10 reasons for picking Austin in a previous post… These were more driven from an outsider perspective but as Paul Graham outlines in the essay, you can’t really hear the city’s message till you’ve lived in it for a while.

Another test, is to see which class of people are most respected. For example, in Cambridge it’s professors, in New York it’s hedge fund managers, in Silicon Valley it’s founders. Who do we respect most in Austin? Athletes, musicians, film makers, founders ? Creatives, in general…

Austin is definitely more laid back and easy going than any city I’ve lived. I would say the message Austin sends is to live better. Agree/disagree, any others?

We’ve been asked a lot lately where the media files shared over CloudFire live… Do they stay on the desktop or get uploaded to some server(s) in the cloud? The answer is both.

Our definition of the “cloud” isn’t limited to just servers and cloud services like Amazon S3, but also includes end user devices like desktops, set top boxes, game consoles, and other appliances - which are really the primary sources of our media.

When you share your photos/video/music from your desktop, it is immediately accessible from any web browser and served directly from your desktop. CloudFire additionally starts caching your more recent and popular files in the background on to our servers and Amazon S3. This happens automatically and doesn’t require you to configure or specify anything. We do this to i) speed up the overall viewing experience, and ii) make your content accessible even when your primary machine is offline.

So when a request is made, say for a video file, CloudFire dynamically determines all the possible sources for it which might exist on one or more user desktop machines, our server’s local cache, or Amazon S3. CloudFire then swarms across all these sources to deliver the file to the browser…

Our raison d’etre is to simplify the steps to sharing and consuming media. We don’t make you manually upload, or reorganize files. All you need to tell CloudFire is what you’d like to share and it should take care of the rest…

Scratching your own itch is the most satisfying and effective strategy for building products - as you are your own best advocate and critic. In the case of CloudFire though, while I wanted to build it for myself, my wife really made me build it for her… which made it an even better product…as she is an even harsher critic than I…

I would have built it with a more “early-adopter” and “social” slant but my wife wanted just one thing:

To share all her existing and new photos and videos FAST!

Being new parents, we are now constantly hounded by friends and family for photos and video updates. We also all of a sudden have a new perspective and appreciation of FREE time and didn’t want to WASTE it UPLOADING. Hence, CloudFire was born.

Now getting photos and videos to friends, is simply a matter of plugging in the camera and syncing up the photos. CloudFire takes care of the rest. I am sure I will bake in some “social” features in the coming weeks but I love the simplicity and focus of the current app. We have opened up CloudFire to a larger audience and will let them help drive the product from here…

Whipped up my favorite pasta dish (Spaghetti Carbonara) for lunch today.

Recipe source: Molto Italiano by Mario Batali

This was my first attempt (second pasta dish ever) and it came out pretty okay. It was moist, well seasoned, and the pancetta was delicious. And of course, the egg makes the dish. I think the pasta could be improved but I’m always very picky on the basics. I tend to hone around a few dishes I love and manically try to drive it towards perfection while keeping a balance and continually moving into new areas.

It’s been quite a journey so far from “just the grill”, into the kitchen…

I’ve had my head down for the last month intensely focussed on the CloudFire private beta. Specifically, we’ve been listening intently to our early users and doing everything we can to make this app rock… Let me rephrase - doing as little and only what’s required to make this app rock… which ironically is more work than saying yes to everything…

We should be ready to pull the wraps soon…very soon… More invites will be on their way at the end of the week.

For me, unitasking is always super productive but a trade-off as other “softer” but just as important things get put on the wayside…like this blog…

I’ve had uVerse long enough to form an opinion… Actually, the internet portion didn’t take long at all. My wife really couldn’t notice a difference browsing the web the day after it was installed. The reason is that even though we now get 3 times the throughput on the downstream (close to 10Mbps), the latency seems to be the same and I could almost swear in some cases worse than before. Latency is what drives the browsing experience and gives the appearance of pages loading faster (or not). Of course, once you download a file, or stream video, it flies.

The TV portion took a little longer getting used to… The interface is better but a little slow too - latency seems to be a common theme here… Having a single HD stream is a little annoying. It means we can’t be recording one HD show and watch another. This is supposedly a temporary shortcoming.

All in all, the service is good… It’s priced very competitively once you start adding up all the little costs of leasing modems and extra set-top boxes (which is all included in uVerse), and it works…but it could be a whole lot better if the latency issues were addressed.

UPDATE: Sure enough, right after I said it “works”, we sufferred a uVerse brown-out last night at 9:30pm, where both the TV feed and Internet went out. It all came back sometime in the middle of the night.

Recipe source: Roast chicken and other stories by Simon Hopkinson

Over the weekend, I made 3 simple dishes from Simon Hopkinson’s “Roast chicken and other stories” book, which is a masterpiece of a cookbook (more on that in a later post).

The main entree was “Poulet saute au vinaigre” or “chicken sauted in vinegar” who as Simon points out is a great dish whose success lies in the fact that there are very few ingredients. I served it with a “Leeks vinaigrette” salad and a side of “Piedmontese peppers”. The only thing missing was something like rice or good crunchy bread for mopping-up purposes…

It’s now time to make a case for non-FREE. The best example that comes to mind is the SmugMug photo sharing service. SmugMug has never been FREE. They started with a 7 day trial taking your credit card upfront and have grown to 100,000 paying users and 10 Million in revenues as of 2007.

When asked about a no FREE plan in 2007 interview, CEO Don MacAskill, answered that when people actually pay for something, they inherently assign a much higher value to it. He attributes this as the main reason SmugMug is not littered with poor quality or inappropriate images - something competitive sites hire people to manage (full-time).

Going FREE never really was an option in the beginning, as SmugMug was bootstrapped, but even Don acknowledges that it was an experiment. Even though it paid off slowly and over time, it was readily apparent that they were on to something based on the fanatic user base that was slowly growing. Just that feedback and support was enough to keep going in the early days. They do have the option of supporting FREE now but still refuse to do so for the reasons outlined above.

Another bootstrapped business that DOES incorporate FREE is 37 Signals. They definitely list a much higher 1M user base but while a FREE option does create a lot more users, typically 80-90% of them are passive, and so true traction is a lot harder to gauge. I am not saying this percentage applies to 37 Signals, but I do have four 37 Signals accounts - one for each service that I don’t use. I can validate SmugMug’s 10M a year claim easily from their numbers but how much 37 Signals makes a year is anyone’s guess. Freshbooks also has a FREE plan but with a policy of automatically canceling accounts that have no activity in 2 months which is a smart way to keep a low passive to active user ratio.

The other nasty side effect of offering a FREE plan alongside paying plans is the irrationality factor. With BoxCloud, I’ve witnessed users doing some very crazy things to avoid paying for something that in end probably would have cost them just 25% of one hour’s pay per month.

Sure Smugmug will probably never grow as quickly as a Photobucket but it’s a real business that’s been built up incrementally versus one built on gathering the most number of eye balls which is both risky and fickle - one day it’s MySpace, the next Facebook.