52 Bicycles

Or something. One premise of SAAS type services is that computation takes place in the cloud, so software does not need to reside on the client. Single-instance, multi-tenant blah blah blah.

But we have these increasingly powerful computers with a lot of cycles at their disposal. SaaS may be a powerful trend, but there is going to be an increasing opportunity to utilize those local processor resources to do more work, either by re-integrating the local computer into the cloud computing infrastructure, or doing more user-centric processing. In this way one ca get a lot more cycles for users, at minimal cost. The question is how to deliver the software to exploit those local MIPS.

Douglas Burke wrote a neat article extolling the virtues of “Bootstrapping 2.0“.

Ask a group of Silicon Valley veterans to talk about their early mistakes and you’re likely to hear a boom-era story about a small, VC-funded start-up that failed because it was more focused on spending capital than finding a stable way to turn a profit.

I learned the hard way when I was brought in to help turn around a New York-based company during the dot-com bust. Although the founders had a good idea, they were better at raising money than taking the idea from the drawing room to the showroom. By the time I joined the company, they had worked their way through 10 different business models and over $20 million of venture and angel capital. Most of the money was spent on infrastructure, marketing and advertising.

Looking back on the experience, I can’t imagine it being much different from owning a fully-loaded luxury car without a steering wheel.

(my emphasis added)

This seems about right. Venture capital means you make your play, and speed is the order of the day. If you’re confident enough in not only your idea, but your particular business plan, this may be the right model - a drag race!

It’s possible that my biases are reflected in my favorite motorsport sport, F1:

• A focus on keeping the car as light as possible
• The key differentiator is in the brakes - braking late is what can let one pass in a corner more than the acceleration heading out of it.
• Top speed takes a back seat to handling (e.g. steering)  

Can definitely put that metaphor to work more, but  if you think your business has twists and turns yet to go, taking out the steering wheel is dangerous. That said, if the track is straight, you’d trade that wheel for rocket fuel.

Hoping that this will get me back in the habit of writing. Appreciate those who decide to take a read after my multi-month hiatus.

Byron Deeter at Bessemer has Top 10 Laws for Being “SaaS-y”:

1. Your key business metrics are: CMRR (Contracted Monthly Recurring Revenue) and Cash - “Bookings” is for suckers.
2. It takes at least $300K of CMRR to climb the Sales Learning Curve - Stop at three sales reps until at least two of them are making $100K MRR quotas.
3. Separate your “hunters” and “farmers” - As soon as you’ve climbed the Sales Learning Curve, begin ramping your sales force by hiring renewal-oriented account managers. Keep the hunters moving, and let farmers tend to the crops.
4. It’s a whole new ecosystem - Channels are very hard for SaaS companies to build, so don’t base your plan on SIs and traditional ISVs. You will need to sell directly for a long time.
5. Stay local - Prove your business in North America first. Only after reaching $1M in CMRR should you consider hiring European sales and services execs behind customer demand. Save Asia for post-IPO.
6. One datacenter - Invest early in backup and disaster recovery, but stick to one data center, at least until well after IPO.
7. Single instance, multi-tenant - Have only one version of the code in production. Really. “Just say no” to on-premise deployments.
8. By definition, your sales prospects are online - Savvy online marketing is a core competence (sometimes the only one) of every successful SaaS business.
9. Constantly trade off cash vs. growth - If you must replenish supplies while still crossing the desert, optimize your growth rate (sales rep recruitment and marketing spending) so that you maximize your recurring revenue run rate when you need to fundraise next.
10. Be prepared to cross the desert - SaaS requires R&D and sales expense up front for a multi-year stream of revenue, so it demands enough investment capital to fund 4+ years of runway. Load up for the long trip and pace your consumption of calories!

He suggests that a company can skip one of these rules, but the others are pretty hard-and-fast. My notes as a business attacking this from a different perspective:

1. Beware rule 9 - cash buys time, so don’t go spending it like water when crossing the desert. His counter is likely that one should get the supplies you need before starting the crossing, which is an argument for major up-front financing - i.e. VC. If you are not going to get VC, keep spending down and replenish through revenue.
2. The fundamental assumption of rule 10 is that this is a major front-end investment for a back-end annuity. However, SaaS and software in general has a low point of entry, so consider that you can make a splash with limited front-end investment, and that the security of the back-end annuity is in jeopardy. A successful firm is constantly adjusting, so one should preserve cash to facilitate those adjustments.
3. Finally, if the model is back-end revenue stream, one should shift one’s cost structure to the back-end. This means minimizing the cost of customer acquisition, and the cost of handling the first customer. Scalability means the ability to expand capacity when it is needed - achieving that is more about flexibility than building a huge infrastructure up-front.

Of course, there are risks to the bootstrapping approach - if you grow too fast, you will outgrow your limited infrastructure. But that’s a high-class problem ,and software can be adjusted relatively rapidly and for only moderate expense to accomodate scaling. Further,  if you are in that position, getting financing at more attractive terms or valuations will be an option, because it’s muchBut  easier to invest in expansion than starting.

But either way, check out the article, because it makes many smart points and is asking excellent questions.

Dharmesh has another always-insightful post, Why Startups Fail: Run Out Of Cash, Run Out Of Commitment, which is leading him to ruminate on VC vs bootstrapping. He comes to an “interesting idea:”

Perhaps startups should simply be trying to give themselves enough time to figure out what will work. And, the time available is not a function of the amount of cash raised, but the amount of cash being consumed. Profitable startups don’t consume cash — they generate it. Hence, they’ve got more time.

Absolutely. VCs often talk about speed and growth at all costs, but that’s the nature of their investment profile. For the entrepreneur looking to succeed, one needs to survive, and live over time.

Dharmesh says that cash and commitment are the keys to time - get a lot, spend a little. Good advice.

Alex Muse has a new post, Getting Small Fast Revisited, that illustrates three requirements for investors that are all about focus:

1. Investor Requirement #1: You’ve got to know your market
2. Investor Requirement #2: You’ve got to have a practical business development strategy
3. Investor Requirement #3: You need a plan that you can execute

His point is that all three of these require a lot of focus, and that focusing on a very limited, small market is the key to getting started. Good advice.

It occurs to me that advertising has a high beta - it does really well when the economy is up, and totally craps out when the economy dips. Brad Feld’s musings on What Happens To CPM’s When The Economy Turns Down? made me think about my friends running businesses predicated on the idea of selling ads on their platform, and what would happen if the market for that inventory declined. A secondary revenue stream to monetize the userbase that either counters the market (as a hedge) or has a positive alpha (evergreen) is called for.

I am apparently missing FOSSCamp because of my high workload, but I offer these notes from the excellent MIT Enterprise Forum Event on Open Source from last year, excerpted from my post MITEFC Open Source Conference Notes

Open Source licenses gemerally fall into in three buckets (Props to Simon Philips for his clear description):

1. “A”: market forming models allow free use, modification and redistribution. Examples: BSD, Apache.
2. “B”: Community based (or file-based) models allow free addition, but restrict modification. Such that adding a new file is external to license restrictions, but changing a file triggers the license (i.e. must open-source the changes you made). Examples: Solaris, maybe LGPL?
3. “C”: Competitor threatening: any innovation needs to use same license: GPL

Almost all open source licenses only trigger when you redistribute - so you can use your modifications or addons internally all you like, as long as the code doesn’t leave your shop.

The previous phenomenon may be a big driver of SAAS/on-demand - a legalistic driver for this new delivery mechanism, rather than the economic/technical drivers often discussed.

Business models for open-source

1. Support: give away s/w, sell service contracts, subscriptions. Examples: JBoss, RedHat (Is this really an insurance product?)
2. Dual License: Have a second license that allows commercial use/redistribution. Example: Mysql
3. Two products: Have a free software product that drives value of a second, proprietary software or content. (interesting)
4. Hosted: offer as a service/on-demand (per the previous main point above)
5. Hardware sales / appliance (give away FOSS, sell all-in-one (IBM’s motive?)

Hope all the folks are doing well at the Hilton @MIT, just one block away from my apartment!