Compare Ebola to the common cold. Ebola kills its victims within a few days, so outbreaks are rare and contain themselves quickly. The cold, however, is a low level irritation. It’s far more successful in achieving its goal (reproduction.)

If the cold were to suddenly become deadly, it would not only limit its own success by killing its hosts before they could infect others, it would also inspire massive efforts to prevent its spread, further limiting its success.

The purpose of most worms like Conficker is to create a botnet for sending spam. The less impact it has on your machine, the less likely you are to notice it, the less urgently you’ll patch your operating system, the more likely it is to keep sending spam on behalf of its masters.

Zombies or Citizens?

Infected PCs are referred to as zombies. But do you call a person with a cold a zombie? Of course not. Because a cold imposes a much lower tax on its host than zombie-ism does on its victims: a few days of Oprah, rather than staggering, brain-eating and putrefaction.

In fact, a well-run botnet is not only less evil than it seems, it actually isn’t evil at all. It is simply a tax, and therefore amoral. A tax can be either good (levied in a way generally viewed as fair, and used for worthwhile purposes) or bad (levied unfairly and used for unworthy purposes.) But the fact of taxation is no longer even debated. With the exception of a few fringe groups, tax arguments are about the level of taxation and the use of proceeds, not on the legitimacy of taxes in the first place.

The Patriotic Botnet

So, could there be a good botnet? There already are. When millions of people download a screensaver that devotes spare CPU cycles to finding large prime numbers or to folding proteins, they voluntarily participate in a single-task, centrally coordinated network: a botnet.

So it’s a short leap to the following scenario: the US government decides to tax computing, so it distributes a worm that places a claim on 1% of the CPU cycles, disk space and network bandwidth of all US computers connected to the internet. The resulting botnet would be equivalent to a few million CPUs working in parallel – far more powerful than the largest supercomputer. These resources could be used to model nuclear explosions (reducing the need for nuclear testing), simulate global warming scenarios, perform basic research, tabulate population data, crunch tax returns, even test the encryption of a proposed electronic voting machine.

Would people find this acceptable? Not today. But remember that we have reached the point where the government intercepts 40% of your paycheck before you ever even see it, and the response is grumbling acceptance.

The Botnet Business Model

So we’ll reach acceptance of it at some point, and probably though commercial forces. Some entrepreneur will create a general purpose botnet that allows users to voluntarily donate computing resources to the effort of their choosing, and generates profits by re-selling some percentage of the computing resources it allocates. We’ll see this model in the next few years. In fact, it’s virtually indistinguishable from torrent-based filesharing. The government will simply tag along with a required plugin.

See you in the future! It’s a lot closer than you think.

• Technology risk — the risk that you can’t make the product you want
• Adoption risk — the risk that customers don’t want the product you make

In my experience, startups usually overestimate technology risk and underestimate adoption risk. There are two reasons for this.

1. Technology Risk is Declining

As the software market matures, technology risk is actually declining for all but the most ambitious projects. Simply put, there is now a lot of really good infrastructure in place (development languages, open source tools and platforms, hosting infrastructure, turn-key cloud computing platforms, and of course the worldwide web itself). Just as important, there are also a lot of really smart technology people who know how to use these powerful building blocks to make pretty much anything you can dream up. Don’t get me wrong, software development isn’t trivial or a slam-dunk, and many ships have foundered on those shores, but the risk is declining and generally over-emphasized.

2. Disbelievers Don’t Build Products

Passion for an idea is the bedrock requirement of being an entrepreneur. Entrepreneurs work with investors and employees who also believe strongly in the mission. After all, who would hire a marketing director, a developer, or a QA person who said “It’s kind of a dumb idea, but what the heck, I’ll give it a whirl.”?

This makes adoption risk that much more dangerous. It’s typically underestimated by the entire organization.

This organizational blind spot can play out in three different ways:

1. Bad ideas that should never have been funded
2. Good ideas that were flawed in execution (this is often cleverly disguised as technology risk, as the dev team scrambles to revise the product to match actual customer needs)
3. Well executed ideas without a clear path to market (or without a realistic level of funding to get them to market)

Fortunately, a well-run research effort can reduce the risk of all of these types of failures. More on that in a future post.

Nothing's perfect

(Part one of this series was called SaaS: First, the good news)

While there are many inherently good things about the SaaS (Software-as-a-Service) business model, there are also some unique challenges.

Higher cash requirements (Please deposit $24 for the next minute…)

SaaS companies can be very cheap to start, but because they tend to produce recurring (usually monthly) revenues rather than large lump-sum payments, it’s hard to use current cash to finance future growth. VCs often assume that it takes about $50 million to build a profitable SaaS company. Assuming a 4 year path to profitability, you’re burning through $24 a minute. This problem is made worse by…

…Lower price points (including FREE)

Because they’re typically based on a subscription model, SaaS companies try to maximize lifetime customer value through a low entry price and a long subscription duration. This approach, combined with intense competition and strong customer expectations that online content and services should be free or extremely cheap, means in order to reach profitability…

…You gotta have scale

If you’re entry price point is free (or something close to it) and you’re relying on some advertising revenue plus conversion to higher priced offerings (a conversion rate of 1% of your total customer base is considered typical) then you have to be able to attract and retain a very large number of customers before you’re in the black. That means your infrastructure and application need to be engineered accordingly, of course. But more importantly…

…Customer acquisition will be a major expense

While the online nature of SaaS suggests a pure online selling strategy, the reality is that SaaS companies tend to be very omnivorous in their drive to acquire customers. Between PR, search engine optimization, keyword advertising, webinars, email campaigns, reseller compensation, and other direct and indirect sales efforts, you could easily be spending 2/3 of your money on sales and marketing expense. To get a return on this investment, especially in the early days before your market matures, you’ll need to reach and convert a pool of savvy and influential people who prefer your vision of the future to that of your many competitors. In other words…

…You’re competing for the attention of Influencers, not just customers

Early customers are different. They know that others look to them for guidance about which technologies or services to adopt, and that their opinion counts. They will be very hesitant to risk their reputation on a company that doesn’t deliver a good experience and good value to their friends and colleagues. To be successful, you’ll need to build specific programs and efforts for reaching, influencing and recruiting these early adopters. These programs may be very different (and much more expensive, on a cost-per-customer basis) than your “typical” customer acquisition programs.

More on this in a separate post.

Whoopee!

There are a lot of great things about the Software as a Service business model. Here are 7 of them…

Cheap to start

SaaS companies can be very cheap to start because you can leverage existing infrastructure (Amazon’s EC3, Google, etc.) and great open source tools and languages to deploy powerful applications fairly quickly. This means you can keep your team small and agile while you figure out your offering and business model. This is closely related to…

…Faster development cycles

You can add functionality incrementally, see how customers respond, and react accordingly. It’s also fairly easy to run “beta” features side-by-side with production code, allowing the adventurous ones to get the new stuff early and give you feedback, without having to worry about upsetting the “just keep it working and leave me alone” segment of your customer base. You can also create a…

…Rapid “Find it, try it, buy it” sales cycle

This not only helps reduce your cost of sales, it’s a great way to understand what customers are responding to. You can test new messages and benefit statements, target new types of customers, even quickly roll out new offerings. This direct sales cycle, coupled with your faster development cycles, makes it much easier to…

…Stay close to your customers

Through methods both active (blogs, online surveys, email) and passive (measuring ROI on AdWords, reviewing log files to see what features are used most) you can get lots of information about what your customers like and want, and what they don’t care for. This not only helps you use your engineering resources efficiently, it means you can do a better job than the next company of meeting their needs and keeping them happy, so you can benefit from…

…WOM economics

Simply put, word of mouth is the least expensive and most desirable customer acquisition strategy. There are some very smart people who specialize in WOM marketing (for example, Seth Godin), but the key idea is simple: recommendations from existing customers are more influential than your own marketing messages, and cost a lot less money.

While we’re on the subject of money, SaaS companies are very well suited to…

…Recurring revenue models

It would be hard to convince someone to send you money every month for a piece of software they installed on a desktop computer, but customers are comfortable with the idea of paying a recurring fee for a recurring service. Part of the reason is simply psychology (customers don’t feel like they own the application), but another aspect is the understanding that you’ll continue to develop and enhance the system over time.

Recurring revenue is good not only because it’s predictable, but also because it compounds your growth tremendously. With the exception of churn, your year 2 revenue starts at your ending revenue for year 1. With traditional software, your year 2 revenue starts at zero. It’s also good for the customer because your interests and theirs are closely aligned. You have to keep them happy every day, not just when a new version comes out, or you’ll lose them.

This recurring revenue stream also helps you to…

Build partnerships quickly and cheaply

Recurring revenue is a powerful inducement to potential partners. You can both share in the wins, and no one has to pony up a bunch of money up front for something that may not produce a lot of results. It’s not only much easier to put together meaningful marketing, distribution or customer acquisition partnerships, but the technical advantages of web-based integration (usually much faster and cheaper than for installed software) mean you can deliver a wide range of functionality to your customers without having to build it all yourself.

So much for the good news. Next, let’s look at some of the inherent challenges of the SaaS business model.

(Part two of this series is called And now the Bad News… )