Browsing articles from "February, 2008"

Why Software As Service Businesses Are So Interesting

It’s that time of year again. No, I’m not talking about spring cleaning. No, not taxes either. It’s the time of year when every MBA student thinks about his or her career plan. Today I had the privilege of meeting with one of Stanford GSB’s finest scholars. His number one question: should I get into venture capital?

While I won’t answer that question in this blog entry, I will talk about one of the items we discussed at length. Here’s what I like about SaaS companies:

  • Recurring revenue model. One of the most compelling aspects of any SaaS company is its revenue stream. Although it takes some time to develop, once a SaaS company has a customer base and has reached a steady churn rate, it has a predictable, repeatable revenue stream. Then it all comes down to continuing to acquire customers while reducing churn.
  • Low cost of sales. Another incredibly compelling aspect of the SaaS model is the low total cost of sales associated with these companies. The product is available for demonstration, evaluation, and long-term use right over the web. While some customer accounts may still require in-person sales, the vast majority can at least get up and running without an on-site customer visit. As a result, a great SaaS company can be very successful at a “land and expand” strategy.
  • Ease of delivery. Unlike the old installed software model, SaaS companies, like Consumer Internet companies, get to deliver updates when they want to. Customers lose some control in this model — whereas they used to be able to decide when they would roll out software updates, now the company delivering the software decides that for them. But companies benefit immensely because they get the most up to date capabilities without the overhead associated with traditional software delivery and maintenance. Companies can also deliver these releases more quickly and more efficiently than before.
  • Stickiness. Most if not all SaaS offerings require that the customer input or import data into the service. For example, a human resources management offering would likely require a customer to import its organizational data. Having the customer complete the initial work is critical, and sometimes challenging. But once the customer does this, they are likely to stick as long as they are getting some value from the product. Having gone through the process once, they are not likely to switch to another system. Moreover, extracting and exporting data from service based offerings is harder than it was with installed software.
  • Measurable growth. Finally, with SaaS business model, a company can really measure its performance. Once the revenue stream is established, a company can focus on a few key metrics. Monthly Recurring Revenue (MRR) indicates how much recurring (vs. new) revenue a company has each month from its customer base. Companies can track their customer Payback Time to see how long it takes to recoup the cost of acquiring a customer and getting that customer up and running. Then companies can focus on reducing this time.

It’s always fun meeting with the next great MBA graduates. Thank you to the student who came by today for the thought-provoking discussion.

Feb 25, 2008

S3 Goes Down – A Post Mortem

When Sun began using the slogan “the network is the computer,” it was about selling more hardware. But to mix metaphors, this is not your father’s compute cloud.

Today’s cloud is about delivering business critical services over the Internet. As a result, if there’s one thing to be learned from Amazon’s outage, it’s that operational excellence really matters.

Gorillas such as Amazon and Salesforce rely on their existing infrastructure to leverage cost and operating efficiencies. Outside of differentiating in what promises to be a crowded space, the real challenge for startups will be delivering on their operational promises.

Both DIY and partnering have their challenges. With DIY, a company has to build out the infrastructure and operate it, even before it reaches customer scale. But it has as much control as possible. Partnering means significantly lower cost up front, but also lower control and potentially lower margins at very high scale. A hybrid model may make sense — some companies delivering cloud computing services fail over to Amazon or other underlying providers.

Especially when critical business infrastructure is outsourced, transparency is key. One thing both Amazon and Joyent seem to have had in common was a lack of communication. Customers had no idea whether the problem was with the hosting provider, the network, or something else, nor when the problem would be solved.

As with many markets, there are two classes of startups entering the fray, those with a vertical approach and those with a horizontal one.

  • Vertical players typically focus on a particular technology as their point of differentiation. They may enhance the underlying platform by making it more scalable or more reliable. On the plus side, they can be more nimble in the short-term due to their focus. But technologies come and go quickly, especially on the Web. It seems like just yesterday that PHP was the new wave”¦ or was it Java?
  • Horizontal players are building out a wide variety of services. In the short term, it’s more difficult for them to execute as quickly. But as we have seen with Amazon, if a large enough community evolves, the community itself may lead the way in asking for and ultimately delivering the most leading edge services on top of cloud platforms.

Regardless of which approach succeeds, and there may be room for both, one thing is certain. No, it’s not that you should think twice about hosting in the same place that Twitter does (although that may very well be a good idea).

It’s that while cloud computing services are delivered virtually, customer expectations are very real indeed. Fail to deliver on those expectations — or worse yet, fail to communicate with your customers about what is going on and why — and face the consequences at your own risk.

Feb 20, 2008

Google’s Ad Revenue: No Real Surprise

“While announcing disappointing fourth-quarter earnings Thursday, Google executives said the company was having a harder time than it expected generating ad revenue on social-networking sites and figuring out the best ad formats for its YouTube video-sharing service,” the Wall Street Journal reported last week.

But the truth is, this really isn’t surprising. Eyeball aggregators like the large social networks have cracked the efficient user adoption code, but they’ve also known for years the simple truth that when you show billions of impressions, users will get burned out on them. Low CPMs are the well-known norm; they make it up in volume.

When it comes to video, there’s no surprise there either. Extracting enough contextual information from short video clips to present relevant ads in an automated way is incredibly difficult. Moreover, the correct video advertising user experience is still under development. After all, consumers just got used to skipping over ads on their televisions using Tivo. Plus, the necessary analytics and measurement for online rich media is just now being put in place by companies like Visible Measures.

There are several parts to a successful monetization strategy:

  • Advertising. Clearly this is the core way to generate revenue. While users suffer from seeing too much of the same thing, the fact remains that when you’re at huge scale, big brand advertisers can’t get enough of you. Soft drink companies, wireless carriers, car companies, and the like spend billions promoting their brands. Many sites simply aren’t at a large enough scale to gain the interest of these advertisers, but the large social networks are. Different forms of advertising are beginning to appear as well: digital coupons and product offerings that are contextually targeted provide the potential for monetization lift.
  • Payments for digital items. One behavior that has been widely established on virtual worlds sites is the willingness of users to pay for digital goods. These include items like clothes and property. In the social networking world, this translates into special badges, gifts, and other forms of “currency” that enhance a user’s status or allows them to interact with their friends in a unique way. As in the offline world, establishing the value of authenticity is critical. If a third party’s free gifts, for example, become seen as an equal to site provided gifts, the site’s gifts may become devalued. But users also value scarcity — items that can only be purchased in limited quantities, or are only available on certain dates, for example. Thus sites are beginning to drive very real revenue from digital items.
  • Mobile.Mobile is yet another interesting source of revenue for social networks. This is because mobile networks provide a built in way to monetize since a billing relationship has already been established with the user. SMS messages and digital downloads offer a vast monetization opportunity.

Social networks have the eyeballs. Where there are eyeballs, there are advertisers. Computers are great at determining context when large amounts of differentiated, long-tail text are involved. Videos and eyeball aggregators, however, are different.

It was years after Internet search was widely available that text-based contextual advertising became the de facto way to monetize it. Untargeted banner ads were the norm. Without user specified search for context, it should come as no surprise then that high value monetization is harder. The only real surprise is that Google thought it would be easier than it is.

But if there’s a silver lining in this cloud, it’s this: Internet entrepreneurs are hard at work delivering solutions to this monetization challenge. Digital goods payments, mobile solutions, and new targeting technologies are just the beginning of a new wave of monetization innovation. The surprise won’t last long.

Feb 10, 2008