Pacific Connection(英語)

Software without the Software: SaaS

この記事を読むのに必要な時間:およそ 11 分

Last year, Hummer Winblad Venture Partners got attention when it was revealed that 12 of its 30 funded companies were offering SaaS: software as a service. The ratio remains pretty much the same. The firm is invested in 11 pure SaaS companies, versus nine based on the conventional client-server approach. One startup, a security company, is involved with both.

If investor interest is an indicator, SaaS is the hot entrepreneurial category de jure. There are SaaS companies, companies to service SaaS companies, SaaS conferences, SaaS blogs and SaaS books. The best known SaaS provider remains, which opened its doors to considerable industry skepticism and has never looked back. The San Francisco company has taken on such entrenched players as SAP and Oracle in the customer relation management arena. Salesforce⁠s New York Stock Exchange symbol is CRM, and at this writing, the stock is trading at a P/E [profit-to-earnings] ratio reminiscent of the headiest days of the boom: 10,750. In other words, there are a lot of people out there who think the company is going to continue to encroach on the competition.

If Salesforce helped blaze the path for SaaS, smaller companies are looking to emulate its success―either providing SaaS or at least banking on the name. Ann Winblad has warned of “seeing companies coming up that are fake SaaS," as she put it at a conference last September. To be an authentic SaaS company, she said, the application has to be truly multi-tenant―meaning that a single instance of the software serves multiple organizations. A true SaaS application should offer access privileges and be able to integrate with other products. She has in general expressed a preference for products that deploy the LAMP stack: Linux, Apache, MySQL, and Perl/PHP/Python.

But perhaps Winblad⁠s definition is too narrow. What about Google Docs & Spreadsheets, two basic office-class applications that run not from a local hard drive but over the Internet via the browser? While few would predict that such hosted applications will soon replace Microsoft Office, both could be useful for collaboration: I⁠m in San Francisco, you⁠re in Tokyo, but we both have equal access to this spreadsheet. The advantages of collaboration via SaaS can also be seen in online accounting packages, such as Intuit⁠s Quickbooks Online Edition, in which a small company⁠s financial data can be reached from any wired location, which is especially useful for business owners who spend a lot of time in the field.

Timothy Chou, who once headed Oracle⁠s on-demand business and wrote the 2004 book “The End of Software: Transforming Your Business for the On Demand Future,” predicts that SaaS will eventually squeeze client-server out of the enterprise, not necessarily replacing legacy systems, but stepping in for new applications. In an August 2006 podcast, Chou said these were still the “early days” for the software service model, which is following a similar trajectory to PC software in the 1980s. Back then, “people forget, there were tons and tons of vendors.” We are also in a period where SaaS applications are mostly the online equivalent of client/server software. But Chou believes that long-term, vertical SaaS applications represent the future. “In the old days when the cost of gaining a customer was so high, you had to give them much more horizontal capabilities.”

Treb Ryan, CEO of the Santa Clara, California company OpSource, has been giving some thought as to how to succeed in the SaaS market. OpSource is essentially a SaaS company for SaaS vendors―providing infrastructure and services. The company sponsored the SaaS Summit last March in Monterey, California. “Anything new and interesting is happening over the Web,” Ryan says. That doesn⁠t mean that client/server packages are going to disappear overnight: “Oracle, SAP, and Microsoft have humungous embedded bases, mature applications, and years to go,” he says. “But the transition will occur over the next 10-15 years, especially as a new generation of people who are used to doing everything on the Web become more representative of business users.”

That bullish outlook seems to characterize the SaaS world, everywhere you look. Give us time, they say, and SaaS will replace client-server apps as surely as the mainframe has largely given way to networked PCs. This vision harkens back to the earliest days of the Internet itself, when researchers wanted a way to access the processing power of a remote computer. With broadband now widely in place, with Ajax blurring the line between local and remote graphical interfaces―and with IT⁠s job not getting any easier in this era of security breaches―why not let someone else host the application?

Ryan says that his most successful customers don⁠t even see themselves as software companies, but as Web companies who have business customers. He cites Webex, which was recently purchased by Cisco and specializes in Web-based meetings and presentations. “Many of these companies have a lot more in common with Google and eBay than they do with SAP and Oracle.” For example, he says, SaaS companies tend to upgrade more like a Web company―they make regular, more-or-less continuous updates rather than grouping changes in release cycles. “The process is very iterative, based on observing what users do―then acting on those observations.” Another difference, says Ryan, is that SaaS companies don⁠t think in terms of sales cycles―of having a sales team with quotas trying to land the next big sale. Instead, they concentrate on attracting a lot of people to stop by, sign up, and try out the site. “This doesn⁠t mean you don⁠t end up getting big customers, but the model really is different. A site will attract, say, 200 users from an organization. They become the internal evangelists who help deploy it across the rest of the organization―moving from 200 to 2000 subscribers.

I asked Ryan about what has traditionally been considered a sticking point for SaaS: the subscription model itself. Will people really pay a monthly or per-use fee for software services, rather than pay once and be done with it. Ryan argues that here, the consumer market is influencing the business market. As people become more accustomed to subscription-based websites, they are becoming ever more comfortable viewing software as a pay-as-you-go service―or at least one that charges per transaction. In the U.S., such models include eBay, where sellers pay on a transaction basis; Netflix, the subscription DVD service; and several subscription music services now competing against iTunes.

“I⁠m a firm believer that this is increasingly the way people are buying their technology―especially the next generation of users who were raised on the Web,” Ryan says. “They think about going online and doing things, not about long term ownership of the underlying intellectual property. They are saying: ‘I⁠ve got a function, I⁠ve got something I want to accomplish, I⁠m going to go online and do that. And I⁠ll pay for it in a number of different ways―by being exposed to advertising, or through a subscription or transaction fee.’”

Ryan notes that some companies are moving into SaaS slowly. Business Objects, an OpSource customer, is still clearly in the packageware business. On the company⁠s website, its applications are each illustrated with a shrink-wrapped box. But the company has at least dipped its toe in the SaaS waters with, a report-sharing service. The service is free to customers, with premium upgrades available for a fee. “That⁠s a very Web-type of thinking,” Ryan says.

Like Winblad, Ryan also thinks that Web services are having an impact on the SaaS market, though movement is still slow. “Someone will write an application that can be viewed via a Web browser, but the app is also readable to other applications via a Web services layer. Those applications can integrate that application back into their presentation to the end user.” The prominent example: Google Maps.

I asked Ryan what sorts of barriers to SaaS entry came up during the conference. “Integration is still one of the biggest problems for these guys,” he says. “While much is out on the Web, there is still a lot of legacy data that remains in the enterprise. Companies are asking: how can these two worlds meld together? How can someone who is using or do their reporting using information that still resides in a database that sits behind the firewall. Almost every one of our customers had issues about that. The panel we did on the topic turned out to be among the most popular. And it will be fascinating to see how companies are going about trying to make that happen.”

SaaS: the latest disruptive technology?

One of OpSource⁠s customers is Coghead, which provides an application development environment for non-programmers. “Our target audience is tech savvy business people who are much closer to the business problem than a professional programmer would be,” says CEO Paul McNamara. “And while they aren⁠t professional coders, they are comfortable enough with the technology to develop their own applications.”

McNamara has been around. He was working for the venture capital firm El Dorado when it was approached by Greg Olson, a Stanford PhD, whose company Versai Technology was researching Web-based architectures that involve a rich client and a back-end server. Olson joined as CEO, and the company reincorporated as Coghead in early 2006. Before that, McNamara was an early member of the Red Hat management team, and earlier, was at IBM around the time it was developing the PC. “I⁠ve been in a couple of interesting points of disruption,” he says. “I saw how the PC disrupted the mainframe business and how Linux disrupted Microsoft. In both of those situations, what you saw was technology enabling a broad class of people to do what only an elite class could have done previously.” He views SaaS and the Coghead service doing something similar.

I asked McNamara if he sees a link between the rush toward open source on one hand, and the raising visibility of SaaS on the other. “I do. I talk about the concept of ‘information factories⁠―places where sophisticated professional programmers take industrial strength raw materials in the form of open source components and assemble them into a useful information service that they provide to their customers. Examples are Google and Travelocity, which essentially are Web-based applications. That⁠s what we do. We use a bunch of open source components to build a Java stack. These are sophisticated, commercial grade, Web-scale components that are fairly difficult for an end user to deal with.” Coghead uses the Sleepycat XML database, now owned by Oracle, the Apache TomCat servlet container, and OpenLaszlo, an open source toolkit for creating rich clients. OpenLaszlo programs are written in XML and JavaScript, and compiled to Flash and DHTML.

Like Ryan, Olson sees the client/server software era giving way to the “webware” era. He uses the term emphatically: Olson is no fan of the label “SaaS.” “When radio came out, we didn⁠t call it ‘phonograph as a service.’ Radio could do so much more than just play records over the air. It could do news, weather, live programming―all things that were fundamentally different. Similarly, I believe webware applications will be fundamentally different from software applications. Software applications tended to be characterized by a fixed and rigid approach, where business adapted their processes to match the software. With webware, we⁠re seeing the ability to customize information services to match the way companies work. That⁠s going to be fundamentally different.”

The Coghead SaaS application is part flow chart, part forms―with the intent of allowing people who are close to business problems to create the application without having to know programming. But are there enough such people to build a business? Olson thinks there are about 30 to 50 million of them in the U.S., alone. “We call them the ‘tech savvy’ business guys. To find them, you have to look one or two tiers down below professional programmers. There, you⁠ll find people who are very comfortable with technology and who really understand the business problems. Today, they may be writing Excel macros or setting up blogs and wikis for their work team. We think those people are in abundance: they tend to be in operating roles or in consulting companies, where they are very close to colleagues or customers.” Coghead gives away single user accounts for free, then charges after that. A five-person account starts at $50/month. If you come up with a useful application, Coghead will pay you when other subscribers use it.

And what about a personal database like Microsoft Access―which in a sense is the desktop competition to Coghead⁠s SaaS model? “There are 9 million Access developers, so the Access market is large and healthy. But there are significant limitations around both it and Filemaker. First, they are inherently single user. Secondly, you build an application and it sits on someone⁠s desk where it is vulnerable to intrusion and lost data. Webware changes that model. You can add people to your application. We have a state of the art data center where the data resides, so you don⁠t have to think about networks and servers. All of that is transformative.”

Moreover, says Olson, “There are 50 million businesses in the world today that don⁠t have a server and will never have a server. So those businesses cannot be consumers of client/server software because they don⁠t have the infrastructure. We have a customer who is a rural farm co-op in the US, that is, a group of family farmers that have pooled their resources to handle payroll, health insurance and back office kinds of things. They don⁠t have a server room on the farm, but they are business people and need to do forecasts, manage suppliers, and deal with equipment maintenance.”

I asked him about business logic, the kind of thing Salesforce is doing with its Apex Code. “We have an expression: ‘simple is hard.’ We embed a BPEL [business process execution language], which is this really complicated, exotic, commercial-grade piece of technology. In the past, only the most sophisticated customers could even think about using BPEL. But we⁠ve done it in a way that⁠s really accessible to people. We said we⁠re not going to try and let people do every possible kind of process that they could possibly think of. Let⁠s constrain it and shoot for the 80 percent that everybody needs. For example, if your business has inventory, and if the inventory level falls below a certain threshold, you can pick out a purchase order and reorder inventory. Or, if an event doesn⁠t take place at a certain time, you can escalate to the next level of management. Those are the kinds of business processes that people want to do, and we give them the ability to do them.” beyond CRM

At, the company⁠s customers range from small ones like Zagat, the restaurant guide, which has 20 seats, to customers like Merrill Lynch, Cisco and Dell―companies who may have each licensed over 10,000 seats. These large installations show how a SaaS company can infringe on the traditional turf of the client-server vendors. “There⁠s no question that we⁠re capturing large customers from SAP and Oracle,” says Adam Gross, Salesforce⁠s vice president of developer marketing. “That⁠s a core part of our business. There⁠s also no question that on demand and software service is ready for the largest companies to adopt--and is actively being adopted by the largest companies. A good example is our win at Japan Post, with 4,000 to 5,000 users.”

Gross notes that the same product, Salesforce, is used throughout its customer base. By contrast, he says, SAP “has a product for big companies and another―which they are trying to make a lot of noise about―for medium size companies. These are totally separate code lines, different developers, and totally different sets of innovation or lack thereof. With Salesforce, there⁠s one version of Salesforce. Whether you are Zagat⁠s or Dell, you can have the same functionality. We think that is very democratizing because in the old software model, the small to medium business or enterprise version would be dumbed down.”

Gross says that even four years ago, the doomsayers pictured Salesforce and its SaaS model as a passing fad, one that would disappear. What did the critics miss? “It sounds kind of odd for us to say this now, but they missed the impact of the Internet.” Back then, there was no YouTube, FaceBook was just getting started, as was the iTunes store, and Google hadn⁠t gone public. But the age of widespread broadband changed everything. “Companies got lulled into a sense that the Internet wasn⁠t going to affect them.” As people realized they could put up a blog in minutes and have it read all over the world, or write an entry for Wikipedia and see it instantly become global conventional wisdom, their sense of what is possible online became heightened. Salesforce was a beneficiary. “Never bet against the web,” Gross says. “People would look at the impact of Ajax and say, ‘yes that⁠s all well and good for things like word processors and spreadsheets, but you⁠re never going to do things like online photo editing online.’ And then Adobe announces an online version of Photoshop.”

I asked Gross where Salesforce is headed beyond CRM. “Our mission is ‘no software.’ We are trying to bring the same kind of innovation, disruption and customer value that we brought to the CRM market to every other business application market. Our strategy is not to build those applications ourselves, but to entice developers to build those applications on our platform and make money doing so.” A step in this direction is the Salesforce Platform Edition, essentially a platform license that covers Salesforce applications―built in-house or by third-parties―beyond CRM. Another step: the company⁠s incubator, a facility that can house 32 startups engaged in AppExchange development, which recently opened just south of San Francisco. Others are planned for Tokyo, London, India, and, perhaps, Singapore.” Indeed, Salesforce⁠s strategy seems to mirror what Timothy Chou advocates: creating an intensely vertical set of services.

That would seem to open the door to small, entrepreneurial developers. “Developers like innovation, new tools and new toys,” he says. “We⁠re able to deliver the fun stuff ― the APIs, the ability to develop and code and program. And we take care of all the stuff they don⁠t want to worry about. I⁠ve never met a developer who wants to wear a pager.”


Bart Eisenberg

Bart Eisenberg's articles on the trends and technologies of the American computer industry have appeared in Gijutsu-Hyoron publications since the late 1980s. He has covered and consulted for both startups and the major corporations that make up the Silicon Valley. A native of Los Angeles and a self-confessed gadget freak, he lives with his wife Susan in Marin County, north of San Francisco. When not there, he can sometimes be found hiking with a GPS in the Sierra, traveling in India, driving his Toyota subcompact down the California coast, or on the streets of New York and Tokyo.


1980年代後半より,『Software Design』や『Web Site Expert』などの雑誌に,アメリカのコンピュータ業界のトレンドと技術に関するレポートを執筆しています。シリコンバレーで,スタートアップ企業から大企業まで幅広い分野でコンサルタントを務めました。

ロサンゼルス生まれで,自称ガジェットフリークです.現在,妻のSusanとともに,サンフランシスコ北部のMarin County在住。また,SierraのGPSを携えてハイキングしたり,インドを旅したり,カリフォルニア海岸をドライブしたり,NYや東京の街中を歩いたりしています。