| The SaaS Revolution - Part III: Making Your Move |
|
|
|
SaaS is good, but it isn’t for everyone Feature September 19, 2007 With Software as a Service now the fastest-growing segment of the business applications market, how should traditional software makers respond? CMU professor Bob Monroe, a veteran of SaaS pioneer Ariba, has been tracking the evolution of online service delivery. In this, the third in a series of four articles highlighting Monroe’s analysis, he explains how software companies can assess whether SaaS presents an authentic business opportunity for them. By Peter Longini, Managing Editor Let’s say your company makes really great software which you license or sell to customers. They take your CD out of the box, have their IT people install it on their PCs, data centers and in-house networks, and voila! Your company’s support requirements remain modest, and business is good. But you’re starting to see a rising tide of others in the industry using a different business model – Software as a Service, or SaaS, in which the customer remotely accesses software that actually runs on the seller’s equipment and for which the seller is paid according to the volume of customer use. So what should you do? Ignore it? Embrace it? Straddle it? Something else? Bob Monroe, a professor in Carnegie Mellon’s Tepper School of Business and former Applications Architect at Ariba, has been keeping a close eye on the evolution of SaaS-based businesses for some time. And in his analysis, the right response hinges on several key market and product factors. But any decision to move toward a SaaS-type delivery system carries with it a great deal of baggage that no one can afford to take lightly. Different Design For one thing, the architecture of the software itself will almost certainly be different for hosted solutions than for locally-installed ones. If you are creating it for just one customer, you would take a different approach than if you were writing it with the idea of supporting a range of different users. “There are fundamental, deep architectural decisions that get made very early in the system design, where you optimize one way or the other,” Monroe observed. “If you don’t build that into the beginning, it’s extremely difficult to retrofit it – very expensive, very difficult, and it tends to introduce a lot of bugs.” Take, for example, configuration. “If it’s an enterprise system and your system can be configured to support a variety of business processes and rules in that specific organization, you might build in the assumption that there’s a specific way in which this particular rule is done for this corporation. So you end up with lots of different points in your code where it checks for the value of that rule,” he said. “If you need to support ten different companies using this software on your servers, you need to be able to configure it so that there can be ten different ways of doing it based on who is logged in at any particular time, and which companies they work for. “There are hundreds of places in the code which make that check and ask for the status of that rule, and you’d need to go and change every one of them,” Monroe noted. “In a big piece of software there can be thousands of things like that which you’d have to redo. There are ways that you can build it in from the beginning, but it adds complexity, it adds costs, and it slows things down. If you assume it’s only going to be one company per installation, you don’t do those things; it’s a waste of time and money. However if you later decide to offer it to multiple customers, you’d have to come back and fix it, and that becomes difficult and costly.”
Go generic But suppose you’re prepared to accept those costs and consequences as long as the business really requires it. How, then, do you know if yours is really the type of software business that would profit from making the leap to SaaS? Not all of them are, according to Monroe. “The kinds of software that seem really well suited to a service model are processes with low-competitive value,” he said. “For example, legal compliance processes are not something you consider a competitive advantage, it’s a cost of doing business. You need to abide by these laws and there is software that will help make sure that you do so in areas like HR Benefits, Payroll, and Equal Opportunity. The laws are complex, but they’re pretty similar regardless of what company you’re in, what industry you’re in, or how big you are; the laws are the laws. And for most companies in your industry, how they go about meeting those requirements probably isn’t a big competitive advantage. “In the traditional model, you would probably just buy it as a package; you wouldn’t want to write it from scratch. And the laws are frequently updated. So instead of having to constantly update and reinstall the software, you could just get it online as a service. It’s not a key competitive advantage to your business,” he said. “The types of software that are winning early on with SaaS are ones that enable fairly generic processes – applications where you have a lot of potential customers.” But for core strategic applications, it works differently. “If you’re an airline, your scheduling algorithms are fairly important – how you schedule pilots and planes and crews and availability of airports. It’s proprietary, it’s important that you do it well, and it’s absolutely essential to running your operation effectively,” he said. “A yield management system would be another good example – coming up with a way to get the most dollars per unit of input that you possibly can. You’re probably not going to be eager to outsource that to any random provider. You’re probably going to want it highly-customized; you’re probably going to want it really tuned to your needs because it concerns one of the fundamental things you do in your business.” Ignore it As a result, any company contemplating the transition from shrink-wrapped software to SaaS has a range of options. “The first option would be to ignore it,” Monroe said. “There are two obvious situations where you would decide to ignore it. The first is that you decide you simply can’t develop that capability; you’re in a niche, you do it well, and building the capability to provide this Internet-based service doesn’t make any sense. You might be a device maker who can’t come up with any interesting ways to hook it into an Internet-based service. So you might decide to continue being the best maker of this device that you can be and just keep your eyes open to see if your competitors start coming up with a SaaS model. But if they don’t, just keep on doing what you’re doing. “A second reason to ignore it would be the demand side. If you can’t come up with any interesting or compelling reasons your customers would want this, then you probably want to ignore it. Operating systems are a good example; they’re essential, but nobody wants to download an operating system every time they use their computer – just the applications they want to run on it.” Grow a hybrid That doesn’t mean operating systems can’t benefit from SaaS, however. “Microsoft has actually come up with a very useful software service that’s fundamental to their operating system – Windows Update,” Monroe pointed out. “The operating system itself isn’t the service, but Microsoft has a real problem trying to distribute updates. If they had to send a CD out to everybody whenever they wanted to improve it, or patch a bug, or fix a security problem, it would be a nightmare. So they’ve built into the operating system the ability to send configuration information about your machine to their Windows Update servers which then run some checks and figure out which updates you need. Then it downloads them, and your operating system installs them locally. It’s a nice example of a hybrid SaaS model that makes a lot of sense.” Make the switch Another option is to drop the shrink-wrapped business model altogether and move entirely over to a SaaS business strategy. This is an extreme move, and one that may be irreversible, so it needs to be considered very carefully. “The time where you would want to do that is if it becomes painfully obvious that this is the model your customers are demanding – that it’s where the market is going,” he said. “Search engines would be the best example. Search is something that would clearly be a lot more interesting if it were done as a service across a broad set of computers rather than just a local thing running just on your machine. “You make the jump when it’s inevitable and you have to do it, or when you see that there’s a need out there that would be really well met with a Web-based service instead of installed software. In that case, you just go straight to offering the Web service – don’t even bother with an installed version. And you can probably build it a lot more quickly than somebody with a big investment in an installed version can. Think of the social networking sites – they are interesting as a service, but they’re not very interesting on the computer as a stand-alone. And people have been able to start them with very little initial investment.” Do both Of course, you could always consider doing both – presenting your offering as a locally-installed product as well as an online service. “X-Box Live is a good example of that, and all three of the big video game companies are trying it,” he said. “A game console is still fundamentally a device; you buy a DVD at the game store, you put it in, you play it locally. Microsoft has been very good at providing the service of letting you play against other players over the Internet by hooking it up to the service they call X-Box Live. When a new Microsoft game comes out, you can still go to the store and buy it. But they now make older games available as a service – you lease them, they’ll download what’s necessary onto your X-Box, you play them for a month, and after a month it erases itself. They don’t generally do that for the brand new games; it’s more for the five-year old games that stores won’t keep on their shelves anymore, but at least now the company can get $5 a month for it. “Another example is Salesforce.com,” Monroe noted. “Salesforce.com is much, much smaller than Oracle. But they had a very compelling value proposition: If you don’t want to spend millions of dollars to license Oracle’s CRM, that’s no problem; for $65 per user per month they will configure the system however you want it and it’ll run your business. For a lot of companies, that was very hard for Oracle to sell against. So Oracle was forced to come out with their own online service since it’s better to lose the sale of your big, expensive, enterprise-wide installed system to a less expensive SaaS that you offer than it is to lose that sale altogether to a competitor.”
About the Author: Peter Longini is the Managing Editor for Inside Product Strategy™. He can be reached at This e-mail address is being protected from spam bots, you need JavaScript enabled to view it .
To read the second article in this series: SaaS-The remote revolution, Part II - click here. |
|||
