by Colin Kingsbury
As the chief executive of a SaaS provider of talent management systems, I might reasonably be expected to disagree strongly with John’s recent three-piece j’accuse, revisiting the pitfalls of Software-as-a-Service (SaaS). In fact, I more or less agree with the substance of all of it. But that doesn’t matter because SaaS is still going to take over the world, and I’ll explain why.
The food here is practically inedible–and such small portions!
There’s a joke among boaters that there’s nothing more expensive than a free boat. John’s first major point that “free” often doesn’t mean “cheap,” or even “useful,” when the full costs of adoption and ongoing usage of technology are considered, is perfectly valid. While free and so-called “freemium” (where users pay only for more advanced features) models aren’t entirely new in the software industry, the ability to deliver sophisticated applications through a web browser has led to an explosion in the number and capability of such tools.
In my view, there is a vast and important difference between “freemium” tools like Dropbox, and completely free services like Facebook. In the former case, you’re a customer; in the latter, you’re the product. When Netflix introduced sweeping service changes that outraged its subscribers, management was eventually humiliated into walking back much of its plan. Ditto Bank of America’s short-lived $5 monthly debit-card fee. Facebook, however, has a more complex relationship with users, whose primary value is as targets for advertisers. The lesson is simple: If you want to be valued as a customer, then you have to be a customer.
Traveling alongside is the complaint “I am tired of feeling like understanding the latest toy is preferable to getting my job done.” The good news is, now you have a ‘seat at the table.’ The bad news is, this is what it feels like! If you could teleport a VP of marketing from the late 90s to today, he or she would be confronted with a world almost unrecognizable to even the most tech-savvy. SEO, SEM, PPC, and social media have replaced a century-long world of trade shows, printed brochures, and mass-media advertising. Sales and product development are constantly forced to evolve with shifting customer priorities. IT has been forced to reduce headcount even as the demands placed on it have exploded. Even legal and finance have had to keep up with a changing world. A full-time software geek can’t fully keep up with it.
But, if you think it’s bad now, just wait five years. While business hiring has been effectively flat since the economy began growing again two years ago, spending on software rose 26%, according to the US Department of Commerce. SaaS has driven the cost of software down dramatically (in many cases to an initial acquisition cost of zero), and it enables vendors to deliver innovation many times faster. Companies are choosing to substitute technology for labor. And when the price of anything falls, demand for it rises. But success won’t come simply by adopting whatever latest shiny toy flits across the front pages of the Wall Street Journal. Make your case by showing how a technology will, or won’t, deliver a measurable positive ROI. That’s an argument even C-level managers can understand.
Why lease the cow when you can buy the farm?
John’s second major point uses an analogy of deciding between buying a house (traditional on-premise enterprise software) and renting an apartment (SaaS). In this view, buying gives you stability and the ability to customize to your heart’s content. Renting is less commitment and gives you flexibility, but leaves you subject to the whims of the landlord. Leaving aside the problems of condos, co-op boards, HOAs, zoning rules, and planning committees, the point is fundamentally valid. But it tells us more about where we came from than where we’re headed.
One area where I think John is completely right is stability of user experience. SaaS vendors do love to push changes out system-wide. While we all think hard about not disrupting users’ lives when we do this, we’ve all been guilty of underestimating how disruptive any change is. While some of this is simply a cultural learning process for vendors, it’s also an area where I expect to see more technical resources deployed to enable vendors to make changes more incrementally on a per-customer basis. Once again, the experience will vary based on whether you are a consumer or a paying customer.
Regarding customizability, it’s easy to say that owning the code and the servers it runs on gives you more control, in principle. The devil is contained within those last two words. At one extreme, you have Apple products which typically offer afew personalization features. After that, your options are “take it” or “leave the store.” At the opposite end, the Linux operating system offers users access to the raw source code and all the tools needed to rebuild it to their personal specs. But this flexibility comes at the cost of extreme complexity, while Apple products are famed for “just working” with almost no user intervention. Customizability, then, needs to be considered both in terms of possibility (i.e., can it be done at all?) but also in accessibility (can we afford to set up and maintain this customization?)
While best-in-class SaaS products now routinely offer embedded scripting languages, APIs, and other tools which make them deeply customizable, on-premise systems do, at the margin, still support more aggressive alterations. But this ultimate flexibility comes at a price extracted at two different points. First, SaaS products, to the extent that they can be customized, are often much easier to customize than comparable on-premise products. Second, when a customer customizes an on-premise product, they often become responsible for maintaining their custom code and resolving any conflicts with future updates. SaaS vendors, by contrast, typically own responsibility for making sure that upgrades don’t break customers’ functionality.
For many smaller customers, SaaS solutions already offer effectively far-greater customizability simply because they can afford to have a power-user or ordinary IT staffer tweak and tailor the system, versus having to hire a consultant to extend an on-premise package. There is no fundamental technological obstacle preventing SaaS systems from offering the same ultimate customization capabilities as on-premise systems; that they are not there yet is due only to time. Ultimately, the vastly greater efficiency of R&D in the SaaS model will enable these companies to drive well past what on-premise vendors can support today.
One area where this is already happening is with integration across systems, something which remains difficult and costly for most on-premise customers. Some time back, my company needed a better solution than Excel and email for expense reports. A 30-second Google search yielded a SaaS product which cost $5 per user per month. In 15 minutes we had it connected to our accounting system through a cloud-based interface. On the user end, employees were able to enter their credit card information, and the system automatically downloaded transactions for import directly into their expense reports. This level of integration basically wasn’t possible ten years ago, at any price. Now, it’s so easy it didn’t even require an IT person. While hybrid SaaS-on premise models such as Intuit’s will extend the useful life of some on-premise systems, the long-term direction greatly favors SaaS. To the extent that integration across systems is the most powerful feature of all, this will be a huge driver of value for the switch.
When Buyers Attack
John’s last major point concerns the disconnect between vendors, buyers, and users. He writes, “They generally preferred to brainstorm new product features to gathering coherent market data. Their relationships were more likely to be with buyers than users.” And it’s true: successful vendors tend to focus on things that customers ask them to do, whether directly, or by virtue of their observed purchasing history. To bring this up in relation to SaaS gets it, in my view, almost perfectly backwards. If you ask a SaaS vendor what’s so wonderful about SaaS, it’s true that a lot of what you hear will sound like benefits for them rather than you.
While some of this is a recapitulation of the old feature statement versus benefit statement divide, there’s a much more important point behind it. More than anything else, the demand for SaaS was driven by the failures of on-premise systems: enormous acquisition costs, consulting budgets that equaled or exceeded license costs, customers who couldn’t afford to upgrade systems due to customizations needing to be rebuilt, development of information silos with systems that couldn’t be integrated, and worst of all, shelfware. Many if not all of these begin with the fact that enterprise systems, due to their cost and complexity, were sold to buyers rather than end users, as well as to the fact that vendors’ incentives were aligned mostly with winning the initial purchase decision.
No matter how many sins SaaS providers commit, the customer lock-in is, in the most important way, much lower. However, this goes more or less entirely out the door as contract terms creep to three years and beyond. While not every SaaS product can or should be priced on a 100% utility, pay-as-you-go basis like electricity, the shorter the feedback cycle, the more customer-focused the vendor must be to survive.
Perhaps the single most consequential long-term shift in the b2b software industry over the past decade is the transformation of users into buyers. Ten or fifteen years ago, most people, even well into the executive ranks, bought very little software. If you bought a home PC, you might get Office and a few other things. At work, you accepted whatever IT gave you. Very few people had significant practical experience evaluating software systems. Today, my 9-year-old nephew regularly buys software for his iPod on the App Store, as his 73-year-old grandfather does for his iPad. This cannot help but have a dramatic effect on expectations and requirements.
Likewise, we’re beginning to see new models of adoption. Products like Yammer or Dropbox are purchased by teams or even individual employees and slowly spread throughout large organizations, until reaching a critical mass where the IT department purchases a company-wide license. While we probably would never want to see a world where every employee buys their own payroll system, the trend to drive purchasing decisions closer to the people who have to live with them can only result in more user-centric software.
And it is true that SaaS today remains largely the province of groupware in larger companies and enterprise-wide systems for SMBs. But we’re closer to the beginning of the S curve than the end. The development of every industry is driven above all by the needs of its customers, and that is the foundation SaaS is built on.