SaaS: A revolt against Enterprise Software

How many failed enterprise IT projects will it take, till too many people have died?

OK, the Bob Dylan reference may be overkill, but SaaS represents a revolt against the sins of enterprise software past. The challenge today selling enterprise software is that buyers are jaded; they have heard it all before, and they no longer believe. They want to see proof that you can deliver what they need, and they are extremely wary of projects with large up-front capital expenses.

What enterprise software deliversSeveral years ago while I was working at an enterprise software company, we had some great software, but it was incredibly expensive to deploy across the enterprise. It was highly configurable, which made it perfect for the early market of visionaries, but the configurability meant that it was overly complex and unmanageable for wide deployments.

We set our strategy to invert this cost and effort pyramid. When we deploy into these enterprise customers, there were many patterns that we should learn from. Further, the software should be able to discover its environment, and configure itself, or at least provide an intelligent and minimal list of options to enable a human operator to configure it more easily. Our motivation was only partially about the end customer! We knew that if we could achieve our goals on configurability, we would also achieve a competitive advantage and be able to sell more enterprise license agreementWhat enterprise customers wants. Which is what we really wanted.

We did achieve many of our goals, and I believe that all Enterprise software should undergo a similar kind of inversion. How many new features do customers really need? Yet how much do they complain about managing your software? This is partially why I believe design (not usability) is so crucial for software today.

The next step along this path is SaaS. Why have any infrastructure when none is (nearly) possible? Why provision infrastructure up-front when you can have it provisioned on demand? And why spend precious human capital learning about domains that are not core to your business, when there is a growing number of providers who will focus in these areas?

As David Gearhart points out, it won’t be all or nothing. But I still say: Enterprise software is dead. Long live Enterprise SaaS.


2 responses to “SaaS: A revolt against Enterprise Software

  1. The advent of SaaS is not simply related to the sins of the past. It has a lot to do with the continual maturing of the software industry and changing business models. (SFDC) was the clear early innovator here, and their success coupled with an appetite for change from corporations has helped fuel the SaaS market. But software has been through these changes before. From single user apps to multi-user apps. From client-server to 3-tier architectures and ASPs. And now, from enterprise software to SaaS. The models keep changing, and the software can do more, but it’s fundamentally the same things repackaged.

    I went to a SaaS conference earlier this year and heard Dave Duffield speak. Dave was the founder of PeopleSoft and he was talking about his new company Workday. See for some detail.

    Workday is essentially a SaaS version of PeopleSoft. It was interesting hearing Duffield spout about all the issues and costs in implementing enterprise apps and why SaaS was so much better. It was funny to hear him say this, given that he benefited significantly from those issues and costs.

    We’re at another change point in software’s ongoing evolution. After the hi-tech binge of the late 90s, then the crash in 2000 and 2001, there has been a long slow recovery for the tech sector. During that time, Web technologies matured (bandwidth, security, interactivity, robustness etc.) and enterprise software became bloated. Another “perfect storm” is brewing with the right conditions for more tech spending.

    One of the real appeals of SaaS is reduced risk — both financial and technical — for the customer. The try-then-buy model has a lot of appeal, particularly when the vendor is eating much of the upfront install/config/manage costs. But, switching costs still remain high once you’ve started using a SaaS application. How much effort and cost would it take for someone to switch from SFDC to NetSuite, particularly if they are heavy users of SFDC?

    In the end, it’s all about the business goals on both sides. Vendors smell the money, and buyers are looking for savings and value.

    Call me cynical, but in about 10 years, we’ll be looking at SaaS as some old model that had it’s benefits but whose time had ended. A new model will be proclaimed as king, and the same enterprise software will be reinvented again and sold to thirsty IT buyers.

  2. I am often intrigued by the “it is a revolution” argument versus the “same old thing” argument. What I like about it is that, of course, neither argument is quite true. What is meaningful is usually in the gaps between the two.

    The revolution argument is typically based on technology that has been already proven, in some cases for quite some time.

    The same-old-thing argument proposes that the universals still apply, like death, taxes and SQL.

    The problem is that it can be hard to get a useful handle on what is actually new. What is actually different? What is an incremental change? What is a leap?

    What we are genuinely fascinated by is the new — therein lies the opportunity, and the risk.

    For my part, I will stake this claim to what is actually new. I see the shift as this: computing withing many different disconnected private networks has shifted to computing over one massive public network. That is the ONE BIG CHANGE. Everything else follows.

    Other things, like XML or SOA or AJAX…not really all that new.

    XML? A text file with embedded meta-data? Yawn.

    As the man said, it’s the network.