SaaS (Software as a Service) is certainly not a new idea. If you include a more general model of processing for pay, it precedes the Internet by decades. Using a baseball metaphor, in an era when eye-popping capital investments were required to get you in the ballgame, not to mention the needed support, maintenance, and training costs to get you in the batters’ box, it may well have made better sense to hire a good pinch-hitter than to attempt to develop your own batting skills. However, things have changed! Oh yes, in an interconnected world, there is now a long list of components with various moving parts, but anything close to the standard array of tools is not going to involve the financial commitments of yesteryear. In fact, because there are many more relatively independent components, each may offer an opportunity to either save money, reduce risk, and/or increase efficiency.

Many are tempted to only look at direct costs when considering outsourcing some service, whether in whole or in part. TCO (total cost of ownership) calculations have to look much deeper than that. Of course, SaaS vendors and other service providers, like to factor in other costs in not always realistic ways to tip the equation heavily in their favor. Systems and application management overhead is something only you can accurately estimate based on observation in your environment. So what questions should you ask yourself to arrive a reasonable TCO or ROI calculation for a particular application?

  1. How much would it cost to purchase a comparable solution outright, including hardware, software, and both hardware and software maintenance costs?
    • How does this compare to the projected costs from the SaaS provider over the expected lifecycle of the proposed in-house solution? If it comes close, you may already have your answer, especially if you face cash-flow limitations that make the large initial capital investment more painful, not to mention the extra accounting and asset management tasks associated with owning something outright.
  2. How much time do you spend maintaining an application and the associated formal change management processes for modifications and upgrades?
    • If an application requires lots of hands-on management for user account management or constantly changing configuration data:
      • Does the SaaS provider offer additional tools to alleviate this pain?
      • Do they cost extra, or are they included in the base price?
      • If they cost extra, does it make sense only if you have a large number of users?  How does this affect your incentive?
      • Do they help one time, during initial installation, but then require on-going manual maintenance?
    • How frequent are major upgrades required? Bottom-line: the initial installation and configuration is just the beginning. Since the SaaS provider handles upgrades, this becomes an important part of their value proposition.  Attempt to quantify this!
  3. What skills are required to operate and maintain the application?
    • This is where risk mitigation may factor into your decision. If the skills required are somewhat rare, or they are not readily available in your geographic area, it might be difficult to recruit new employees should you experience turnover. If the required skills are not viewed as generally desirable, etc., then retention of qualified employees could be more difficult.
    • Are there training costs involved with operating a service in-house that would not exist, or be reduced, if you used an outside service provider?
  4. Are lost opportunity costs involved?
    • Are there other areas in which your business would be better served if you focused your attention there instead?
    • Are those other areas closer to the core competencies of your team? Re-tooling may not make sense.
  5. Does moving an application incur additional costs for resources to ensure adequate response time?
    • If an application is mission-critical, it may not be a good candidate for moving out of your data center. But mission-critical is almost a loaded adjective. Many organizations outsource e-mail, but still consider it crucial to their success. However, if your response time requirements impose additional costs that offset your other gains, you may want to think twice. If you have to triple your bandwidth to the Internet, purchase a new router and firewall to handle the extra load, then you had better be sure that there will be a payback for it within the projected lifecycle of the in-house service.
  6. Does moving an application, and, more importantly, the associated data introduce unacceptable risk?
    • Don’t trust someone else with your crown jewels! Where is your intellectual property? If a lot of it is in e-mail, or something like Microsoft Sharepoint, then perhaps those would not be good choices to outsource. If you trust your partner to do a better job of securing your information than you believe you can do, then perhaps, but have a clear understanding of can happen to you if that trust is ill-founded.
    • If the service holds PII (personally identifiable information), e.g. employee records, make sure you feel comfortable with the vendors’ security controls.  Do they make audit results available for your inspection?
    • How much would losing information cost you, either in cold, hard cash, reputation and/or confidence in your brand?
  7. Does the placement of an application positively affect those who use it?
    • If an application is primarily used by road warriors, then moving it to a location that provides better throughput and response time, e.g. as compared to your corporate Internet connection, will serve your users better.  This may not mean a SaaS provider though — you could choose to collocate that service in another data center.
  8. Does an application have interconnections to other applications?
    • If an application requires connectivity to your accounting system, or some other source of data, how will you make that information available to the application you wish to entrust with a SaaS provider?
    • If the application has many moving parts, all of which would be difficult and costly to implement in-house, then using an outsider provider that can coordinate and maintain these pieces becomes very attractive.

In a follow-up posting, I’ll look at a couple of example scenarios.