Perhaps one of the first centralized services in many organizations, email -- along with associated services such as calendaring and contact management -- is quickly becoming a staple cloud service with offerings available from many companies, including Microsoft, Google Inc., IBM and even VMware Inc. I'd be surprised if there was a company that has yet to consider whether to keep this service internal or move it to the cloud.
Working in higher education, the topic of email outsourcing is constantly discussed and is being actively undertaken by colleges and universities as both Microsoft and Google provide free services under Live@EDU and Google Apps.
Obviously, not every vertical enjoys free offerings from these companies, so money plays an important role in the outsourcing decision. Among the financial factors that play a role are:
- Infrastructure reductions. For some organizations, the hardware that supports email and related collaboration services cost millions per year. Organizations need to factor both server and storage space and associated power, and cooling and backup needs for equipment when determining the total financial impact of the hardware side of the service.
- Software licenses. It goes without saying that almost any enterprise-grade email/collaboration service carries software licensing fees. If you're using Exchange, for example, you need to consider the total cost of Exchange Server licenses, Exchange user Client Access Licenses, Windows Server licenses and Client Access Licenses.
- Staff time. Every system requires administration. When it comes to email and collaboration systems, this consists of managing the aforementioned infrastructure and licenses, but also includes managing the individual user accounts and servicing help desk requests associated with the service.
By moving to cloud computing, some -- yet not all -- of these items are eliminated -- but there does remain a workload need. On the infrastructure question, there is a clear return since an organization can eliminate the servers and storage that formerly supported the email infrastructure. Once that equipment is retired, there is also a resulting drop in energy and cooling bills. Software licensing may also be simplified now that there is one less service to include. On the staff time front, however, the return is probably less clear. Although the actual systems administration tasks go away, an IT staff's contract negotiation and management skill set needs to be beefed up in order to ensure that organizations don't suffer from poorly defined service-level agreements (SLAs), or worse -- from bad vendors.
Although the financial benefits may be pretty clear, organizations need to carefully consider the following factors before deciding to offload what is a business critical function:
- Bandwidth. Can the organization's current bandwidth support what could be a massive increase as a critical, continually used service is moved off-site? For highly centralized organizations, does the increase in bandwidth cost negate the financial benefit for outsourcing the service? For highly decentralized organizations that depend heavily on a distributed and mobile workforce, there could actually be a financial benefit to email outsourcing. After all, if the users no longer need to use the central pipe to access email and instead use a provider, the organization may be able to avoid bandwidth increases in the future.
- Business processes. How will business processes be affected with a move to off-site hosting? Is the existing email and calendaring system tightly coupled with other on-site services, such as third-party room or resource scheduling services, mobile device synchronization and customer relationship management systems? If so, these organizations need to undertake a process-by-process review to determine how to handle each and every need before pulling the trigger on an outsourcing contract.
- User training. If an organization is simply moving from on-premises to hosted Exchange and users will continue to use Outlook, the training issue is likely a moot point, although users need to be made aware that their experience may vary depending on the quality of the connection to the Internet (a move from LAN to a wide area network could have an adverse impact). If, however, you're making the jump from Exchange to Google Apps, users will likely notice the change and will need to be prepared for it, even if you use the Outlook Connector.
- Service criticality and acceptance of risk. Some organizations prefer to control their own destiny and not be at the mercy of a provider, no matter how solid the SLA. How critical is email to your organization? While major providers can offer more solid infrastructures, outages affect massive quantities of customers and organizations are left simply waiting with no control over the resumption of service. Although keeping email internal necessitates a highly available architecture, some organizations prefer it this way as they can react when a need arises.
Westminster College and the email outsourcing dilemma
As far as email is concerned, for now we've decided to keep email as an internal service at Westminster College. First, we're moving to virtualize anything and everything under the sun, and we've simply included Exchange 2010 in these plans. While there is still a resource cost in the form of licensing, server resources and storage space, it's not a significant enough sum to make a major financial case. Remember, Microsoft licensing for colleges is pretty inexpensive. Further, we're working hard to provide users with a highly integrated experience and feel that keeping this service internal helps us move toward that goal. Next, as a highly residential -- that is, very centralized -- college, we're concerned about the bandwidth impact and potential issues that we'd experience in the event of a connection or service outage, although these are relatively rare. Over time, we're working on moving to multiple providers for Internet access, but in a rural area, that's a bit of a challenge. Once we're at that point, the decision may change.
We've also put forth considerable effort into building a "private cloud" in our data center. While it will never be, on paper, as available as a commercial data center, we've experienced great uptime success (knock on wood!) in recent years, partially due to the redundant nature of our architecture.
Although there are obvious upsides to outsourcing, we're at a point at which the potential for downside -- in our opinion -- currently outweighs the benefits. Any organization considering undertaking a similar initiative needs to do an analysis -- financial and operational -- to determine need, fit and whether or not the expected outcome of the project meets the business need.