George Milliken, Director of SaaS Hosting at Nimsoft.

There’s a rapid shift underway; infrastructure is moving into the cloud and applications are moving from on-premise to Software as a Service (SAAS). Software vendors, MSPs and enterprises are embracing the change as a means to cut costs, improve margins and serve new markets. The trend is not new, in fact, the Software & Information Industry Association first published their “Software as a Service: Strategic Backgrounder” in February 2001. This document defined 11 years ago the trends we’re all now experiencing.

Now that the trend is clear many organizations are moving rapidly to deploy SaaS offerings. There are some key pitfalls to be aware and avoid that affecting enterprise applications migrating to a SaaS model. This is true whether the target infrastructure is cloud based or hosted in a traditional data center. Over the last 11 years many best practices have identified, which will help MSPs and enterprises deploy SaaS successfully.

Here’s a quick overview of some key considerations and pitfalls to avoid.

#1 Remember you’re offering a service, not just software

Define the whole service, which is more than just the software. Start with the SLA and work into an architecture that takes into account uptime and other commitments you’re making to the customer. Calculate the cost of meeting those commitments and then determine your pricing strategy. Skipping one of these steps or executing them in the wrong order can lead to offerings that lose money and customer satisfaction issues.

#2 Define the service catalog

In addition to the SLA and customer contract you need to identify the service catalog. What can your customers request? Can they request backups? Restores? Refreshes? A test environment? Special configurations? List all items that differentiate your service from your competitors and then answer these questions:

• “How will I provide these items?”
• “Will I provide them through the software itself or will I provide these manually (via human labor)?”
• “How will I measure delivery of the whole service?”
• “What are my critical success factors (CSFs) and my key performance indicators (KPIs)?”

#3 Decide How to Handle Metering & Billing

Now that you have defined not only the SLA, customer contract and the service catalog, you must determine how you can measure, meter and bill your customers. This is especially important for those items you’ve decided to offer as part of the SaaS solution that may not be built into the software itself.

For instance, if you’re going to offer to store 5 gigabytes of data for your customers and bill for any usage over that amount then you need to make sure you have a mechanism to measure usage and generate bills. Failure to build this mechanism will lead to an awkward manual process or lost revenue and customers that consume disk space with impunity.

Examine carefully the service catalog items you defined during that exercise. Ask yourself these questions:

• “How often are we willing to perform each service catalog item?”
• “Will we charge for this item? If so, when and how much?”
• “Does the software automatically render each service catalog item and log the consumption of this item?”

#4 Identify your “as a Service” Gaps

Once you identify the service catalog you’re now in a position to examine the entire SaaS offering and answer some key questions.  Most importantly do you have software and automation in place to automate the entire SaaS offering? Can you automate these standard service catalog items?

• New Provisioning
• Upgrades
• Capacity Increase
• Decommissioning
• Environment shutdown and startup
• Customer outage communications
• SLA tracking & reporting

If the answer is “No” then you need to do some development work to fill in these “as a Service” gaps. Failure to build the necessary software and automation will prevent you from scaling and will impact your profit margins.

Experience shows that single tenant applications moved into a SaaS offering tend to have more as a service gaps than multi-tenant offerings. For example a single tenant enterprise application (like CRM for instance) may be multi-user but it doesn’t contain the software and automation needed to manage multiple instances of itself. That’s because the original developers never envisioned data centers full of these applications; they were thinking about a single instance sold to an on-premise customer.

All the things you need to truly render the service to multiple customers are what we call “as a Service” (aaS) gaps. These gaps are very expensive if you have to service them manually.

#5 Be Aware of the TCO Your Assuming

When you decide to provide SaaS to multiple customers you need to house those applications in a robust data center. Your customers want to avoid on-premise perpetual license fees and capital expenditures. You as the data center provider will need to carefully plan out the required infrastructure and make sure you can meet your service commitments and still make a profit.

Regardless of whether you scale in the cloud or scale your colocation facility it’s important to accurately calculate and understand the true total cost of operations for your SaaS offering. You will find that while there are economies of scale in building out your infrastructure there are also “scaling points” that can be quite expensive. You might sell your customers disk space by the gigabyte but you’ll have to purchase disk space from your storage appliance provider in large (expensive) chunks. (Just try to buy half of a NAS and you’ll see what I mean).

In summary, just remember that Software as a Service is about services, and not just the software that makes up the core of the offering. Think about the big picture from the customers’ perspective and you’ll be on the right road.

George Milliken is the Director of SaaS Hosting at Nimsoft. He has over 20 years of IT experience. Follow George online via Twitter and LinkedIn.