This continues a series of columns from practitioners I respect. The category "Real Deal" describes them well.
This time it is Matt Haller, a Senior Principal at Baker Tilly, who leads the Enterprise Solutions group in its Management Consulting practice. He has been advising clients on strategy and implementing business systems for over 25 years, and has been working with cloud software since 2006.
BTW Click on graphics to enlarge
“We are often asked to build Total Cost of Ownership (TCO) models for companies considering on-premise and cloud business software. It is commonly believed that although the cost of on-premise software is higher in year one, costs are lower in subsequent years. Thus, over time, on-premise software is a lower cost alternative. This hypothesis is understandable; however, often flawed because the cost of software and performance upgrades are not considered properly. Although we call upgrade cost a “pinch”, they can take a huge bite out from the IT budget.
True multi-tenant cloud software brings significant benefit as software performance and infrastructure upgrades are performed by the software vendor as part of the subscription cost. Subscribers automatically receive functional updates, and have the ability to “opt-in” to these new features. As compared to on-premise software, this approach dramatically lowers and even eliminates the cost of upgrades, while minimizing business disruption. With on-premise software, upgrades can be a major event - with many upgrades requiring a full implementation cycle.
Even though upgraded software is included with the annual maintenance payment for on-premise, the upgrade process becomes a costly project, in many cases costing 50% or more of the initial implementation. There are many complexities including testing and re-programming modifications to the core software for compatibility, and making existing hardware, operating system, and database versions function with the upgrade. This cascading dependency becomes even more complex when multiple software products comprise the entire business system. An upgrade to one on-premise software product forces upgrades to multiple on-premise software products and their respective infrastructure components. Costs for upgrades further increase when factoring “soft” costs for the project team and business disruption.
Hosting and off-shore application management services for on-premise software have over time decreased in the cost of on-premise. However, these services lack the economies of scale true multi-tenant vendors bring with the inclusion of these services as part of the subscription.
Why the big TCO difference? Enhancements are simplified when performed by the cloud software vendor, or on a Platform-as-a-Service (PaaS), as the coding is performed on the same development platform using the same set of software code. By using the same platform, enhancements are instantly compatible. If third party software is part of the business system but comes from a cloud vendor’s application ecosystem (e.g. AppExchange), compatibility issues are eliminated. Cascading dependencies for hardware, database, and operating system do not apply to the cloud model.

The proper way to compare on-premise and cloud-based software is to examine cumulative costs and model costs at least 6-10 years. Most companies upgrade software every 3-6 years, and own software for at least 10 years. The on-premise upgrade cost creates major cost spikes in the years the upgrade is performed, (the “pinch” as illustrated in figure 1 above). Cumulative costs actually diverge over time. (figure 2 below) The gap in future years becomes more prominent as modifications, and cascading dependencies grow.
We applied these factors to a $300M automotive manufacturer considering an upgrade to their on-premise ERP software. TCO over 7 years for on-premise was nearly $13M. We found the cloud alternative was 46% less than on-premise, and 35% less than hosting the on-premise software.
When building a TCO model comparing on-premise to cloud software, make sure the model properly factors upgrade cost, and extends far enough into the future. Recognize that hosting is not the same as multi-tenant cloud software. Cloud software typically provides a very compelling value proposition.”

Matt can be reached at Matt dot Haller at Bakertilly dot Com.
The Real Deal: Matt Haller on the “pinch” when comparing Cloud/On-premise TCO
This continues a series of columns from practitioners I respect. The category "Real Deal" describes them well.
This time it is Matt Haller, a Senior Principal at Baker Tilly, who leads the Enterprise Solutions group in its Management Consulting practice. He has been advising clients on strategy and implementing business systems for over 25 years, and has been working with cloud software since 2006.
BTW Click on graphics to enlarge
“We are often asked to build Total Cost of Ownership (TCO) models for companies considering on-premise and cloud business software. It is commonly believed that although the cost of on-premise software is higher in year one, costs are lower in subsequent years. Thus, over time, on-premise software is a lower cost alternative. This hypothesis is understandable; however, often flawed because the cost of software and performance upgrades are not considered properly. Although we call upgrade cost a “pinch”, they can take a huge bite out from the IT budget.
True multi-tenant cloud software brings significant benefit as software performance and infrastructure upgrades are performed by the software vendor as part of the subscription cost. Subscribers automatically receive functional updates, and have the ability to “opt-in” to these new features. As compared to on-premise software, this approach dramatically lowers and even eliminates the cost of upgrades, while minimizing business disruption. With on-premise software, upgrades can be a major event - with many upgrades requiring a full implementation cycle.
Even though upgraded software is included with the annual maintenance payment for on-premise, the upgrade process becomes a costly project, in many cases costing 50% or more of the initial implementation. There are many complexities including testing and re-programming modifications to the core software for compatibility, and making existing hardware, operating system, and database versions function with the upgrade. This cascading dependency becomes even more complex when multiple software products comprise the entire business system. An upgrade to one on-premise software product forces upgrades to multiple on-premise software products and their respective infrastructure components. Costs for upgrades further increase when factoring “soft” costs for the project team and business disruption.
Hosting and off-shore application management services for on-premise software have over time decreased in the cost of on-premise. However, these services lack the economies of scale true multi-tenant vendors bring with the inclusion of these services as part of the subscription.
Why the big TCO difference? Enhancements are simplified when performed by the cloud software vendor, or on a Platform-as-a-Service (PaaS), as the coding is performed on the same development platform using the same set of software code. By using the same platform, enhancements are instantly compatible. If third party software is part of the business system but comes from a cloud vendor’s application ecosystem (e.g. AppExchange), compatibility issues are eliminated. Cascading dependencies for hardware, database, and operating system do not apply to the cloud model.
The proper way to compare on-premise and cloud-based software is to examine cumulative costs and model costs at least 6-10 years. Most companies upgrade software every 3-6 years, and own software for at least 10 years. The on-premise upgrade cost creates major cost spikes in the years the upgrade is performed, (the “pinch” as illustrated in figure 1 above). Cumulative costs actually diverge over time. (figure 2 below) The gap in future years becomes more prominent as modifications, and cascading dependencies grow.
We applied these factors to a $300M automotive manufacturer considering an upgrade to their on-premise ERP software. TCO over 7 years for on-premise was nearly $13M. We found the cloud alternative was 46% less than on-premise, and 35% less than hosting the on-premise software.
When building a TCO model comparing on-premise to cloud software, make sure the model properly factors upgrade cost, and extends far enough into the future. Recognize that hosting is not the same as multi-tenant cloud software. Cloud software typically provides a very compelling value proposition.”
Matt can be reached at Matt dot Haller at Bakertilly dot Com.
August 06, 2012 in Cloud Computing, SaaS, Industry Commentary, The Real Deal: Guest Columnist | Permalink