Every solution is limited by budget and investors look for maximal ROI. The solution architect needs to consider cost-saving during architecture design. Cost should be optimized from pilot creation to solution implementation and launch. Cost optimization is a continuous effort and should be continuous process. Like any other constraint, cost-saving comes with a trade-off; it should make a point of determining whether other components such as the speed of delivery and performance are more critical.

Often, cost increases due to over-provision resources and overlooks the cost of procurement. The solution architect needs to plan optimal resources to avoid excessive underutilization. At the organization level, there should be an automated mechanism to detect ghost resources, which team members may create dev and test environments, and it may no longer be in use after completion of the implementation task. Often, those ghost resources go unnoticed and cause costs to overrun. Organizations need to keep a record of their inventory by applying automated discovery.

During technology selection, it's essential to evaluate build versus source cost. Sometimes, it's better to use a third-party tool when your organization doesn't have the expertise and the cost of the build will be high, for example, by sourcing log analysis and business intelligence tools. Also, we need to determine the ease of learning and the complexity of implementation when selecting a technology for solution implementation. From an IT infrastructure perspective, we need to evaluate capital expenditure versus operation expenditures as maintaining a data center requires high capital investment upfront to meet unforeseen scaling demands. Since multiple choices are available, solution architects can select options from the following— public cloud, private cloud, and multi-cloud. Alternatively they can take a hybrid approach. 

Like all components, cost needs to be automated and alerts need to be set up against budget consumption. Cost needs to be planned and divided between the organizational unit and the workload so that responsibilities can be shared with all groups. The team needs to continuously look at cost optimization by optimizing operation support and workload as more historical data is collected.