Although you can get cost estimates using the Pricing Calculator, it helps if you understand some of the biggest factors that affect Azure costs. One of the most important is the region where Microsoft hosts your Azure resources. Many customers prefer to use resources in a region that is closest to them. For example, a customer in South America might choose to put its resources in one of Microsoft’s regions in Brazil. However, for many Azure services, the cost is significantly higher in the Brazilian regions than in the US regions. For some services, it can even be twice as expensive. So if network latency or data residency requirements aren’t important for your application, you might want to consider using one of the cheaper regions.

Another important factor is where you transfer your data from and to. When you transfer data to an Azure resource, it’s known as ingress traffic, and the cost is free. When you transfer data from an Azure resource, it’s known as egress traffic, and there’s often a cost. If you transfer within the same region, it’s free, but if you transfer to another region or to a location outside of the Azure network, then there’s a cost. So it’s usually best to put your resources together in the same region unless you have a good reason to split them across multiple regions.

One common question is how resource groups are affected by regions. For example, if you have a resource group in the West US region, and it contains resources in the Canada Central region, will that cause any problems, and will it incur any charges? Fortunately, it won’t cause any problems, and it won’t incur any additional charges. The location of the resource group is irrelevant, and resource groups themselves do not cost anything. The resources contained in them incur charges.

Aside from choosing your regions carefully, there are many other ways to reduce your costs. One way is to purchase reserved capacity. Normally, Microsoft uses metering to charge you for your Azure resources. That is, it tracks your usage and charges you based on what you’ve used. But if you know that you’re going to be using an Azure resource for more than a year, then you can save money by committing to a long-term contract.

Only some Azure services can be purchased this way, and the amount you save varies, but if you commit to a one-year or three-year term for a virtual machine instance, for example, you can save up to 72 percent.

When you make a reservation, you need to choose which region the VM will reside in. If you need to move it to another region, you have to make a new reservation and exchange the old one for the new one.

Whether you buy reserved instances or regular pay-as-you-go instances, you can also save money by using the Azure Hybrid Benefit. If you have existing Windows Server or SQL Server licenses, and they’re covered by a Microsoft Software Assurance plan, you can move those licenses to Azure. Then you can use those licenses to reduce your costs on Azure VMs, Azure SQL Database, or Azure SQL Managed Instance.

Another way to reduce costs is to use Azure Spot Virtual Machines. These VMs can be up to 90% cheaper, but there’s a catch. A spot VM can be removed with only 30 seconds’ notice. So you can’t use them for mission-critical applications. They should only be used for tasks that can be restarted, such as batch processing, development and testing, or video rendering.

That isn’t the only way you can save money on VMs, though. You can also resize underutilized virtual machines to a cheaper option. For example, if one of your VMs typically uses only 10% of its CPU capacity, then you could switch it to a VM size that has less CPU capacity, which would be less expensive.

You don’t even have to go searching for these overpowered VMs yourself because you can just go to the Cost tab in Azure Advisor, and it’ll tell you about sources of potential cost savings, including any underutilized VMs you’re running. It’ll also tell you if any of these underutilized VMs could possibly be shut down entirely.

Bear in mind that if a VM is stopped (which is also known as deallocated), it could still incur charges. You won’t have to pay for the compute resources since it won’t be running, but you will have to pay for any data disks and static public IP addresses that are attached to the VM.

Even deleting the VM won’t necessarily delete all of the associated resources that you have to pay for. By default, data disks and public IP addresses persist after a VM has been deleted, so you’ll need to delete them separately. Once again, Azure Advisor can help because it’ll tell you if you have any unused public IP addresses. 

And that’s it for reducing costs.