top of page

Cloud Cost Management? I can't afford bill is too high

What is Cloud Overspend?


Simply put it's spending more than the forecasted budget!  Slightly more difficult though, how do you budget for a dynamic cost model in the first place? Also just because you are in budget, doesn’t mean you aren’t overspending, just means you can afford it.


Public Cloud is not an investment model, there is no benefit in oversizing machines to meet future growth, the only way it makes sense is to use the applications that meet the needs of your business and size the infrastructure to meet the needs of the applications, everything else could be considered overspend.


Why is there Cloud Overspend?


Much of the reason for cloud overspend comes down to organisations failing to understand and accommodate the fundamental differences between the CapEx cost model and the OpEx nature of the cloud.  


Simply stated, the costs of running a data centre are front-end loaded into equipment procurement, but the marginal cost of operating that equipment once it’s up and running is relatively trivial.


In the cloud, there are no capital expenses.  Rather, costs accrue over time based on the size, duration and other characteristics of the workload.  That means budgeting for and managing cloud resources is a constant ongoing process that requires unique tools, oversight and governance.


What are the top 3 causes of Cloud Overspend?



  • This is really a Cloud Operations (CloudOps) issue, where the cloud instance selected for the application workload has too much resource and is therefore underutilised.

  • This is likely caused by one of three things:

    • Lack of understanding of the application requirements, to accurately match the instance type to the workload.

    • Lack of understanding of the instance type profile.

    • A legacy behaviour from a CapEx model where infrastructure was provisioned to meet the needs of the application through the depreciation cycle of the equipment.

Forgotten Instances;

  • This is where an instance has been provisioned against a workload requirement, the requirement for that workload no longer exists, but the instance is retained.

  • One of the problems here is responsibility, surely whoever provisioned the instance in the first place is also responsible for terminating it.

Storage Costs;

  • This can be for a number of reasons:

    • As we all know our storage requirements just keep getting bigger and bigger so we have natural growth.

    • We can also have the issue of orphaned storage, where storage has been provisioned as part of application workload requirement, the instance associated with the workload has been terminated, but the storage remains. Obviously in an OpEx model the user will be paying for this every month until it is deleted.


What are some of the Frustrations of Managing Cloud Spend?


  • Bill is over-budget

  • Too much data for me to analyse

  • Bill is too technical to understand

  • Different terminology used across cloud providers

  • Multiple systems required to analyse multiple clouds

  • Lack of visibility

  • Unused or over-provisioned services

  • Ever-changing ecosystem; bill is changing by the second

  • Expensive IT personnel cost to analyse bill

  • No budget constraints means no accountability

  • Cannot allocate spend to a division/cost centre/dept/project/person

  • No governance means no responsibility

  • Nearly limitless capacity, means potentially limitless bills

  • No built-in functionality to manage cloud cost

  • Integration and security issues in using 3rd party tools to analyse

  • No global view of cloud costs


What are the Implications?


One report of Aug 2023 states, that worldwide spending on public cloud is forecasted to reach $1.35 trillion in 2027. (IDC).  That’s $540 billion in overspend.  More than 90% of companies overspend by more than 40%.


What’s the solution?


There are fundamentally two ways of solving this problem;


  • Create a Target Operating Model and enforce compliance to prevent the problem from happening in the first place.

  • Catch the problem as quickly as you can after the event.


Whilst the target operating model is nirvana, in reality it is challenging to implement in a democratised environment, as you would need to get all stakeholders to agree on the model before implementing it.


OpExx has been designed to retrospectively identify problems as quickly as possible, as this is by far the easiest to implement, is non-intrusive and therefore invariably the most cost effective way of identifying anomalies.


OpExx is designed to build a common language between the cloud consumer and their chosen channel partner without the necessity to grant access to the primary cloud account. This quickly makes available the technical and financial resources required to avoid unnecessary overspend.


Unless you are one of the 5% of companies that appear to have solved this problem, return on investment with OpExx can be measured in minutes, as any costs will be seen as a percentage of savings.


To see a demo of OpExx please call Neil Gomes on 0203 856 8520, or email

34 views0 comments


bottom of page