Do you still remember how much time and how many approvals you needed to order new hardware for your organisation 10 years ago? The entire process was managed, in many cases, by a single person working in procurement. Today, the cloud is a natural choice for hosting your solutions. Every developer can now order a new cloud service on behalf of the organisation. These actions often happen without any approval processes.

For the enterprise, it means that it is harder to predict and control their cloud infrastructure spending and manage their cloud costs. FinOps was introduced as a way for companies to deal with this challenge. It’s a cultural practice where every team member collaborates, takes ownership of their cloud usage and manages their cloud costs. The whole framework implementation can seem like an immense challenge for an organisation, especially if it’s already using cloud resources. If you’re only starting to play your cloud journey, I have a simple recommendation — onboard the FinOps approach from the day 1.

Source: https://www.finops.org/projects/adopting-finops/  

Some of the most commonly asked questions are: How fast can we see the first benefits of FinOps? Do we have to onboard FinOps fully to see the positive impact? Meanwhile, the most discussed benefit usually is the cost optimisation. While onboarding FinOps is a long-term, ongoing process, the first results can be seen quite quickly. Of course, it depends on your cloud solutions architecture.

The implementation of FinOps is a large process. It’s a cultural change that needs to happen in your organisation. For such a change to happen, the enterprise has to be well prepared. It would be best if you can run this transformation as a project. At the beginning, you will need to acquire knowledge about FinOps and create an initial plan for the implementation. The cultural aspect of it can’t be overstated. You should establish a community and find advocates who will educate people in your company for you. This group should propose a way of working and the first KPI. At the same time, they should investigate tools that can support your organisation in the FinOps journey. Later, find the first area for implementation, and incrementally spread the approach to the rest of the organisation. You should remember that this is an ongoing process and think about its continuous evolution.

You might be thinking that introducing FinOps to your organisation can take some time, and you have a problem that needs to be solved right now. This can be true, it’s a massive change and the pace of implementation will be affected by the people in your organisation, your current approach to IT and your organisational culture. For short-term goals, you can narrow down this approach and use only the elements that will help you achieve the first benefits and solve the urgent problems. However, you have to remember that this won’t get you full benefits of FinOps. The three steps that can help you satisfy the short-term needs are: define your goals, establish monitoring, and deliver quick wins.

Step 1 – Define Your Goals

Cloud cost optimisation is a heavily used term in the scope of FinOps. It’s natural because this aspect is easily visible to the organisation. Sometimes organisations over-focus on cost reduction and forget about business needs and the revenue that’s being generated. If the income is growing and your profit margin is not decreasing, then an increase in cloud costs should be acceptable for your organisation. You should only worry when your margin starts decreasing. If this happens, you’ll definitely need to take a closer look at your IT solutions.

Before you jump into the costs, you will need to define the business goals of your organisation. Then you can think about cloud strategy that supports these goals, and how like to monitor it. You need to focus on a KPI that measures the efficiency of your operations. Please remember that the most efficient operations are not always the cheapest. You can start with simple KPIs like:

  • cloud cost/revenue
  • cloud cost/business operation
  • cloud cost/customer.

Starting cost optimisation without alignment with your business goals is not a good idea. In the long term, it can cause more problems than benefits.

Step 2 – Establish Monitoring

The next step is establishing monitoring. You want to see how the changes you made impact your costs and defined KPIs. The first thing you should do is focus on your cloud spending. Before looking for an external tool, check the built-in solutions. Most of the clouds vendors have tools available for their platforms (e.g. Azure Cost Management and Billing; AWS Budgets, Billing or Cost Explorer). You can use them to see your actual costs, and historical values and group the data using different criteria.

To get more information from a grouping, you need to create a tagging strategy in your organisation. A well-designed tagging strategy will help you get more relevant information, and it will allow you to connect your costs with your business outcomes. When you’re thinking about tags, remember about organisation units, business lines, products, solutions, and regions. Use cloud policies to ensure that every new resource uses tags.

The next thing you should do is define budgets and alerts. They will both help you notice unpredictable behaviours of your cloud and react to those events quickly.

Source : https://azure.microsoft.com/en-us/services/advisor/  

When you establish monitoring, you can’t forget about infrastructure elements. Most cloud services already have some built-in monitoring capabilities supported by AI. They can help you understand the usage pattern of the service and allow for adjusting the size to your actual needs. Moreover, these tools can provide recommendations that can help you optimise your product and reduce the demand for resources.

Step 3 – Start With Quick Wins

FinOps assumes that you will set up a cross-functional team acting as a Cloud Cost Center of Excellence (CCoE) in your organisation. This is important because this unit can measure the impact of every, even the smallest, change on the organisation as a whole. Small optimisations are rarely implemented by individual teams because of the small impact on the teams’ individual responsibilities. However, when those changes are applied in multiple places, the impact on the entire organisation becomes much more noticeable. Moreover, these changes create new pathways for growth. In the later stage of FinOps adoption, the CCoE will be responsible for interacting with different parts of the organisation to define and manage cloud strategy, governance and good practices.

In materials related to FinOps, you will find a lot of recommendations about reserved instances. Their impact on cloud economy can be significant (up to 70% cost reduction). But before going along that way, there are several other options you can look into.

First, check for resources that you’re not using at all. Sometimes your team tests something, delivers proof of concept, or changes your architecture and nobody cleans it up afterwards. The easiest way to find these resources is to create a dependency graph and find ‘orphans’. Another good practice is to create a separate place for those temporary projects and clean it automatically every month.

In the previous step, I mentioned monitoring and recommendations provided by AI. Please review those suggestions. Some of them can be very easy to implement — add a key to the database table, turn on indexing, etc. In certain cases, those recommendations will increase the performance of your application, which will, in turn, decrease the need for resources.

Then, you should right-size your resources to your application and business needs. There are several different possibilities. You should start by changing the direct size of resources to your actual usage. The information gathered from monitoring can be help you with that. The decreasing size of resources can impact the included features. For example, it can reduce the availability or redundancy of services. Then start using autoscaling options and adjust the service to usage patterns. Think about periods in which your application is barely being used. Maybe you can turn it off for the night or use a serverless approach. You might also be able to create an environment for a specific task, run it and then deallocate everything later.

In the cloud, you are paying for booked resources, and it’s very easy to order new ones or return the ones you don’t need anymore. Automation is key, and it allows for supporting scenarios that are not built into the services. Please remember to include business stakeholders in such decisions each time.

After these simple actions, you should be able to see the first results in the form of cost savings. This is a good time to start working with resource reservations. In parallel, you can start negotiations with cloud vendors about special pricing offers based on your commitment.

Conclusion — Step 4 — Continue FinOps Adoption

I hope that the first visible benefits will encourage you to continue your journey towards FinOps. The cost savings often are the first advantage, and they can be achieved very fast. The quick wins I talked about can open new possibilities that will generate bigger benefits for your organisation. As FinOps adoption in your enterprise continues to grow, you will start noticing its increasing impact on strategic activities. It will help you focus on cost allocation in a way that supports your business model, improves the accuracy of forecasting and business effectiveness. As for the business stakeholders, they will be able to make better, more informed decisions that will help your organisation profit even more.