When it comes to cloud every organisation wants to reap the promised benefits – cost-effectiveness, stability, and high levels of performance. It also enables businesses to take advantage of innovative technologies; think about the possibility of using large language models (LLMs) without the cloud. However, not everyone is realising the value of cloud.
The reason? Cost.
Managing cloud costs can be problematic, and to be more efficient and truly gain the benefits of the technology, organisations need to understand the drivers behind those costs.
In essence, this is our approach to building a cost-effective organisation. We focus on the sustainable nature of cost optimisation by analysing past data, embedding efficient behaviours, and predicting future spend.
Our methodology is based on The Frugal Architect’s approach to software development, which was created by Werner Vogels, CTO at Amazon. We look at how to solve the problem of out-of-control cost growth, how to speed up development, and how to take advantage of the immense potential of cloud.
Awareness
Your teams need to know how much they are spending on cloud. More than that, they need to understand that spend in relation to the overall IT budget or business revenue. We recommend focusing on spend relative to the IT budget as looking at it in terms of business revenue could present a false sense of efficiency as cloud spend could be eating into the IT budget, which is a problem.
In addition, your teams needs to have a broader perspective – that is they need access to data about previous years spend and keep an eye on predicting future spend.
Taking that down a level, engineers need to know how much their own service or system fits into the overall cloud budget.
And finally, they need to know how their cloud costs affect performance of the service. For example, too much cost suggests the service is not delivering as expected and increases the risk of reliability issues in the future.
There are cloud cost management tools that can help – such as Cloudability, CloudHealth, AWS Cost Explorer, and Azure Cost Management – but if your teams have the awareness of costs, they can better use those tools and get the outcomes they need.
Time/Prioritisation
Teams need to prioritise the activities around cloud cost optimisation – looking for unnecessary capacity usage or cost which does not have an impact on the reliability of the service. Importantly, they also need the time to carry them out.
The same goes for housekeeping activities which are often neglected. They are, in fact, vital in cloud cost optimisation and in helping your engineers understand the costs of the cloud services and how they work. Building these activities into existing sprint cycles or creating dedicated cycles for them can help.
Observability
Businesses and teams need the right level of visibility of cost, performance, and utilisation data over both the short and longer term. This in turn is indicative of the team’s understanding of and confidence in their systems. To put that into perspective, a less confident team might log everything, use logs rather than metrics and keep data longer than required.
There are three areas to consider:
- Tagging – cost ownership can be tracked if cloud systems are tagged accordingly.
- Metrics vs logs – metric data is easier to analyse and is also cheaper when it comes to ingestion and storage costs.
- Granularity / data retention – system data requires one minute granularity, while cost data needs one hourly granularity.
Understanding
This principle ties back to awareness because not only do teams need to be aware of their cloud costs and how they fit into the wider IT and cloud budget, but they also need to understand this in the context of their services. They need to be able to quantify, measure and articulate the business drivers of capacity utilisation, cost, and performance when it comes to their services.
It is all about understanding whether the usage is valid – especially for organisations where data-driven workloads are becoming more important. In some organisations this might be easier than for others; for example, there is a clear relationship between business demand and system capacity in ecommerce businesses. However, data-driven systems may experience cost increases at a rate higher than revenue growth or demand signal, which is why team understanding of the trio of utilisation, cost and performance is vital in cost optimisation.
Confidence
Your teams need to have confidence in their ability to make changes to your cloud infrastructure quickly and without a negative impact on service reliability. This area is often one of the biggest in terms of cost optimisation opportunities. Lack of confidence in or reliability of the service equals excess capacity. Teams should know exactly which resources their service needs and what the “safe” limits are for them.
With this increased confidence, the need for excess capacity is eliminated. In addition, there is an added benefit of being able to deliver quality code or products into production more quickly.
Product
Your teams need to have a strong understanding of the value your product delivers to users and / or your business, and what is needed (in terms of level of performance) to realise that value. When looking at this value, there are two elements you need to consider; performance and obsolescence / relevance. Both areas can lead to rising costs and risk.
Performance – your business and engineering teams must be aligned on the level of availability and performance that your customers should get, relative to market conditions.
Obsolescence / relevance – again, your business and engineering teams must be on the same page on the value that the effects or attributes of your product currently delivers.
Predictability
The final principle deals with the accurate predictability of cloud spend over the three years – based on business demand. Importantly, your teams need this long-term view on cost and service performance, as well as understanding what is driving the costs. In addition, understanding what value is delivered to the business by these costs allows for better conversations between the business and engineering teams on the value of cloud services.
Knowing future cloud spend also plays a major role when negotiating minimum commitments with the likes of AWS (EDP), Azure (Enterprise Agreement), and GCP. But work off your own calculations as cloud providers assume a simple, straight-line forecast which does not take your unique business requirements into account.
What comes next?
There is, of course, more to cloud cost optimisation and building a cost-efficient organisation. But this is the perfect place to start. These seven principles are how we structure our engagements and help our customers realise the value of their cloud estates, while saving money without compromising on performance, and continuing to make gains in the long-term. We help our customers create and embed an efficiency mindset so that the benefits of working with us continue to deliver long after the project is complete.
To understand how we can help your business become a cost-efficient organisation and realise the true opportunity of cloud, get in touch with us today.
Find out more about cloud cost optimisation and building a cost-efficient organisation, download our whitepaper.
