Worldwide end-user spending on public cloud services is forecast to grow 18.4% in 2021 to total US$304.9 billion. This is up from US$257.5 billion in 2020, according to Gartner, Inc.

Whilst these numbers are encouraging and perhaps overwhelming, this does not give a true sense of “value” spent by companies on cloud. The emphasis is on “value” because it is estimated that only 70% of the money spent on cloud is efficient and the rest wasted. There are various strategies that organizations have initiated to avoid these wastages, including adopting a “Cost first” approach to their cloud journey.

When taking a cost first approach, you become proactive when it comes to cloud spend. It is understandable that when an organization embarks on their cloud journey, adaptability and feasibility of their applications on cloud is the prime consideration rather than cost. However, once that has proved successful, cost must shift left.

To understand the cost first approach, let’s first look at the functionality of the centralized Cloud Governance Team. There are many roles of this governance body, but the main ones include:

1. Choosing one or more public cloud providers.
2. Certifying a set of services that can (and cannot) be used by the teams within the organization.
3. Monitoring cloud spent by each business unit in the organization.
4. Centralizing access management on cloud.
5. Defining standards and best practices for all to use.

 “Monitoring cloud spent by each business unit” has a direct reference to cost, but it’s not necessarily a cost first approach. If there is no cloud budget set out for business units, then it’s a reactive approach. If business units are billed based on only what they consume, there is no way of knowing if they put an emphasis on optimizing cost.

To put a cost first approach at the fore, budgets must be defined. These budgets should be broken down into the following categories:

1. Running Cost – The cost that will be incurred month over month for the resources that are already running on cloud.
2. Onboarding Cost – The cost that is expected to incur on newer cloud applications (migration or green field) and planned to be onboarded during the current year.
3. Exploration/Experimenting Cost – The cost for conducting Proof of Concept’s (PoC’s) with the newer or un-certified services. Businesses need to create PoC’s to show the cloud governance team why they should invest in extending the list of certified cloud services.
4. Training Cost – The cost for upskilling the team. This includes attending summits, webinars, certification reimbursements, eLearning platform subscriptions fees etc.

All four categories listed are related -- without training, there won’t be exploration, without exploration there won’t be any optimization of the running cost and without optimization onboarding, onboarding new applications would be like putting money into a blackhole.

The category that requires the highest budget is the running cost, closely followed by new application onboarding, training and exploration costs. 

Fig 1: Cost first approach

The purpose of putting the entire cost into separate buckets is mainly to introspect at any given point in time. If the running cost is increasing and the exploration and training costs are decreasing to maintain budget, it’s normally an indication the business unit is not investing much effort to optimize the cost.  Ideally, it should be training costs increasing (promote P2P knowledge sharing rather than individuals gaining knowledge externally), exploration costs increasing and running cost either maintained at the same level or decreasing. The “cost first” practice at this point will ensure that you are using the budget efficiently.

One key point to remember is that the “cost first” approach should only come into effect after at least 2 years of cloud adoption by the organization. It takes time to understand and adapt to the cost aspect on cloud. This is primarily because of its complexity as Capital Expenditure (CapEx) vs. Operating Expenditure (OPEX) are good for discussions but in practice, is very different.  Secondly, it changes from one cloud service to another.

Once this approach has been followed for some time, the previous years’ performance should govern the current years’ budget. Onboarding newer applications onto the cloud should be the main factor for the budget increase but if there was no improvement shown for the existing workload on cloud, there will be more wastage. Incentivization for business units working within the limits and questioning the ones that missed it is an approach to take. The essence of a cost-first approach is that, if you work within your budget, you will shift left.