Cloud, CapEx, and OpEx: Reframing the Conversation

January 14, 2021 Phil Le-Brun

The function of economic forecasting is to make astrology look respectable.

—John Kenneth Galbraith


While financially literate, I would not be presumptuous enough to give financial advice. That said, I do know dated beliefs when I see them. One is the belief held by some companies that switching from capital (CapEx) to operating expenditure (OpEx) is the fundamental barrier to cloud adoption. When there is a clear path to significant savings, reduced risk, and improved agility, it’s frustrating to be told that CapEx is the “wrong sort of money to save.” As one CIO commented, “We need to change the conversation: we are comparing CapEx apples to OpEx oranges.”

The CIO was right. If we were talking about converting $1 of CapEx into $1 of OpEx, I might understand the concern. Companies focused solely on the short term gravitate towards capital expenditure to boost perceived profitability, given that popular measures such as EBITDA conveniently ignore depreciation. We can be lulled into believing that CapEx is an almost-free financing option. I’ve been there, spending CapEx while someone else had to worry about the depreciation. Taken too far, this myopia can rob organisations of deeper financial control, growth opportunities, and even their future.

So, how can we counter some of these financial concerns when it comes to the cloud? I share six myths here. I hope they will help you reframe this financial perspective on the cloud, and to think deeply about the balance between short-term profitability and sustainable prosperity.

Myth: CapEx gives me more predictability in budgeting

Isn’t it wonderful that you can predictably know that the same amount of money can be wasted every year! We can casually wave through annual budget requests for core systems such as the ERP system as long as the requested money doesn’t deviate too much from previous years.

We easily fall into the trap of believing that these systems should cost the same every year due to their criticality and need for stability. We further attempt to create predictability by sizing the hosting environment to accommodate peak demand. A system in demand at the end of the month idly floats in a sea of excess compute capacity for the remainder of the month as there is often no incentive to improve financial performance.

With typical enterprises spending two-thirds of their technology budget on just keeping the lights on, it’s easy to believe we only really have to worry about the other third. While a little painful, you can create an impetus to focus on continuous cost optimisation by going through a zero-based budgeting exercise, forcing a critical look at spend your organisation has become accustomed to. Migrating many of these systems to the cloud and using the tools available to free up this year-on-year financial bleeding can save both CapEx and OpEx.

Myth: I have more control with CapEx

While we might exercise control of capital through purchase orders (POs), what happens after the PO is approved? How do we track the efficacy of a new software license or data centre post approval, and what’s to be done if the purchase was a mistake? Do the approvers of the PO really understand that they will be asked to sign more POs in the future for upgrades, maintenance, and hardware replacements? This last year has highlighted the challenges of not being able to flex capacity with demand, improve liquidity, and quickly experiment without being encumbered with long term purchase decisions.

Consider too the psychology of CapEx, namely the much-studied sunk cost problem. If I buy a data centre or database license, or I spend a bunch of capital on a project, what happens if my business needs or environment change, or the initiative starts to go sideways? History shows that we tend to put aside the cold logic and double down on investing more in the hope of rescuing and justifying our original purchase. We find reasons to use the database license whether it’s the right software or not. We innovate based on what we have previously acquired rather than what makes the most sense.

The visibility of cloud expenditure and level of control is far superior to that achievable through capital-intensive equivalents. Agile, cloud-based organisations pace their investments with value realisation. Their iterative investments are managed in near real time, with failures quickly addressed, pre-empting further unwise investments. Now couple this with higher productivity in agile organisations, increased investments which are meaningful to customers and the bottom line, and a reduction in the invest-now-hope-for-a-return-in-a-year mentality. You can see why these agile organisations have a magnitude higher workloads in the cloud.

Myth: OpEx impacts EBITDA, CapEx won’t

Arguably this myth is just a restatement of accounting practices, but it belies the true cost of CapEx. While dispelling the last two myths might reassure you that comparing technology CapEx and OpEx expenditure is apples and oranges, the move to the cloud does still increase OpEx. What was CapEx now becomes OpEx. But wait, is this true? What does system unreliability cost you today in profit and reputation? How much manual effort from your (expensed) employees and third parties goes into remediating or preventing these problems annually? How much (expensed) undifferentiated work are your teams and suppliers engaged in from purchasing licenses, installing and patching software, monitoring data centres and contracts, and so forth? How much business do you lose because you can’t scale to meet a peak demand due to a successful promotion? How much is the (expensed) annual software assurance for purchased licenses?

You get the gist. Very few companies with traditional data centres can or do invest in the level of performance or disaster recovery they need for critical systems. For those that do, their OpEx and CapEx increases considerably. The cloud might increase OpEx, but rarely do we account for the positive impact of fewer outages, less undifferentiated work, and the reduced cost of risk mitigation.

Myth: My like-for-like OpEx cloud costs are higher than my CapEx data centre

This might be true if you were to run an EC2 instance twenty-four seven at 40% utilisation with a licensed database and software assurance just like you might in your data centre. Why would you do this, though? This myth is the equivalent of comparing your personal car to a rental and hence deciding to buy a new car at every holiday destination. These comparisons are common but misguided. Cloud pay-as-you-go elasticity, more efficient data centres, and automation will, in most situations, prove to be more frugal. Some level of rearchitecting can further improve security and reliability at a reduced cost.

Myth: A “private cloud” gives me a good CapEx compromise

An equally vexatious myth is that enterprises can create a best of breed solution, building fixed infrastructure rebranded as a “private cloud,” and using virtualisation to emulate cloud scalability. This so misses the point of the cloud that it’s difficult to know where to start a rebuttal. Yes, this approach can help utilise more of the compute capacity, but it nowhere near reflects the level of scalability, security, or reliability investments made in the AWS cloud. Nor does it come close to providing the over 200 available services ranging from storage and compute to machine learning and blockchain. At best, it is a slightly more efficient on-premises data centre. At worst, it imbues a level of complacency that you are somehow competitive and agile, while haemorrhaging OpEx on patching, monitoring, and managing your pseudo-cloud.

Myth: Signing big CapEx deals is what leadership is about

Well, this is the equivalent of me buying a tractor because (1) I have a lot of earth to move from my garden one day, (2) you never know, I might need the tractor again, and (3) it’s on sale, and good leaders buy things at a discount. I know, it makes no sense, so let’s not even expend time on explaining this fallacy and why limiting future options is not at all leadership-like.

At the heart of the CapEx/OpEx conundrum is the same philosophy that we see with digital transformations. Traditional CapEx investments are based on a presumption that a business can predict the right answer for the majority of investments. Agile companies invest based on an assumption that they can often be wrong and yet are able to learn and pivot cost effectively and quickly.

This isn’t an exhaustive list, and I’d love to hear your thoughts on what other myths we need to dispel.


This blog is provided for informational purposes only, and is not intended to provide financial or accounting advice.

Previous Article
The CIO-CFO Conversation: Capturing the State of IT
The CIO-CFO Conversation: Capturing the State of IT

It’s important for CIOs to be able to convey the current state of IT, particularly in conversations with th...

Next Article
3 Keys to Innovation in Insurance
3 Keys to Innovation in Insurance

There’s a path to digital transformation for companies in every industry, including those that have strong ...