As a term, cloud shock has been with us a few years now, but it’ use is not yet quite yet sufficiently widespread to safely assume that everyone knows what you’re talking about when you say it. Cloud shock is, in very simple terms, the moment of realization that money being spent on cloud services (be that software and/or infrastructure) is significantly higher than anticipated/budgeted and/or cannot be explained and attributed. It can be a rude awakening.
Cloud shock can be experienced by a wide variety of stakeholders across the organization from individual business managers (a recent report from Blissfully suggested that in companies with 1,000+ employees, there can be as many as 100 SaaS billing owners), through to functional leads (Chief Marketing Officers now spend more on technology each year than they do payroll) and even the CIO.
In fact, the CIO is one of the most likely stakeholders to experience cloud shock, especially as more of the organzation’s technology spend starts beyond their visibility and control.
But by all accounts, the worst-case scenario is when the CFO has their cloud shock moment. It’s often the occasion when an individual invoice or aggregated spend with a single vendor reaches a certain critical mass that it hits their radar. And in large organizations, that typically means it’s a big spend (with many reports of organizations spending a whole year’s worth of SaaS or IaaS budget in a single quarter, it’s not difficult to imagine how that can happen).
Let’s say it’s an Azure or AWS quarterly invoice and it’s for a million dollars (say it in your best Dr Evil voice). It’s the kind of bill that ends up with the CFO for approval. And if your CFO is anything like mine, they’ll ask questions. Difficult uestions like “why is it so high?”, “who’s responsible for the spend?”, “what did we use this for?”, “can it be attributed to revenue?” or “did anyone actually check this?”.
The initial cloud shock generated by the sheer scale of the invoice is compounded by the inevitable shrugs of shoulders and blank looks when the CFO asks various staff (not least the CIO) to answer the questions above. Shock quickly turns to frustration, in no small part because your Azure or AWS bill won’t answer those questions for you. A single invoice could represent hundreds of people across the organization all doing different things, using different budgets, working on different projects and generating value (or not) in different ways.
But your Azure or AWS bill – or Salesforce or ServiceNow or whatever – won’t tell you who, what, where and why. Just what’s owed for the aggregated consumption.
How to avoid cloud shock
There are many ways to try to avoid cloud shock (or, more likely, to try to avoid it happening again), some of which require fundamental changes to the way stakeholders across your organization procure and onboard software and infrastructure. If your organization is willing and able to implement such major technology governance changes, all good.
If not, there are still some initial steps you can take that will have a big impact on your cloud spend.
Start discovering and identifying SaaS spend
Estimates vary, but it’s a safe bet that at least 30% of your software spend is now going towards SaaS software. It’s also increasingly likely that these apps are being selected and paid for either at a business unit level or even on credit cards. This can lead to all kinds of inefficiencies, from duplicated and redundant subscriptions to failing to realize volume price breaks.
There are three main ways to discover SaaS consumption: Hook into finance systems, discover it at the point of use or undertake a manual survey.
There is a growing collection of ‘SaaS management’ tools that claim to integrate with finance and expense systems to identify SaaS spend. There’s some promise to this methodology, but it’s not as easy as you might think to integrate with those data sources or tell which lines on a credit card claim or invoice are for software versus sandwiches!
For SaaS discovery at the point of use, you need a specialist inventory solution with a browser or gateway plug-in. Potentially painful to set up, but the theory is sound. The key limitation here is that those plug-ins can’t (and shouldn’t) track all internet use, so they need to be told exactly what to monitor. And if you don’t already know what you want to monitor, you’re potentially in a Catch-22 situation.
Finally, there’s a manual survey. It might sound antiquated, but actually asking key stakeholders what SaaS apps they’re using can yield some interesting results (in a previous life, I was dealing with a SAM team that thought there were five key SaaS apps in use across the organization. We did a manual survey and discovered there were 78!). It might at least give you a sense of scale for actual SaaS use across the organization.
Don’t simply migrate like-for-like to the cloud
One of the great myths about cloud is that it is, without exception, cheaper. Not only is that not true, it’s especially untrue if you just replicate your on-premise licensing in the cloud. For many applications, the licensing rules not only change in the cloud, but change depending on which cloud host you use.
Microsoft SQL is a great example. Not only are the licensing metrics different in the cloud compared to on-premises, there are different rules in play depending on whether you host your cloud-based SQL in Azure or AWS or Google. It’s very easy to get it wrong and over-spend.
Make sure you’re on top of IaaS right-sizing
Undoubtedly, a proportion of cloud shock can be attributed to cloud sprawl, specifically cloud instances that are spun-up, used for a time and then left languishing. Unused but still causing a spend because they haven’t been deactivated or retired.
But there’s another, perhaps even larger overspend risk. All too often, cloud instances are over-powered. In other words, the computing power (CPU, memory, disk space) is disproportionate to what an application, database or workload really needs.
It’s important to understand where CPU, memory or disk utilization on cloud instances is low (you’ll have to decide how low is low). On rare occasions, you might have a legitimate reason to over-power an instance, but mostly it’s just money down the drain.
Head off cloud shock before it hits the CFO
If you’re lucky enough that your CFO hasn’t yet had their cloud shock moment, there’s a small window in which you could implement one or more of the strategies outlined above. And if it has, perhaps the need is even more pressing.
Regardless, cloud shock isn’t inevitable and isn’t something you can’t do anything about. With relatively simple steps, you can take positive actions to reduce your organization’s cloud spend.
Whether you’re looking to move the cloud in a cost-effective manner or ensure that what you have deployed isn’t creating over-spend, speak to a cloud licensing expert at Livingstone today.