Tech Your Facts: Is Cloud More Expensive Than On-Prem?

Like any field, tech has its share of myths (and we’re not talking about the Sudo Dragon, which commands lesser mortals to make them a sandwich). In Tech Your Facts, we examine these common tech beliefs to see if they are true or not. This week, Mattias Andersson takes the helm.

There is a lot of confusion about how much cloud computing costs. It’s not uncommon to hear some people proclaim the cloud is more expensive than on-prem, while others say it’s cheaper. 

In reality, both are myths. 

“Ah,” you might say, “Does that mean cloud costs exactly the same as on prem?” 

Not at all! The reason both claims are a myth is because the very question is an oversimplification. It doesn’t take into account the true value proposition of the cloud, and what happens when people start using it.

Back in the 19th century, a fellow by the name of James Watt invented the Watt steam engine. It was a big improvement on the steam engine of the time, which used much more coal to get the same amount of energy. 

So, did coal usage in England decrease with Watt’s new invention? No, it soared dramatically!

Why? The reason was that because Watt’s invention made coal more cost-effective, people could use the engine in a wider range of industries, to solve all sorts of problems. So people used it more — in fact, it helped “fuel” the industrial revolution. 

Now, I’m not explaining this to just give you some interesting historical tidbits. This rebound is known as the Jevons Paradox — and it doesn’t apply only to coal, but any resource, including the cloud!

When an organization adopts cloud services, staff are empowered with a tool that is cheap and can solve a wide range of problems. Once they realize how valuable it is, they start using it everywhere — which in turn means they may well spend more money on cloud than they were on-prem. To paraphrase the old Pringles slogan, “Once they pop, they can’t stop.” 

At this point, you might conclude that the cost of cloud is both cheaper and more expensive: cheaper with the actual cost of the service, but more expensive because of how often a business winds up using it. Again, this doesn’t quite take everything into account either. 

First, imagine two cars, both of which run on gasoline. One of the cars uses half as much gasoline as the other one. So, which would you choose? Purely by cost, you might pick the one that uses less gas.

But what if I told you the car that uses less gas had no steering wheel, and was pointed in the wrong direction?

Suddenly, we’re having a different conversation. It doesn’t matter much if you’re operating more cheaply if you’re not getting to where you want to go, right?

You can’t really compare a steerable vehicle with a non-steerable one: it’s comparing apples and oranges. It’s actually about reaching your destination. 

Businesses often measure things in Return On Investment (ROI). If you are able to do something that generates ten times as much value with two times as much cost, this is a pretty useful thing. 

When people talk about the number one value proposition when moving to the cloud, it’s not cost, it’s agility. Agility means you can do the right thing instead of doing the wrong thing more cheaply, and that’s everything when trying to run an effective business. 

After all, the cheapest solution is to have no tech at all, but your business wouldn’t be very agile!

Another thing that can happen is that you wind up trying to do the old thing in a more expensive way in the cloud, replicating that car without a steering wheel. The underlying technology is more efficient, but you’re using it less efficiently.

As a result, a business that adopts cloud technology can end up costing themselves more to get the same or less value. This can lead people to conclude “everything is terrible since we moved to the cloud”, because they haven’t realized the full value proposition.

It's kind of like hearing that planes are a much faster mode of transportation, so you buy one for your company. But then you get even more frustrated than before when you try to use it because you're still trying to drive on the streets to your destination, instead of flying like you should. That truly is worse, but only because you're using the tool with the old mindset and approach. And your bad experience is not a fair indictment of planes, in general.

Often when someone claims that “cloud is more expensive than on premises”, they’re not comparing the comparatively mediocre systems they build on their own with the robust systems they can get with the cloud. In almost all cases, this makes it a bad faith argument.

When someone makes this argument, they are often not considering things like the cost it would take to replicate the level of physical security offered by having your information in a secure data center, or the effort it would take to replicate the low latency of a content display network (CDN), or any number of other things that are included "for free" in the cloud pricing.

Are you thinking of the cost that it would take to have a data center in each country so you can ensure quick delivery of a digital product to people no matter where they are? The difficulty of getting the hardware to a location, setting it up, sorting connections and electricity, or managing the political situations to set up these centers? This often doesn’t come into the equation, these “uncounted” benefits.

Another example is how when AWS, Microsoft, or Google manage a service for your database, they’re not mismanaging it, as so many of us often are. They’re doing it professionally, they’re doing it at scale — learning lessons broadly to avoid problems that we haven’t even considered…or at least, not until we hit them ourselves. 

When we do hit issues, on-prem, it’s often not a “here’s five dollars and the problem goes away” sort of situation. Rather, it is often an issue that means you can no longer run your business the way you thought you could. Everything grinds to a costly halt until it's brought back online. But with cloud, risk mitigation is built in — sometimes as an automatic thing baked into the managed service (like hard drive redundancy in bucket storage), or sometimes as an option you can take advantage of (like multiple availability zones). This is more than just a swift delivery of resolution — it’s offering continuity of business.

Yes, you may pay more money for it — because you’re getting more value to solve a problem that you didn’t even have the option to solve before. Practically speaking, you couldn’t have another data center (or three) just sitting around in another location waiting for a failure not to happen. So people glibly ignore these sorts of things when it comes to risks and costs, and so they oversimplify the dollar figure.

Conclusion: Cloud is not cheaper or more expensive, but different

When we simplify things, we can remove important context — in this case, the value proposition of what cloud computing offers, which simply can’t be provided by on-prem computing. So the next time someone asks you if cloud computing is more or less expensive than the alternative, remember that the better measurement is not cost, but agility to achieve practical solutions.