Rethinking CAPEX and OPEX in a cloud-centric world
Odd as it may seem to startup entrepreneurs these days, there was a time that if you wanted to start a Web-based company, you had to think about physical facilities, network pipes, and buying your own hardware.
Today, of course, none of that is necessary. You just log into your Amazon account, spin up one of Amazon's many Web services, and you're off to the races. It doesn't cost hundreds of thousands of dollars (or even thousands of dollars). You can launch a Web-based business for about the cost of a cup of coffee.
When I started ZATZ back in the 1990s, I didn't have that option. The -aaS market (as-a-Service) didn't exist. You couldn't just buy infrastructure-as-a-service by billing a credit card. Today, I have an account on Digital Ocean, where I can spin up server images left and right, and never have to worry about bandwidth. When I opened my account with them, they gave me a ten buck credit, and I've yet to use that up.
But back in the early ZATZ days, we needed physical servers. We needed racks. We needed, first, dedicated ISDN lines and then dedicated T-1 lines. Before we could serve our first page, we had to incur thousands upon thousands of dollars of capital expense.
Let's talk about that for a minute. Capital expense (or CAPEX, for the cool kids) is an accounting term for money you spend on physical resources that needs to be amortized over a bunch of years.
For example, let's say you build a physical warehouse building. It's expected that you'll use that warehouse for many years. Accounting rules say that you can't just deduct the full cost of that warehouse in the first year -- you get to deduct a little bit each and every year. That's called amortizing.
This stuff means very different things to bootstrap startup entrepreneurs and big company financial managers. Back when I started ZATZ, capital expenditure meant "where the heck am I gonna come up with all that money?" By contrast, in a big company, a capital expenditure impacts budget, ROI, and potentially stock price.
Here's a slide I often use to show the chunkiness of capital expenditure:
As you can see, back in the days before Infrastructure-as-a-Service, building out was a chunky proposition. When your server ran out of capacity, you needed to pony up the cash for a new server. When your data center ran out of capacity, it was time to break out the big money and build a new data center.
That was CAPEX. It was hell on cash flow, and it often needed divine intervention from a board of directors or investors, but it also provided some benefits: deductions could be taken over a bunch of years. I'll get back to that in a minute.
The other side of CAPEX is OPEX, operational expenses. Unlike capital expenses, which have to be amortized and deducted over a bunch of years, OPEX is "expensed" right in that year. It's like that cup of coffee. You buy it, you drink it, it's gone. You're not expected to derive value from it for a decade.
Today's world of as-a-Service services has made it possible to turn IT operations into an operational expense, rather than a capital expense. If I were starting ZATZ today, rather than buying those servers, those racks, and that network infrastructure, I'd just spin up a VM on AWS. I'd be billed for whatever usage I used, and that would be that. If I had a big month, I'd get a bigger bill. If I had a slow month, I'd get a smaller bill. All would be OPEX.
By the way, I back up a lot of my personal files to AWS. Do you want to know how much I get billed each month for my entire backup load? $0.18. Yep, it's amazing.
From a startup entrepreneur's point of view, it seems generally better to pay-as-you-go rather than to spend it all at once. But that might not be the case once you start looking into taxes, accounting, and the lord-of-all-that-is-holy, the profit-and-loss statement.
From a CXO perspective, you might not want to account for all your expenses as you incur them. You might want to save some for slower months, or a slower year. On the surface, it may seem that if everything is OPEX, you can just control your own budget without board approval, but keep in mind that it's often a pretty good idea to have the buy-in of your board.
I asked Rob Durst — who advised us early on at ZATZ, has been a CTO, and is now CEO at Silver Bay Technologies — how he'd advise a modern-day CXO to think about CAPEX. Here's how he recommends you think about it:
"There are all the usual accounting, cash flow, and reported earnings issues of OPEX vs. CAPEX, of course. From a practical matter for a startup with a shot at a franchise, however, it would seem to be all about speed (establishing share and brand), and to run fast you often need capital investment beyond what you bootstrap from your own operations. How to best spend that capital, and account for it for both investors and for tax purposes, gets down to OPEX / CAPEX decisions."
That's a perspective I hadn't thought about. Sometimes being able to draw on a big war chest and outspend your competitors can give you a serious competitive advantage. After all, everyone has access to AWS these days, so there's no inherent barrier of entry to choosing that technology as a strategy.
Back when I started ZATZ, our investment in infrastructure was a huge barrier of entry to our competitors. We invested not only in hardware, but in specialized content management software. It allowed the company to run on its own power and support us nicely for almost 15 years. We were far from the only online publisher starting in the 1990s, but whenever we had a competitive sales encounter, we generally had a better and more compelling story to tell than those who didn't seem as strongly capitalized.
The title of this article is "Rethinking CAPEX and OPEX in a cloud-centric world" and it's important to realize that how you think (or rethink) about it depends very much on where you are in the business food chain:
If you're a bootstrap entrepreneur, going all OPEX will get you off the ground quickly as a bootstrap -- but if you start making money, expenses you incurred when you were broke won't be able to be deducted when you desperately need those deductions.
If you're a funded entrepreneur, CAPEX may be valuable because it may be able to give you more resources and competitive differentiation. Also, depending on how your investment is structured, some of those costs may provide tax benefits to your investors.
If you're a CXO, you should avoid the temptation to go all OPEX because it gives you the freedom to just use your own budget without oversight. Oversight and buy-in is important, and you may be giving up some strategic benefits by outsourcing all your data operations.
And if you're a big company, keep in mind that if you OPEX everything and outsource your facilities to companies like Amazon, they're the ones growing infrastructure and you're ultimately relying on a company that may well become your competitor.
The bottom line is this: while CAPEX and OPEX seem like boring accounting issues, how you play them may determine just how strong, competitively differentiated, and secure your business becomes over the long haul.
By the way, I'm doing more updates on Twitter and Facebook than ever before. Be sure to follow me on Twitter at @DavidGewirtz and on Facebook at Facebook.com/DavidGewirtz.