The Paradox Of Positive Elasticity

With so many vendors claiming new efficiencies on their various technologies, will we actually end up spending less money on these technologies, or perhaps end up spending more?

In the spirit of eyes-wide-open IT discussions, we have to be prepared for the likely scenario that cost reductions in IT may -- paradoxically -- result in organizations spending more on IT, and not less.

And, in most cases, we'll be better off as a result.

Elasticity Made Simple

If you've ever taken a tour through an economics course, you probably remember the discussion around price elasticity.

The thinking is straightforward: lower the cost of a good, and something will happen.  If less money in total is spent on your widget, it can be said to have negative elasticity.  If more money is spent (in aggregate), the good or service is said to have positive elasticity.

I can make a reasonable argument that most IT inputs are strongly positive elastic -- as we make the product or service far cheaper, we end up consuming far more -- and spending more as a result.

Examples Are Everywhere

By any measure, the cost-per-unit-of-compute metric has been falling like a rock.  Consider a modern Xeon 5500 processor compared earlier Pentiums of just a few years ago.  Hardly a fair comparison, right?

How many organizations are spending *more* on compute these days (directly or indirectly), rather than less?  Just about everyone I talk to is spending more on compute -- in aggregate -- than they were just a few years ago. 

The reason, I think, is simple: as compute prices fall, smart people figure out new ways to exploit the new economics to make money -- usually at a faster rate than the price declines!

Or, let's take network bandwidth -- another service whose historical price curve looks like a Grand Slalom skiing run.  Paradoxically, most IT leaders tell me they're spending more on network bandwidth than ever before.

Put up a bigger, faster and cheaper pipe, and all sorts of clever applications will come out of the woodwork to use up your savings, and then some.

Another example: virtualization allows you to run more applications on less hardware, among other things.  But, paradoxically, many organizations are reporting they're running far more than ever before.

Finally, consider my personal favorite topic -- storage. 

Media prices have been in a free-fall for decades. Layer in all the newer space-saving technologies, and you'd think that the storage industry would be in dire straights from a revenue perspective.

Quite the opposite: short term economic fluctuations aside, the underlying trend is that storage industry revenues continue to grow, and grow and grow.

Somehow, IT organizations figure out how to use all those storage savings, and then some.

Simply put, it's hard to argue that -- long term -- aggregate IT spending is increasing, not decreasing, despite order-of-magnitude declines in the cost-per-unit of the underlying technologies.

IT Expenditure Growth: Good or Bad?

I'm probably going to get flamed for this, but I'll argue that -- in many circumstances -- increases in IT expenditures can be a good thing. 

No, I'm not nuts, and I'm not pimping for the IT industry as a whole.

I know, the optics aren't good.  Spending an ever-increasing amount on IT resources may create the appearance that something is amiss in IT land.

But we need draw the bigger picture. 

IT exists to support the needs of the organization.  And if the organization can find a way to make more money from decreasing IT unit costs (and end up spending more on IT as a result) that can be a good thing in my book.

All that new server capacity isn't sitting idle: it's helping to run the business.  All those newer virtualized applications are doing useful work.  All that extra storage capacity isn't sitting their empty -- it's filled up with data that the business thinks is important (hopefully!).  Ditto for network bandwidth.

Make something important better / faster / cheaper, and the business is likely to use more of it -- with good results.

Sure, there's inefficient and wasteful consumption we'd all like to root out, but that's always been the case with business use of IT resources, hasn't it?

One quick example, if you don't mind. 

Recently, EMC installed a bunch of Cisco telepresence units to connect our various operations.  I overheard one conversation where someone was bemoaning the fact we were now spending a ton on network bandwidth, and "something should be done".

Well, those telepresence units are replacing lots of long-distance business trips.  We're now collaborating in a way we never could do so before, and good stuff is happening as a result -- I can see the business value.

Sure, there's the occasional wasteful use.  But that doesn't mean that -- on the whole -- we're far better off as a business spending all that money on new network bandwidth.

The Consumption Controversy

One of the more controversial tenets of most cloud models is convenient consumption: allowing users to consume what they need, when they need without a lot of fuss.

The idea that users would be permitting (largely) unfettered access to computing resources is downright scary to many IT organizations.  Just think how much IT they'd all consume if it was faster, better and cheaper!!

But, being on the other side of the table, let me offer a different perspective.

In the company I work for, I pay for what resources I use: facilities, phone, travel, contractors, etc.  Show me the costs, and I can usually make a good decision around tradeoffs in my budget.

I don't have to go through an elaborate process to get a meeting room.  Nor make a phone call.  Nor take a business trip, or hire a contractor.  It's all pretty straightforward stuff.

I am, however, expected to stay within my budget unless there's a damn good business reason.

For example, if our HR group found a great source of contractors who were good / fast / cheap, I could end up spending more on that category.  Would that be a bad thing?

Why shouldn't computing resources be any different? 

Make IT resources cheaper for me -- and easier to consume! -- and I may choose to allocate more of my budget to that particular bucket, vs. some other one.

This may mean that I want to spend more money on IT, rather than less.  It's a business decision. Congratulations, I've decided to hand more of my budget over to you in the IT department.  You win.

Stepping back, I would look to the providers of these internal corporate services (facilities, phone, travel, IT, etc.) to offer me continually improving good / fast / cheap for whatever internal service they're providing.

And I may end up consuming more of whatever as a result.

Where Does That Leave Us?

I despair somewhat when I see those of us on the vendor side tout the latest money-saving technology.  And we're all guilty of it.

Sure, the cost-per-unit of whatever product or service delivered goes way down, but our customers usually end up consuming far more of it as a result.

I think we're creating the expectation that -- yes -- by using our latest-greatest magic tech, your IT bills may actually go down.

When, in reality, the evidence shows that they'll likely move in the opposite direction.

I just can't imagine IT vendors putting a disclaimer on their data sheets: "Warning: Use Of This Money-Saving Technology May Lead To Higher Aggregate IT Expenditures".

That'd be asking too much, wouldn't it?

:-)

Comments

[Nick Mehta]

Hi Chuck,

Great post and something I think about a lot. I believe and hope this is true - that technology commoditization actually yields increase in technology use and net increase in purchasing. However, two things concern me:

1. While price drops may increase overall unit sales and revenue, how does it affect margins? For example, a web-based version of Microsoft Office or Exchange on a per user per month basis may yield MORE Microsoft users and MORE revenue but LESS operating profit and cash flow for Microsoft.

2. There are a number of economists who theorize that technology and the Internet in particular will actually lead to a massive wave of deflation in the coming years (theory being technology increases market transparency, commoditizes products and services more quickly, etc.)

It's hard to say which way it will go, but as they say, it sure will be interesting to watch!

Nick

[shiningarts]

I have observed human nature for quite some time, although I didn’t formally study economics. Everybody gets accustomed to their current situation or lifestyle fairly quickly. I, for one, attest that my current lifestyle in the United States is rather different than my early life while I was growing up in another country. Regardless of a person’s economic condition, I believe very few feel completely satisfied with their own financial circumstances. This is normal condition of humanity because we are not accustomed to comparing our current stage with our prior lesser condition. You feel always inadequate because you compare yourself with something more than you have.

I don’t think the economists and pundits are necessarily declaring what will happen—they are just reading tea leaves, albeit with an experienced eye. What they really can’t predict are psyche of the massive population and specific future usage demands. I believe people will find a way to consume more IT recourses if it helps meet their desires and further their expectations--even if this will lead to much more automation and less human intervention. However, we can’t predict where this eventuality of this mass atomization will lead because of the falling resources being dedicated to IT expenditure.

All of things we are talking about now including virtualization, private computing, or whatever will eventually behind us and we will continuously pursue greater and greener pastures. It would very interesting if some scientist claims that they have discovered the “Paradox of Positive Elasticity” deep down of our chromosome someday. Well, the wisdom of the wise and the intelligence of the intelligent will be never sufficient for us all.