Mixed Results For IT Spending Forecasts
July 17, 2017 Timothy Prickett Morgan
We have made it through the first half of 2017, and now is a good time to take a pause and count the money that the IT vendors of the world have amassed from the hardware, software, and services they provide to the datacenters of the world. And so, IDC and Gartner have done just that even ahead of the reporting of financial results for the second quarter. The results from these two firms generally agree, and they show that, as always, different parts of the business are growing even as other parts slow or even shrink.
One big reason why these IT spending projections matter is that they provide a backdrop sentiment against the economy at large and within different technology areas, industries, and global regions. IT managers and the bean counters that are sometimes their adversaries use such metrics to steer a lot of their own spending. If people have money to spend, and their peers and rivals are spending, then it is easier to argue to spend, as was the case during the dot-com boom; when some companies cut back after the irrational exuberance ran its course and the dot-com bust hit, then everybody tended to go down – with the exception of the hyperscalers who were just getting started back then and a few companies that had no choice. The Great Recession caused a boom-bust-boom cycle, too, with companies spending like crazy as the recession began, then pulling back hard as it got really bad, and then loosening the purse strings two years later as a recovery was beginning.
These days, we are in an era of seeming prosperity, although no one would call this a boom. It is more like a slow grind, with people being hopeful that housing prices and Wall Street continue to grow. So IT spending growth is slowing, and not the least of which because China’s economy and its voracious appetite for everything is abating a bit at the same time that it is getting more interesting in buying local – and particularly not buying IT gear from vendors headquartered in the United States.
In constant currency terms that take out the effect of changing local currencies against each other and gauged in U.S. dollars, IDC believes that IT spending worldwide will rise by a pretty respectable 4.5 percent, to $2.1 trillion, and it is projected to grow at another 4 percent in 2018 according to the company’s prognosticators. Telecommunication services are under severe pricing pressure and represent nearly half of the total pie of Information Communication and Technology (ICT) spending, which is expected to grow at only 2 percent in 2017 to reach $3.5 trillion.
IDC reckons that spending on servers, which we consider both a leading indicator when things are going well and a lagging indicator when they have gotten bad but we are just not sure yet, will increase by 4 percent this year and by 5 percent in 2018. Product refresh cycles for processors, memory, flash and other persistent storage, and networking are all driving the hardware spending rise that will start right about now, according to IDC. And infrastructure as a service, which means hardware for rent, will hit $25 billion this year and crest above $50 billion by 2020. This is a sizeable market, and it implies that hyperscalers and cloud builders will represent a large portion of expected server sales by the end of the decade. (Probably still not a majority, though.) Software spending is much larger than hardware spending, as you might expect, and is expected to increase across all types by 7 percent to hit $471 billion in IDC’s models. IDC said that IT services (not including those telecom services mentioned above) would see 3 percent growth in 2017, but the company did not give out the figure publicly.
In releasing its own 2017 IT spending projections, the analysts at Gartner absolutely did give out their figures. Take a look at this table:
As you can see, Gartner is less bullish on spending on datacenter hardware, which includes servers, storage, and switching, with an anticipated growth of three-tenths of a point in 2017. That is better for sure than the three-tenths of a percent decline that Gartner saw for datacenter iron last year, and the growth rate is expected to quadruple in 2018. But still, the revenue stream will be stuck right around $170 billion. Enterprise software revenues, on the other hand, are expected by Gartner to show stellar growth, opening up the gap between datacenter hardware and the software that runs on it even further. As for devices sales, the market for PCs, smartphones, and tablets – in the aggregate – is looking to be stronger than expected, with a 3.8 percent bump up to $654 billion in 2017, which is better than the 1.7 percent growth in Gartner’s earlier forecast. Tablets are in decline as phones get richer and smarter.
The upshot for Gartner is that it expects for IT spending on a global basis to rise by 2.4 percent to $3.5 trillion (including the telecom services), which is up from its previous 1.4 percent growth forecast mostly due to the decline in the US dollar’s exchange rate against other currencies. We suspect that a lot of the growth that IDC is showing in its forecast is from the same cause.
This is not real growth, of course. But the money spends like more when you bring it back to the States, so there is that.
One last thing: Whenever we talk about IT spending and its growth or decline, the one bit that is always missing from these projections is the cost of people. You can get relative salary data, but I have never, in three decades of writing about the IT sector, seen a forecast talk about the spending on IT staff, which is a way of talking about the cost of in-house application programming and the services provided internally to maintain systems and the software stack. I have no idea how much this costs, but let’s take a stab at it on the back of the old envelope. There are about 10 million Java programmers in the world, and this is by far the biggest single pool of programmers. Let’s assume there are another 20 million programmers using PHP, C/C++, C#, RPG, COBOL, Fortran, Ruby, and whatever. Let’s also say there is one system admin for every five programmers, or around 6 million of them worldwide, and another 1 million CIOs, system architects and database managers, and such. That is 37.2 million people. If the average salary across these people is $65,000, that works out to $2.4 trillion, and if you add in the benefits overhead of maybe 50 percent, then you are talking something like $3.6 trillion.
In other words, the actual IT sector might be at least twice as big as these forecasts show. It is a wonder no one has figured out how to present the full picture.
Finally, the fundamental issue and problem in all IT, and in all corporate computing, is revealed by Timothy Prickett Morgan.
Thank you Timothy Prickett Morgan
————————————————————————————-
One last thing: Whenever we talk about IT spending and its growth or decline, the one bit that is always missing from these projections is the cost of people. You can get relative salary data, but I have never, in three decades of writing about the IT sector, seen a forecast talk about the spending on IT staff, which is a way of talking about the cost of in-house application programming and the services provided internally to maintain systems and the software stack. I have no idea how much this costs, but let’s take a stab at it on the back of the old envelope. There are about 10 million Java programmers in the world, and this is by far the biggest single pool of programmers. Let’s assume there are another 20 million programmers using PHP, C/C++, C#, RPG, COBOL, Fortran, Ruby, and whatever. Let’s also say there is one system admin for every five programmers, or around 6 million of them worldwide, and another 1 million CIOs, system architects and database managers, and such. That is 37.2 million people. If the average salary across these people is $65,000, that works out to $2.4 trillion, and if you add in the benefits overhead of maybe 50 percent, then you are talking something like $3.6 trillion.
In other words, the actual IT sector might be at least twice as big as these forecasts show. It is a wonder no one has figured out how to present the full picture.
—————————————————————————————–
Corporations worldwide have indeed figured out this problem and stupendous annual cost by dumping the cost of people, particularly in the IBM i ecosystem, which is why there are perhaps one quarter of the IBM i customers today compared to a decade ago.
The clear (to me) ultimate solution is NOT PHP or Ruby or Java, or any other technical solution, but to enable corporate developers/programmers, and corporate management, to be perhaps Ten Times more capable and productive and valuable (and happy) than now with the ultimate productivity tool; the computer and machine learning.
Timothy Prickett Margan has finally focused on the key issue and problem, but not on the obvious and immediately available ultimate solution.