I have a few thoughts to share with you ahead of our Symposium season that continues in Orlando next week. If you like the idea below, or if you disagree with them, let’s touch base in a 1-1 or over a coffee and explore this interesting topic. Maybe I am on point. Maybe I am wrong. Either way, it’s fascinating…
We are all excited about cloud computing and what it can do for our businesses. If you read the press you would think we are betting the farm on cloud. The press, the hype, the siren calls from Tech Icons and financial analysts suggests Cloud will change everything. And, even if we keep in mind that all investors enjoy hype because it creates profitable opportunities for them, there is some substance to the Cloud frenzy.
Yet over the years, as with virtually all technology, we overestimate the short term impact and underestimate the long term impact of same. To explain this we first we need to agree what cloud really means. And believe me, we are not all in the same page.
The current focus of cloud computing rests firmly on the evolution of Moore’s Law. If you didn’t know, Moore’s Law sits at the center of how most economists think of IT adding value to our economy. Indeed, IT’s contribution to productivity is directly calculated off the inexorable decline in semi-conductor prices and the growth in processor power. This might seem obvious and perhaps a small point but it talks to the total misunderstanding of what cloud means to us all long term.
To prove this point, I draw your attention to an article, “The Productivity Paradox of Information Technology: Review and Assessment”, by Erik Brynjolfsson, from 1994. Erik and his “productivity paradox” is at the heart of this issue. The opening paragraph says, “The relationship between information technology (IT) and productivity is widely discussed but little understood. Delivered computing-power in the US economy has increased by more than two orders of magnitude since 1970 (figure 1) yet productivity, especially in the service sector, seems to have stagnated (figure 2).” Note how “IT” is equated to “computing power”.
Others have suggested that IT Does Not Matter. Nicolas Carr’s troubling assertion in 2004 is with us today in spades. Cloud computing offers the promise of services at elastic scale: virtually unlimited processing power at a declining cost. This IS information technology, right?
Actually, no. This is only one part of IT and perhaps it may not be the most valuable part of IT. Being able to process data faster – day in day out – is useful. But more important than the processing itself is the “what” the processing is actualy describing. This subtle difference might be understood by economists, political leaders and even financial analysts, but you won’t find the difference explained or explored in their literature. There is an attempt in some productivity data reporting that does try to take account of the quality of processing as well as the price, but this is a recent addition and itself is not the point.
CEOs intuitively know the difference I call out here. They know that it’s not the speed of action that wins if that action itself is poorly judged. Speed will help if your response or action is useful. Today the cloud conversation is fixated on speed and performance.
Back in the day MRP was a business differentiator. Material Requirements Planning was a factory innovation. But the first generation of this innovative business process was oriented around a monthly and weekly batch planning cycle. Soon competitors wanted to ‘do’ MRP too. Soon enough every firm was doing (batch) MRP. Then came weekly and daily MRP with faster business planning cycles. Eventually we even achieved net-change and real-time MRP. The point is that the WHAT was more important than the SPEED. And this is pretty much the case with every business process innovation – think more recently of ERP, Supply Chain, Procurement, CRM and so on.
First there is a product (or service) innovation (the what) and sometime later, as other sources of supply come on-line, the focus shifts for them to how that product or service is delivered to market – this is process innovation (the how). The former denotes the effective work; the latter on how efficiently we can deliver it. IT and all it represents is littered with both forms. See Mastering the Dynamics of Innovation.
Since cloud computing is focused on the use of Moore’s Law and processing speed and power, we should see improvements in productivity for a given task. In other words, a given process (e.g. switching from paper-based mail to processing email) can be made more productive with IT. This is self-evident even if it does not always show up in the data. Back to Mr. Brynjolfsson’s paper. He calls out multiple reports and research that at best identifies (if at all) a weak link between IT spend and improved productivity.
We need also to recognize that we can look at how IT is used by different firms. Firstly there is the IT industry itself. Then there are IT-using industries (generally when an industry spends as much as 5-10% of its costs on “IT” it is an IT-using industry) and non IT-using industries. The IT industry itself has transformed its own productivity – think back to CASE tools, Object Orientation, and code generators. This productivity improvement does how up in the data. Additionally IT-using industries, those that use IT to do their work, can now drive a new generation of productivity growth. But this is where we get the paradox: we don’t see much of it in the data and, given the focus on processing, it seems that what those organizations do with IT doesn’t matter. I think it does.
“ERP on-premise versus ERP in the cloud is still ERP”. Only cheaper. Maybe.
Not really. The initial cost is lower, because you can start with “less” capacity and even features. And maybe you spend a little Opex, rather than a large up-front Capex. But, the features are compensated with human labor (paid wages) and the “less” argument only applies as long as your company never grows—is that actually an admirable goal?
After a period of time – the time over which you used to depreciate your capital outlay for an on-premise app, the cumulative amount of opex now spent on cloud services will add up. It may even add up to as much as the original depreciation cycle for the older capex-based acquisition. And don’t forget – as you consume more and more of your opex on cloud services, the opportunity cost for your excess cash flow will increase and you will have less for other purposes.
But there are also warts with this. Cloud apps as services offer less flexibility in customizing business processes, apps and data models. So for a while yet, all other things being equal, cloud based apps will be generally less functional than on-premise apps- until vendors figure out cost effective ways to serve you unique capability as if it were a unique, private cloud offering but in a multi tenement public cloud manner. This is what PaaS and DaaS (innovation and differentiation for business processes and decisions) offer and where IaaS and SaaS are not meant to be focused (IaaS for standardized services for IT and SaaS for standardized processes for business).
So for the time being cloud will mostly change the economics of ‘doing IT’. It may prove cheaper than on-premise approaches and it adds certainly more scale and more resilience. These are all good things the CIO cares about. And if costs come down this will help the CIO meet their number one stated challenge of cost optimization.
But what of the CEO? She wants business growth. She wants business innovation. She wants business differentiation. Here cloud computing tends to fall short, so far. Yes, there are some emerging cloud offerings and approaches that promise some business innovation. There are new business models popping up too. But there is no assured blue-print yet that stems from cloud computing automatically.
The current focus of SaaS cloud apps is focused on standardized business processes and so these will support a lot of the core business processes you need but SaaS will not yield to the CEO the innovation and differentiation she demands. This is where eventually PaaS and DaaS will become more important. But even this long term view wont ring true until enough SaaS is “out there”.
Niche cloud-born SaaS app vendors are popping up but this is not a panacea either. It is not a panacea since someone has to integrate the stand-alone cloud app to the standardized business process-based SaaS solution you use in a different cloud. This will be more complex and harder to do than what you used to do when you had total control of two on premise apps since you won’t likely have the ability to change either applications’ data model, business rules, or metadata etc. So then cloud integration will again come to the fore – but not the standard technical integration – I mean semantic integration. I suspect we will see “information governance as a service” to delineate from integration as a service.
Cloud-native App Dev platforms (e.g. PaaS and DaaS) are forming up. This is a good start. But the hype and economists do not yet generally recognize the difference in how each layer of this cloud stack offers opportunity to business. They are focused on IT vendors (which is, admittedly, an important part of our economy) and they are not focused on what IT-using businesses will do with cloud:
- Today Cloud computing will drive improved economics for ‘doing IT’
- Soon Cloud computing will change the economics of ‘doing new business innovation and differentiation”.
I think cloud will represent two distinct waves of innovation. We are firmly in the first and we have yet to truly engage in the second. Only trouble is, the latter will likely do more for our overall economy than the first. The former is where all the hype is and where IT productivity improvement resides. The latter is where large-scale business productivity improvement resides.