There, I’ve said it. I know it’s a heretical view, but I’m no follower of fashion, whether sartorial, gastronomic, linguistic or professional.
But it’s not a simple issue, and I will qualify my views very carefully.
I’m talking of course, of the IT Cloud, that somewhere-nowhere place where software and data are held for private or corporate use. I don’t doubt that it’s a safe, secure and cost-effective place, (and if it’s not yet those things, it could be). And I don’t doubt that when it comes to certain kinds of ‘utility’ applications – word processors, spreadsheets, simple databases, and so on – it makes it easier for people to share data and ideas. There are plenty of benefits when it comes to certain, limited, kinds of Cloud. My objections are confined to the Cloud as it’s promoted for business applications, a field I’ve worked in for more than 35 years.
Business applications are at the core of all medium-sized and large organisations all over the world. Without them things don’t get made, sales don’t happen, deliveries fail, customers become unhappy, profits don’t get calculated, and vital things don’t get bought or just don’t arrive in time. I’m not sure if they save on paper, but they save on labour, doing those repetitive, communicative and administrative things more accurately, consistently and rapidly than humans can ever do them. They reach out across the internet to enable us to order pizza from our living rooms, or buy tickets for flights from one continent to another. They touch are lives hundreds of times a day.
They are immensely complex. The largest of them, such as SAP or Oracle, contain millions of lines of computer code, and trillions of possibilities, and they’re developed by thousands of programmers over millions of man-days. The largest (and most expensive) of them also do absolutely everything that a company needs its business systems to do- from procurement and purchasing, engineering simulation, to distribution route-planning, payroll and career-planning.
The best of them are configurable and don’t need source-code modification to make them work well for a particular customer. Take Infor’s SunSystems, for example, a beautifully designed accounting system that is easily adapted for all sorts of purposes, or systems@work’s time@work, a professional services and expense management package that works for a wide variety of professional services organisations, from consulting and engineering, to law or oil and gas exploration. Neither system comes with source code, so ‘one size fits all’ and each organisation’s quirks are managed through configuration of the software’s parameters. My company – LLP Group – has worked enthusiastically with both of these products for more than twenty years. We love them, but except when life is very simple, I don’t seem them perching entirely comfortably on the Cloud.
The reality is that there is no core business system in the world that does all the mission-critical things that a company needs only through configuration (those vital things that enable it to be competitive). And there’s no core business software system in the world that does everything. There’s usually a need for the integration of several business systems – time@work with SunSystems and Microsoft CRM, for example – to cover 95% of a company’s needs. Some of these systems, particularly those that bring a particular competitive advantage, may even have been written specifically for one company.
Integration is difficult to do in the Cloud. I take the Cloud, in this context, to be a place where a single instance of a business software system is installed and used by multiple organisations of different kinds – one software version, but configured differently for each organisation. That’s ‘cloud cuckoo’ Cloud in my experience, certainly if we’re considering those essential systems at the heart of a competitive business.
The Cloud is best when it offers either a peripheral function that isn’t mission-critical and is more or less standard across a wide variety of sectors (expense management, for example) , or a function that is essential but can be configured sufficiently to suit most organisations (CRM systems such as Salesforce, for example). But it’s no good for those idiosyncratic areas that make one company stand out from another or when an ambitious fashion-following IT or Operations Director says, ‘I want everything in the Cloud.’
I’ve seen Cloud implementations fail at the first hurdle, especially when it comes to the almost always necessary integration of multiple systems. Integration code is rarely usable by more than one organisation, and as soon as you need something special up there in the Cloud alongside the ‘one size fits all’ copies of your core business software, you’ve descended from the purest Cloud to something that’s merely ‘hosting’, because no one else can use the particular combination of software and integration code that’s been installed and developed for you.
And, even if you’re in the purest Cloud, you’re still vulnerable to what can seem like the arbitrary upgrade of the core software, which is entirely out of your control. Sadly, business software contains bugs, and most companies want to test very carefully before accepting a new version. In the purest Cloud you don’t have that option.
So, no, I don’t believe in the Cloud. I love the idea that you don’t need your own IT department, that to others can be delegated the responsibility of procuring IT infrastructure, and of ensuring safety, security and performance. That’s hosting or the ‘private Cloud’. But when it comes to the mission-critical functions and very specific integrations on which your organisation depends – forget the Cloud in its most idealistic form.