Last month, Salesforce.com made it onto the list of the top 10 enterprise software vendors for the first time. The arrival of a software-as-a-service provider among the elite says a lot about the growing importance of the cloud in business applications.
The cloud may be on the rise, but Salesforce.com's $3.8bn 2013 revenues are dwarfed by Microsoft's $65.7bn, and by the combined $142.9bn of the top four enterprise software vendors — Microsoft, Oracle, IBM and SAP:
Although there's a lot of inertia, the cloud and software-as-a-service (SaaS), along with several other technology trends — including mobile, analytics and consumerisation — are profoundly changing the enterprise software landscape. Here's how another analyst firm, Forrester Research, sees the forces that are reshaping business applications:
Outlook: mostly cloudy
The rise of cloud and the utility model represents one of the biggest shifts in the way enterprise applications are written, deployed and consumed. The global public cloud market will be worth $191bn by 2020, up from $58bn in 2013, with cloud applications accounting for $133bn, according to Forrester.
Cloud technologies can also been seen as the biggest threat to the established order of on-premise software and the vendors that supply it.
"It's sort of scary for them because you can see a world in which people say, 'I'm going to buy order management from him, item master management from him, and vendor management from him and I will make it all work together, or the semantic web will'," Forrester Research VP and principal analyst George Lawrie told ZDNet.
To counter the growing appetite for cloud services, one of the approaches employed by big enterprise software vendors is to work with firms such as Accenture, Deloitte, Infosys and Wipro, according to Lawrie.
"'Why don't you host it? Why don't you provide a kind of managed service — a complete turnkey SAP or Oracle, or whatever it is — and we'll call that cloud'," he said.
"If you talk to them, they'll say, 'We've been doing cloud for ever'. They'll call that cloud. I've just finished working with a company that said, 'OK, it may not be cloud. But what's the difference? It walks like a duck, it quacks like a duck — I'm calling it cloud."
A second approach is the one adopted by SAP with Business ByDesign.
"They're saying, 'No, no, no — this is a native cloud application. We've written it from the ground up'. But they've been really slow to get there," Lawrie said.
"We've had lots of time to prepare for these changes. By the early to mid-2000s the signals were screaming that this change was going to occur, but still companies were unprepared."
— Simon Wardley, Leading Edge Forum
"It has no inventory at all. At first I thought, why are they doing that? It's obvious: they don't want to cannibalise their own business. They're saying, 'This is for services businesses. You're a manufacturing business. That's much manlier. You've got to have stuff on-site and your own IT people'. And actually most manufacturers believe that."
The shift to the cloud and the utility model has caught out not only vendors, but also big users of on-premise software, according to researcher Simon Wardley at CSC's Leading Edge Forum.
"When we talk about enterprise IT and all the changes going on, and ERP and CRM becoming more of a commodity provided by utility services, all of these are pretty predictable changes that have been known about for some time," he said.
"It's a process from the novel and new, to the uncharted, to the commodity, the utility, the industrialised," added Wardley.
"That impacts not only vendors that may have inertia barriers in terms of existing business models; it will also impact companies that consume it in various ways — not just in efficiency and the ability of competitors to build things more quickly, but it can also reduce barriers to entry into your business."
Research into strategy that Wardley conducted among a group of 150 companies highlighted shortcomings in organisations' grasp of fundamental changes in enterprise software.
"They have great big strategy documents full of how, what and when, and very little why. The why, when you break it down, is normally 60 to 70 percent of what other people are doing — like cloud, big data and social media. They're acting in that environment without actually understanding the landscape," he said.
"If you think about why a general bombards a hill, he doesn't do it because he's got some consultant's report saying 60 to 70 percent of generals bombard hills, therefore what you should do is bombard a hill."
That lack of insight causes obvious issues in the application of technology for things like cloud and the shift from product to utility.
"We've had lots of time to prepare for these changes. By the early to mid-2000s the signals were screaming that this change was going to occur, but still companies were unprepared," Wardley said.
"They had lots of inertia due to past practices, and best practices in the product world, which are of course going to be different to best practice in the utility world. The companies being impacted by this are generally those that have very poor situational awareness. They're surprised by the highly predictable."
Custom or commodity?
Utilities such as SaaS have their place, but they are not going to replace everything, according to Ovum principal analyst Roy Illsley.
"It is probably eventually going to get to potentially 60 to 70 percent of the market — that's what most people expect," he said.
But SaaS is also having a wider impact, by changing perceptions about the need for tailor-made applications.
"SaaS has revolutionised things because in some aspects companies are quite happy to say, 'You know what? We need to have that app bespoke. But we can have that as standard, delivered from the cloud. That's where ServiceNow and Salesforce have made a killing," said Illsley.
Forrester's George Lawrie says people are typically buying the non-differentiated systems of record and building systems of interaction.
"Nobody uses the screens that SAP delivers. They always build their own screens, which are optimised for their own experience for their employees."
—Roy Illsley, Ovum"
"Those systems of interaction are the ones that have the consumerisation and which are differentiated for their audience," he said.
"For example, nobody uses the screens that SAP delivers. They always build their own screens — always, always, always — which are optimised for their own experience for their employees."
However, businesses may be fooling themselves about the significance of the customisations they are undertaking, according to CSC's Simon Wardley.
He once asked how many of a group of 100 CIOs had ERP.
"One hundred hands went up. Then I said, 'Well, it's commodity if you've all got it'. Two arms shot up saying, 'No, no, it's not commodity' because of their customisations, which made it special," Wardley said.
"So then I asked them what their customisations were. Of course, they were very secretive but after a bit of badgering they told me one, and so I asked the entire room, 'Does anybody have this?'. One hundred hands went up."
This group of CIOs was spending billions of dollars a year in total customising pretty much the same systems in almost identical ways, creating no differential value for any of them, said Wardley.
According to Ovum's Roy Illsley, the effect of companies like Salesforce.com has been to change what businesses focus resources on. They are now often happy to adopt a standard SaaS CRM system, for example.
"It's the same as everybody else's. It's not the tail wagging the dog like it used to be," he said.
"Of course, you're going to get slight variations where people want bespoke stuff. But in general it seems in this world everybody else does something like this, so how does that differentiate this company from that company?"
It's not the app that makes a difference — it's what a company does that rivals can't see and copy, Illsley added.
"It's not that little bit of software that tells us where customers have bought it or their balance is X — every bit of software can do that."
Inertia and other cloud barriers
Nevertheless, the attractions of cloud-based applications may not be enough to trump business inertia, according to CSC's Simon Wardley.
"Of course existing consumers will often have resistance to that change for numerous reasons — existing processes, political capital. There are often all sorts of inertia barriers," he said.
"It's difficult to say, 'That billion-dollar ERP system I've just implemented, I can now buy on a credit card' without looking foolish."
According to Wardley, inertia is inevitable among user organisations as well as among vendors.
"People will always try and hold on. They always want the past so that it doesn't cost anything to change but they want it like the future."
"Unfortunately, you can't have volume operations customised for you and commodity provided with non-commodity components. It doesn't work," said Wardley.
But there can be sounder business reasons for sticking with traditional on-premise approaches, according to Ovum's Roy Illsley.
"There are still some traditional companies that use apps running specifically for their business purposes. So if you're a bank, your core banking applications are almost certainly still running on a mainframe," he said.
"If a core banking application works, you're not going to throw it out and spend millions of pounds to replace it with something that you're not actually sure of."
— Roy Illsley, Ovum
"They're a completely different set of apps to the ones if you're something like a PayPal, which could be construed to be a bank but it hasn't got all the same baggage that a bank has."
"They're probably running things in a slightly different way because they're a newer organisation that has grown up without having all this legacy stuff that works," said Illsley.
"If a core banking application works, you're not going to throw it out and spend millions of pounds to replace it with something that you're not actually sure of."
However, elements of ERP are breaking away and being used by businesses as discrete software components.
"Because [companies] do want to be connected but actually, as long as they can collaborate and see and share information, it doesn't need to be a monolithic great big app that runs absolutely everything," Illsley said.
"It can be discrete fragmented apps — with the apps being run locally but the data and management being centralised."
As examples, he cited airlines using iPads to locate passengers without printing out reams of paper, and restaurants using iPads for bookings, orders and recording customer names.
"The mobility aspect is being continually rolled out and that needs the apps to work in a different way because you can't necessarily have a full ERP capability running on an iPad. You're going to have to do something slightly differently," Illsley said.
"That 'something slightly differently' may be a front-end app sending information back and then you've got the app at the back end that just does the back-end processing, and that's standard."
"How you crunch numbers, how you docket the numbers, how you present the numbers in a financial market or how you count stock — that's standard stuff that heavy-duty processing can do but it's almost certainly happening in real time now and not anything like batch processing."
Mobility, real-time applications and consumerisation
The increasing provision of services to the device is why mobility and the demand for real-time information are two of the big enterprise software trends, along with the cloud.
"The cloud has chipped away, SaaS has chipped away at elements of enterprise apps and you see elements of it twisting off into cloud and SaaS-like operations," Illsley said.
"But with mobility and real-time applications, what you're now getting is the demand for the tool to deliver faster to more places the right sort of capability for the people who are using it."
"You're not expecting them to come in and sit at a terminal and use green-on-black to fill out a form when taking an order. Those days are well and truly gone."
The rippling impact of consumerisation has affected not only the nature of mobile applications, but also the whole look and feel of enterprise software, according to Forrester's George Lawrie.
"If it isn't shiny and iPady, they don't want to use it. So that consumerisation has driven another leap in the way the vendors need to provide the applications," he said.
"You can't tell people they've got to go on a week's course, and you can't keep telling them they're wrong. Which has typically been the way with enterprise systems: if anything doesn't work, it's your fault."
—George Lawrie, Forrester Research
"And it's much more complicated than it was for client-server because these enterprise applications, as soon as they start to become instantly consumable, you can provide them not just to employees but maybe to your customers' employees and maybe even to consumers."
"As soon as you do that, you've got to have a different kind of infrastructure because you now don't know how many hits you are going to get on your system," said Lawrie.
"First of all you've got to componentise those applications. So all that effort that's gone into making it very monolithic — 'It's all got to go into one database or it will never work' — that all just doesn't really work anymore. It's got to be much more subtle."
Just like consumer applications, business software has to be intuitive and immediately usable.
"You can't tell people they've got to go on a week's course and you can't keep telling them they're wrong. Which has typically been the way with enterprise systems: if anything doesn't work, it's your fault," Lawrie said.
"That's not really acceptable to people who are used to buying a holiday online or ordering something from Amazon."
What has also changed in organisations is there's less time and resources for learning, Ovum's Roy Illsley said.
"So what they need is for people now to be more general and wear more than one hat. You've got people needing to be able to pick up other areas of the business they're not familiar with to be able to use the tool, and to do that it's not got to be shrouded in technology and language, and specific processes that they can't get their head around."
The underlying trend building up across all applications is that software must not only do its job and be easy to use, but it must also provide collaboration and sharing, and be simple to deploy and support — and, where possible, user self-supporting, Illsley said.
"A lot more investment has been in either how it looks and it feels for the people who use it and what they get out of it, and into making it easier to manage from a back-end perspective, so that the cost of using the app is coming down," he said.
Automation, self-service and analytics
The drive for software cost reduction and ease of use, both for business consumers of IT and tech staff, is fuelling a drive for greater automation, according to Forrester's George Lawrie.
"There's an extent to which this stuff gets automated. We've seen that with the virtualisation of servers and networks and storage. You can spin up a new server out of your datacentre without you having to do any manual work at all," he said.
"All that kind of stuff that operations people used to do — that's all automated, eventually it's standardised and then it's given away. You shouldn't even be doing it. You should be pooling it with someone else."
The second aspect of this trend is increasing self-service features for end users.
"All those days when they used to say, 'Can you do a report for me? No. Do your own report. Here's the data. Here's the tool. Have at it," Lawrie said.
As well as breaking applications into blocks, consumerisation is also leading to changes in business process management and rules, so where rules and processes were once embedded in code, that's increasingly no longer the case.
"There will be business rules — for discounts, for example — that would never be in the program now. They'd always be in a table that users can maintain easily themselves. "So you want to change the policy? There's the table. Change it. You want to change the process? We're going to do the credit check before we ship instead of when we take the order. There you are, you can do it. It's not an IT thing," Lawrie said.
That philosophy is also making itself felt in the embedding of analytics into applications, both in mobile and on-premise enterprise software. According to Lawrie, those analytics can involve real heavy lifting, mathematics, computation, and some exploration of correlations.
"Increasingly, the idea is to throw data into a big lake and bring the analytic to the data, not the other way round. When we say analytic we mean, 'Let's take all these observations and let's see what correlates to something else' — the unknown-unknown problem," he said.
"With the elastic cloud and with some new software, that means completely new insights for end users and potentially their being able to orchestrate the rules of their transaction systems alongside these systems that are giving them insights. So they can test and learn in a way they never could before."
Retailers have been doing these types of analytics for years offline, but now they are becoming more common thanks to technologies such as in-memory computing.
"You're beginning to see people doing this inline. So at any utility or telecoms company, the person who is probably the lowest-paid person — the customer service person — will have in front of them something that estimates your lifetime value to that company," Lawrie said.
Ovum's Roy Illsley cites retail as an area that's increasingly seeing the value of being able to deliver real-time dynamic price updates to stores.
"Traditionally, price updates were something that you run overnight. You produce a great big file and send it out next morning to the store," he said.
"But because the technology is there and the applications are now being tweaked to enable it, you can now get real-time dynamic price updates out from a central location to all the stores to say, 'It's hot weather. Let's increase ice cream — bash. Increase the beer price — bash."
Because the apps can take account of external events, decisions can be made faster — and the infrastructure now can support that delivery of information. The app and the infrastructure have come together to enable the real-time delivery.
"If you look at what's happened in the infrastructure, you've gone from the monolithic app running on a physical server in a datacentre with one app per server architecture," Illsley said.
"The cloud and virtualisation have come along and now you've got apps that are infrastructure-aware and that can scale themselves. How infrastructure can be turned into a flexible resource is something the apps have now cottoned on to. That's a big shift in app development."