Is utility computing a mere pipe dream?

Illustration: David Lyttleton

It is inconceivable that any company could seriously contemplate operating
without essential utilities such as electricity or water. Yet, despite the
obvious need for such core services, practically no enterprise would contemplate
incurring the capital expense of, for example, building a power station to
provide electricity or constructing its own reservoir to keep the corporate
water coolers topped up.

And, according to the fast-growing number of proponents of so-called utility
computing, IT should be considered in the same way as these other essential
services and purchased from a third-party specialist provider, rather than being
managed in-house. One of the fastest-growing areas of utility computing for
enterprises is the strategy of buying hosted applications, dubbed
software-as-a-service (SaaS). This model of delivering software over the
internet eliminates the need for companies to buy, build, manage and maintain
key enterprise applications such as accounting, human resources, CRM and ERP.

One of the chief advocates of SaaS strategy is Nicholas Carr, former
executive editor of the Harvard Business Review and acclaimed business writer.
According to Carr, the advantages of buying in IT services from third-party
utility providers are compelling. He believes they include reduction of capital
and operational costs, lowering of maintenance expenses, enabling of rapid
implementation and, ultimately, increasing profits.

“Since the industrial revolution, we have seen numerous examples of
infrastructure technology, such as steam engines, electricity, telegraph,
telephone systems and highway systems, and it is my argument that IT falls into
that category. It is fundamentally an infrastructure technology,” Carr argued at
a recent Microsoft-organised debate in London.

Causing a stir
Carr originally came to prominence when he sparked a storm of controversy three
years ago with his article, IT Doesn’t Matter. While the piece was widely
praised at the time as being the IT article that had been read by more business
leaders than any other, the ensuing debate saw Carr come under withering attack
from members of the technology industry. He was variously described as “brain
dead” and “from another planet”, while at the same time his commentary was
vilified as “a gross cancer creeping across society”.

Paying over the odds
These strong reactions have abated somewhat three years on, but Carr is still
arguing passionately that most companies are paying too much for basic
commoditised IT services that give them little or no competitive edge over their
rivals. He asserts that savvy financial directors should recognise this
commodity nature of IT today and ensure that their companies treat technology as
a basic utility that should cost as little as possible.

“I think we are at the stage where most of IT is so commoditised that the
power in the market has shifted from the vendor to the user. And the user has
very many more options and more leverage to drive down the cost of computing and
the cost of IT in general – sometimes in very dramatic ways,” says Carr.

Because of this commoditisation of IT, Carr predicts that
software-as-a-service computing will become more important, particularly for
firms that are suffering from capital constraints and so find it very appealing
that they are able to treat technology as a predictable and manageable
operational expense rather than a capital investment.

“I think we are just at the beginning of the era in which more IT
capabilities – both hardware and software – will be delivered over the internet
from outside, in essence utility, providers,” said Carr.

“One of the big practical questions for companies going forward is, what do
we have to own ourselves and what can we take from the outside utility
suppliers. And I think the way this question is answered – where the dividing
line is put – will change over time as the capabilities of the utility pr
oviders grow.”

It is hardly surprising that Carr’s clarion call – urging firms to minimise
spend on technology that is commoditised and thus not able to confer strategic
advantage – is anathema to Microsoft, which has long been one of the biggest
winners from the current IT status quo. Any widespread moves towards so
ftware-as-a- service, in particular, could be disruptive for the Redmond giant,
which currently makes the vast majority of its money from sales of traditional
packaged or pre-installed software.

Robert McDowell of Microsoft hit back at Carr’s assertions that IT is a
non-strategic commodity: “I believe innovation, if anything, will increase over
the next five to 10 years and yet I agree that there are pieces of technology,
including our own product set, that are commoditised. Not only do I agree that
they are commoditised, I am proud of the fact. Commoditised implies that, not
only are most people using it, but also that the relative cost is low, that it
works pretty well and most people know how to use it,” he said. “But to compare
IT as a whole to electricity is really a leap I’m not prepared to make.

“One of the things I fear is that the argument that IT is not strategic, but
is merely a commodity and something worthy only of cost reductions, will be
accepted. If CEOs and business leaders back off their engagement with IT, I
think there is going to be less chance to add value.”

Paul White, director of Microsoft’s UK Dynamics Product Group, added that the
software firm was busy repositioning its business to accommodate any shifts
towards SaaS. “Microsoft wants to serve customers in any which way they want to
be served and we will ensure that we can do so. We have the resources to compete
very effectively in the hosted space and the bottom line is that competition is
good for us and the customer. The basis of competition has changed to become
software-as-a-service, which is a new attribute to competition. We look forward
to that; we are not frightened of it and we are not shy of it,” he says.

Outsourcing IT
As Microsoft wades into the SaaS fray, IT industry analysts report that
enterprises are becoming increasingly enthusiastic about the advantages of
outsourcing at least some of their applications to specialist SaaS partners.
David Bradshaw, principal analyst at research firm Ovum, urges finance directors
and other board-level business leaders to look carefully at the potential
commercial benefits of buying in software as a service.

“Software-as-a-service is, in fact, really just a continuation of existing
trends in the applications market for making applications easier and cheaper to
adopt. Finance directors should definitely consider software-as-a-service as a
realistic option. They should certainly not discount it and should be looking at
the functionality that an enterprise application provides and not exclude
software as a service on the basis that it may be perceived as unconventional,”
says Bradshaw.

Resisting change
It is inevitable that there will be a certain amount of natural resistance,
particularly among enterprises in more conservative sectors, such as the City,
Bradshaw adds. “But even for these firms, most are effectively using SaaS
already, in the form of services such as Experian, which, of course, is a hosted
application – even if many companies may not think of it as such. So making the
switch to SaaS is hardly radical, even in this sector.”

This upbeat assessment of the current utility computing market was echoed by
IDC. According to the IT industry analyst, the acceptance of SaaS delivery
models, which includes software on-demand and hosted application management,
continued to gather momentum in 2005, which is expected to further increase in
2006. IDC believes that SaaS providers will pay particular attention to
developing and increasing partnerships to expand revenue streams and customer
adoption in the upcoming year.

“The role of SaaS delivery in the software industry continues to grow.
Traditional software players have recognised the importance of SaaS and numerous
SaaS-focused companies are entering the marketplace,” said Erin Traudt, research
analyst for IDC’s software-as-a-service program. “IDC believes that these types
of disruptive business models will ripple throughout a variety of software
markets in 2006.”

“Current SaaS adoption is just the tip of the iceberg, leaving plenty of
opportunity for providers and partners,” Traudt adds. “Customer demand will play
an integral role in fuelling the development of SaaS solutions and ecosystems.”

Other IT market watchers concur. Pacific Crest estimates the market for
hosted software will grow 25% a year to $10bn (£5.8bn) by 2009. AMR Research’s
2005 software-as-a-service syndicated research study shows that more than 78% of
500 respondents across major vertical industries and company sizes are currently
using, or considering software as a service. Only 18% said they have no plans to
consider software as a service.

Specialist needs
However, even in the face of such bullish predictions for the future growth of
the SaaS model, Lindsey Armstrong, president EMEA for SaaS vendor
Salesforce.com concedes that there will always be a need for specialist software
that is tailored to the needs of specific companies in specific sectors.

“I don’t think there will be a wholesale move across to 100% on-demand – not
in the near future. I think the move will be quite slow. Gartner says that, by
2010, 30% of software revenues will be from on-demand services, which represents
steady progress,” Armstrong says. “What it does say is that there is still going
to be software built on-premises and I think that the difference is that this
software is going to be so specific to your business or industry as it requires
an in-house requirement and in-house expertise.”

Good management
Carr accepts that, for specific companies operating in specific sectors, there
can be valid commercial reasons for developing bespoke and relatively expensive
in-house IT systems. However, for the vast majority of companies, he advises
being as far away from the “bleeding edge” of technology as possible.

“Rather than focus on distinctive IT, I think we are at the stage where
successful IT management comes down to good management. IT has become more like
any other resource that a company needs to manage and those that manage it most
effectively will be the most successful, not necessarily those that are most
innovative and take the most chances,” says Carr.

Calculating the cost of SaaS
To help prospective adoptors of SaaS, Forrester Research has developed a
methodology dubbed TEI (total economic impact) to quantify the respective
strengths and weaknesses of in-house versus hosted enterprise applications. This
system requires analysis of the following criteria:

• Costs – In general, on-premise deployments require greater upfront costs,
especially when comparing licence fees with subscription fees. On-premise
deployments typically require hardware maintenance, upgrades and support costs
on top of licence fees. SaaS deployments avoid charges for upgrades, hardware,
temporary business support and end-user support and administration. Over the
long run, yearly costs remain lower for on-premise software, but increase during
upgrade cycles.

• Benefits – SaaS deployment options bring immediate business benefits, but
also frequent, automatic updates, shorter time to deploy, independence from IT
and improved usability. On the on-premise front, integration benefits are more
pronounced than with SaaS.

• Flexibility – SaaS and on-premise approaches deliver scalability, technical
flexibility and ease of configuration. However, on-premise systems deliver
proven integration capabilities, better customisation options and tailored
configuration capacity. SaaS deployments deliver technical agility by allowing
migration to on-premise, reduced IT staffing requirements and greater deployment
options in bandwidth-constrained environments. Ultimately, the advantages in
flexibility in one model are easily cancelled out by the other.

• Risks – SaaS options tend to centre around impact risks like loss of
control, weaker integration, limited verticalisation and customisation
capabilities. On-premise risks tend to relate to implementation risks, such as
deployment complexities, training needs and support issues. Hence, on-premise
risks remain slightly greater than SaaS in magnitude.