SAAS still unknown quantity

WHEN enterprise software giant SAP reported its latest financial results a month ago, slow progress for its mid-market software-as-a-service play was just about the only blemish on an otherwise excellent quarter.

For a company the size of SAP, even runaway success with its embryonic Business ByDemand offering - launched across a trial group of a half-dozen markets almost a year ago now - would barely have moved the meter.

The struggle SAP has had in getting early traction in SAAS is a hint of the challenges facing incumbent software vendors in shifting their product lines and economic model to a subscription approach.

Nobody really knows as yet what enterprise-oriented SAAS will evolve into, but trailblazers such as Salesforce.com offer a hint.

Comparisons are tricky because Salesforce.com is still in a phase of fairly rapid growth, but on the available evidence it appears likely SAAS will involve higher expenditure on marketing and sales per dollar of revenue, and lower margins than traditional software licensing.

Over the past two years Salesforce.com has consistently spent an amount equal to about 50 per cent of revenues on sales and marketing.

In comparison, Oracle and Microsoft both spend about 25 per cent of revenue on sales and marketing.

Higher spending on attracting and retaining customers in a subscription model presumably reflects a couple of influences - first, that users have to be persuaded to adopt a new and unfamiliar way of meeting IT needs, and second, that hanging on to existing customers will get harder (and more costly) if SAAS makes it easier to switch to competitors.

Both these trends speak, in turn, to the question of margins, which historically have been very high in most of the software industry, but look likely to be a fair bit lower in the emerging subscription-based service environment.

There's no intrinsic reason why a subscription model should lead to lower average revenues per user, but several forces in play could have this effect.

The first is that switching costs are likely to be lower, partly because most users won't buy into a subscription model of paying for IT functionality unless they feel comfortable that they retain some control of their data and destiny, and also because it is harder to build barriers to switching into SAAS offerings, especially if they use open standards and interfaces.

Software companies have traditionally exploited high switching costs through hefty annual maintenance charges, which after a while turn into a vast ongoing annuity stream - maintenance accounts for more than half of revenues at both Oracle and SAP, for instance.

The second reason why average revenues (and margins) per user may be lower in a subscription model is that users only purchase what they use. This may take the form of buying only the features and functions required - unlike the all-or-nothing bundling of features in much packaged software - or paying for actual volume of usage, rather than per seat.

The third force likely to eat into margins is that SAAS marks a departure from the pure write-it-once and sell-it-forever scalability of software.

The idea that it costs literally nothing to make and sell another copy of a piece of software has never quite been true in practice, but because of factors such as support costs and feature creep it's pretty close.

For SAAS, in contrast, a much higher proportion of the costs of acquiring additional customers will rise in line with market penetration, which requires bandwidth, servers to run the service, storage, power and so on.

Finally, much software appears to be becoming less differentiated and more commodified regardless of how it is delivered.

The rise of open source alternatives, the maturing of many parts of the IT industry, and convergence around stable, open standards and technologies are all eroding the mystery and margins of the past.

None of these trends are particularly enticing for the traditional software industry, and least of all for leaders such as SAP, Oracle, Microsoft and Symantec.

The leaders at these firms are too smart to think they can somehow escape the coming transformation, but if they can put it off, they will.