What does the steady drift from on-premise to cloud computing mean for vendors and customers? Much has been written about the advantages for customers. A survey May in 2014 by The Information Difference found that 49% of companies now have operational cloud deployments, with just 12% of companies having no plans to move to cloud computing at all. Some 39% of survey respondents believed that cloud deployments were better value than on-premise, with just 7% thinking the opposite. With just 10% saying that more than a quarter of their applications were cloud-based, there is clearly plenty of potential for this trend to widen considerably as customers gradually replace or migrate their old applications. With a mere 2% of survey respondents reckoning cloud applications cost more to maintain compared to 52% claiming lower costs, it would seem a one way street for applications from on-premise to cloud is likely.

The implications for software vendors may be less rosy. The poster child for cloud, Salesforce.com, has been a fine investment for its stockholders based on its strong revenue growth over the last decade (a 100-fold stock price jump since its IPO). In 2013 its revenue grew 36% compared to an industry average of 11%, but profitability has been elusive, with continuing losses even as its revenue has grown. It is a similar story with other cloud vendors, at least those whose financials are public: heady growth, heavy losses.

Established software vendors face a dilemma. Their customers like cloud computing, with its usage-based pricing and lower maintenance effort. IBM showed 65% growth in its cloud business even as its overall revenues slid 5% in Q4 2013. But entrenched industry giants like IBM, Oracle, SAP and others have traditionally had enviable profit margins from their traditional on-premise software model. This is not hard to understand: customers pay up-front for software, frequently millions of dollars, as well as hefty maintenance fees, whether or not they actually succeed with the software, or even use it at all. A global company deploying a new software application will usually have plans to deploy software widely, and prefer to pay an up-front enterprise software lump sum in order to cap its exposure to licence fees. However this involves the risk that their software take-up turns out to be slow or limited, for whatever reason. The vendor just pockets their large fee and moves on to the next sale. With cloud deployments the usual pricing model is subscription based, based on number of users or some other usage metric, so customers effectively pay just for the software that they use.

If this new world is proving tricky to generate profits for the likes of Salesforce.com, imagine the threat that it constitutes for vendors structured around a traditional on-premise sales model. These companies have large sales teams used to paying high salaries and hefty commissions provided the big deals keep coming on. If these deals start to dry up as more and more customers start to move cloudwards, many will struggle to adjust. Even if they do so, the current economics of the pure play cloud vendors hardly presents a rosy future.

This may partly be a matter of time. One enterprise software vendor always had a subscription-only model, and has done very nicely thank you. That company is SAS institute, whose privately-held share status (just two individuals own it) has meant that they have been able to patiently grow at a sustainable pace, without having to answer to Wall Street growth pressures. This company has had 38 years of growth and profitability, turning over more than $3 billion last year.

The SAS Institute model shows that a subscription model does not necessarily imply a future of ever-deepening red ink for software company revenues. However this is a company that was designed for such a steady as she goes model. At present the software world seems divided between pure cloud companies going for headlong revenue growth (with investors willing to be patient so far) and established enterprise vendors who cannot ignore the cloud steamroller, yet are dreading seeing their profitable on-premise software sales cannibalised by the far less profitable cloud software.

It may be that, as companies adjust to the new world, the business model with catch up with the sales growth, and that profitability will eventually come in the world of cloud. However with stock analysts fixated on revenue growth, adding more and more sales and marketing costs may make this journey a long one. The comparative ease of migrating from one cloud vendor to another, compared with shifting off embedded on-premise software, represents another challenge. If customers can switch software less painfully, then vendors will actually need to deliver sustained value to their customers, not just deliver a great sales pitch and run away to the next presentation while banking their up-front software fee.