I’ve got a .NET product on-premises and I want to move it to Windows Azure. How much will it cost me?
„Show me the money”
– Frank “Cush” Cushman in „Jerry Macguire”
A business of hundreds of millions, if not even billions
Windows Azure, the cloud computing platform launched by Microsoft in 2010 celebrated on February 1st 2013 three years of existence. Public information regarding income exclusively from the Windows Azure business is missing; however, if we are to look at the last quarter (Q2 2013) we notice that the division „Server & Tools” (which Windows Azure is part of) reached an income of $5.88 billion, 33% more than in Q2 2011, the last quarter when Microsoft did not have an offer for the public cloud market. We could assume that part of this growth was caused by Windows Azure, which already places the product in the big league of the Redmond company’s product portfolio.
In one of his recent statements, Bob Kelly, Vice-President of Microsoft, the marketing responsible of the Windows Azure team, stated that there are tens of thousands of subscribers to the platform and that their number is increasing by a few hundred every day. Even if they are not precise, the numbers indicated and the increase in the division’s turnover with more than a third in comparison to the moment when Windows Azure was introduced should inspire confidence to those who are seriously considering moving their products in the clouds. Beyond this, starting at the end of 2012, Microsoft Dynamics CRM Online became a citizen with full rights of the Windows Azure platform. A direct competitor for Salesforce.com, Dynamics CRM Online has been considered by Gartner the CRM product with the fastest growth, which make Microsoft’s move inspire even more trust to those who want to take the step towards the world of the public cloud.
The issue of cost or ROI
Personally, I am a supporter of the cloud movement and I believe that many on-premises products of today are going to find their places at a high altitude in the coming 5 – 8 years, if they are going to try to survive. Furthermore, I believe that if we are to talk about new products, the first question one needs to answer is “Why NOT the cloud?”. It is an already noted trend and experience shows that all companies I have worked with lately have considered at least a form of cloud when it came to the implementation of a new product.
Forrester, a research company, indicates that by the year 2020 the global market of cloud computing will reach the value of 160 billion dollars, 83% being represented by the SaaS solutions. Regardless of the accuracy of the foresight, one thing is certain – SaaS will have a major impact on ISV companies that base their income today on selling licences.
The truly difficult question addressed the value of the investment that must be made by a company that already has an on-premises product that it wants to move in the cloud and finally the rate of return on investment (ROI) on medium and long term. In 2011 Forrester Consulting carried out a study based on the TEI (Total Economic Impact) methodology on 6 ISV companies which have migrated on-premises solutions to Windows Azure in order to determine the answer to this question.
The companies included in this study are developing solutions for energy management, PoS (Point of Sale) systems, systems for reservations and for management of IT infrastructure, have a market experience between 3 and 24 years with locations both in the USA as well as in the EU and before the migration they had on-premises products based on a traditional licensing model. Five products developed using Microsoft (.NET) technologies, only one using Java. The study followed the evolution of the migration projects for three years, during which they all succeeded in reaching the market and obtaining income.
There are at least five important conclusions of this study
- Up to 80% of the target product code could be ported through a simple recompilation. The remaining 20% had to be re-written or adapted in order to use the services offered by the Windows Azure platform;
- The initial code porting, which allowed a first run in Windows Azure of the on premises product (without PaaS specific optimisations) lasted between 8 and 12 man-months;
- The production version (which includes optimisations, including the implementation of SaaS specific multi-tenant concepts) added on top of the initial effort between 5 and 24 man months;
- The operational costs regarding the hosting of the application using Windows Azure varied between $400 and $2,500 / month (the average after three years being of $953 / month). This represented a decrease of between 70% and 80% of the previous hosting costs, based on rented servers or own infrastructure (including administration);
- The annual increase owed to the SaaS solutions implemented in Windows Azure have varied between 20% and 250% in the first 9 – 14 months from the launch;
Naturally, there are many variables at play and the risks is that these numbers are interpreted in a way in which makes Windows Azure the key to success in business, and regardless of the market conditions anyone can get rich with a mere migration to the cloud. Obviously, such a conclusion would be fundamentally wrong and completely untrue. Nevertheless, the results are relevant. They answer with enough precision to those who have an on-premises product they want to migrate to the cloud. But are they enough to write a business case strictly around them? Probably not. But they can be a good starting point for a few initial calculations and surely a base on which one can build a plan for a POC (Proof of Concept). However, before all of these, one must check whether the market accepts the idea of a SaaS product and whether through such an investment business objectives are reached, such as addressing new markets (which otherwise are hard to reach because of geographical, logistical or competition-related reasons) or defining of a competitive advantage based on innovation and operational excellence.
There I’ve shown you the money
Finally, the Forrester report presents the following table, relevant for the archetype of the ISV company that represents the 6 included in the study –
Development costs are based on the medium income of an experienced .NET developer, equated at $85,000, increased by a total cost factor of 1.2 (overhead multiplicator) .It is granted that for Romania these costs are different. Even though the operational profit is not presented, we may conclude that at least in the case of the archetype we are talking of a positive ROI.
Even if in the case of some clients I interact with the fear of taking the step towards the cloud is still greater than the perception on the value of the opportunities it offers, I notice increasingly more often that those ISV companies that decided despite a sometimes reticent market to make this decision have stood to gain. Exact Software, one of the largest ISV companies in The Netherlands has recorded in 2011 a decrease of its income from selling software by 13% in Benelux, while the SaaS based solution increased by 46%, bringing an income of 11.6 million Euro in comparison to 7.9 million Euro in 2010. The same situation is to be found in the case of UNIT4, another Dutch ISV who relied early on SaaS and now is benefiting of income of over 9 million dollars from this solution.
The road ahead
Windows Azure will continue its growth in the coming period, mainly because of the new products that will be developed as SaaS on this platform and of the success of the Microsoft products that have recently adopted the platform. For myself, the fact that the future of software products, as we know them today, is closely connected with the cloud is a certainty. It remains to be seen how many ISV companies will be able to face the change, but independently of this result the numbers that I have presented above indicate the fact that for some sublimation is already a solution generating growth and profit.