The first month of the year is already out. A couple of weeks ago I wrote about where I expect to see innovations in SaaS applications during this year. Before we get too far into this new year I thought I’d write a few words about the challenges that SaaS applications must address during 2010. We need to make headway towards these challenges to be addressed, not only for the adoption of SaaS applications to broaden but also in order for these applications to be considered for replacement of on-premise mission-critical applications.
In a recent Goldman Sachs survey, published on 1/10/10, CIOs listed the following concerns with SaaS and cloud computing (in priority order): security, solution maturity, Total Cost of Ownership (TCO), solution complexity, and solution reliability. Interestingly the priority order of the concerns was identical in the similar survey that Goldman conducted at the end of 2008. CIOs, however, perceive that that during 2009 SaaS vendors have made progress on all but the TCO area. In particular, while 84% of the surveyed CIOs identified security as the top issue, the corresponding 2009 number was 52%. While the difference between the 2008 and 2009 numbers for the TCO issue is not significant, statistically the number is the same, it is interesting to note that the in every other area there has been progress except in the area of addressing TCO.
Based on my conversations with executives of our SaaS portfolio companies and other SaaS companies we considered investing, as well as direct discussions with CIOs and business executives in companies that have licensed or are considering SaaS applications, my list of challenges includes:
1. Security. The recent attacks on Google’s infrastructure and Gmail showcased the risks that SaaS users would want addressed before they feel comfortable migrating mission-critical on-premise applications and data to the cloud. I feel that vendors are starting to make solid inroads in addressing regulatory and corporate governance standards but cyber security is another matter.
2. Application reliability/uptime. We are seeing requirements moving from 3-nines uptime (99.9%) to 4-nines uptime (99.99%). Obviously the more mission-critical the application, the higher the necessary reliability. Salesforce’s recent outages showcased the need for improved reliability. Meeting this requirement will require more advanced data center architectures, an issue that smaller SaaS companies have not started to address yet by and large.
3. Data integrity. The data that leaves a company’s data center to move to the cloud must not be corrupted en route. I would imagine that data integrity has to be at 5-nines. This issue is particularly important as customers are now asking for the integration of data from several applications, some of which reside behind their firewall and others reside in application vendor clouds. Achieving this level of integrity will also require that SaaS vendors rely on each other to individually achieve this level of data integrity.
4. Application performance. The application’s performance should not degrade as more users access the application. Also, as applications become more popular within organizations we find that users are no longer willing to tolerate slower applications. They are now starting to expect identical response to the corresponding on-premise applications. I expect that once this issue is addressed improving the application onboarding process will be the next user challenge.
5. Data movement/throughput. Customers are moving increasingly larger data quantities (now at the low-terabyte level) to the cloud. This has been particularly evident in SaaS BI applications. SaaS companies must address this issue whether they use their own infrastructures or IaaS like Amazon’s.
6. TCO. Thus far SaaS application vendors have been able to demonstrate qualitatively the lower TCO of their solution over corresponding on-premise solutions. This was due to the lack of detailed and extensive history operating SaaS applications. However, now more data exists and SaaS vendors should be able to easily demonstrate the superior TCO of their application.
I look forward to creating a scorecard at the end of the year detailing whether and by how much these challenges were addressed.