SaaS Considerations

Ephraim Schwartz at Infoworld and now Peter Coffee at Salesforce.com have been making some pretty persuasive arguments for the the viability of SaaS – Software as a Service.Even I have succumbed in part to the SaaS siren song. I think of it as Managed Hosting of your apps on steroids. Think of SaaS as offering SAP or Oracle or Salesforce like ERP services over the Web but with two levels of user customization – configuration customization (both initial and ongoing) and then programmatic customization using your own or third party components that plug into the base system with both local APIs and global WSDL based services.

Now some of the chief benefits of SaaS are are typically described as:
1)low to minimal cost server/data center investments for an organization;
2)lower cost network and ongoing support services investments;
3)reduced development costs as core systems come “ready-to-go”;
4)Ongoing development costs reduced because users can build/buy/borrow components being developed for configuration and programmatic customization. Your development team is expanded to all the SaaS user base;
5)Immediate availability of Web 2.0 apps as SaaS operates almost exclusively from the Internet.

One of the most obvious exponents of full SaaS is Salesforce.com with its AppExchange and Apex customization features. Peter Coffee at Salesforce, has agreed from time to time to be a guest commentator, wearing both his Saleforce and broader SaaS hat. With that in mind lets look at some broad macro issues associated with SaaS.

Macro Issues with SaaS

Saas is a trade-off – one gets very fast system development, much lower hardware infrastructure costs, and presumably lower ongoing operating costs but at the price of me-too software (no better than competitors) and lack of control of future directions and updates to the software. If you get into programmatic customizations – arent you back into the development business fulltime again ?

Ephraim raises the issue of ITs increasing need for data integration and collaboration. This inevitably means continually moving data from behind protected firewalls and exposing it to security risks in transmission to the SaaS system. The ongoing costs of adhoc development and added security risk are higher than operating behind the organizational firewall.

Operating in the Cloud which brings the benefits of anyone authorized, anytime, anywhere, and nearly on any device access still has some real downside risks. First, Cloud space is not as fast and performant and responsive to performance issues as Firewall space should be(not necessarily is). Second, Cloud space has third party connection issues (my ISP is down, the T1 line has been broken, massive hack attack on IT organization or SaaS service provider is taking place). How can SaaS provider come close to be being able to offer the 99.999 uptime that more organizations are demanding of their critical IT services ?

Think worst case. A natural disaster or terrorist strike hits your SaaS provider either on site or at immediate connection to the Cloud space. The system will be down not hours but days. What are SaaS providers prepared to do right now in these worst case circumstances ?

It is interesting to note that Ephraim, originally a strong proponent of SaaS has taken writing more saliently about the trade-offs associated with the SaaS macro model. Are these the key issues or have we missed some critical, show-stopper issues ?Let me know by email here. And in the meantime read about SaaS here or attend the free conference being offered by salesforce.com on May 21st in Santa Clara.