This recurrent (famous?) topic was on the agenda of the US CPO’s summit and you can find the executive-outcome here. As titled by Purchasing.com, where I found the article, “the debate continues”, i.e. there is not really an answer.
From my perspective and based on real customer experience, I would say that ‘on demand’ (my definition: SaaS or shared platform, software and servers) is better than “behind the firewall” if the 2 conditions below are true:
- cost is your number-one driver in a software selection process (a real “on demand” solution cost at least 1/10 of a “behind a firewall”) and if
- you don’t need any immediate integration with your backend system. However, don’t forget to clarify the following 3 topics: 1) the guaranteed availability/up-time of your platform (should be higher than 97%), 2) the ownership of data stored on the platform, 3) data-recovery, at the point you want to change provider.
Of course you can integrate backends with an ‘on demand’ solution. It is possible, for sure. Now, here is the beauty of experience. What I’ve seen from my customers – not driven by costs – is that the internal strongest political forces will drive you to go “behind the firewal” when backends integration is at stake, for security/confidentiality reasons. The finance department, obviously strong and concerned, is usually the one to prefer “behind the firewall” as well as the IT department, not ready to open its safe-infrastructure to an “on demand” service provider.
What is best is not my point. I’m just trying to be pragmatic and to help the ones wondering what to do with regard to that topic, saving time, money and energy.
As a consequence, all companies not looking yet for integration should go “on demand”. For the others (looking for integration), if initially your organisation is pro-on-demand, just go for it as you will save significant money…and for the remaining, no option I would say.