Our first baby steps into cloud-based application deveolpment

I attended a Salesforce.com CIO council meeting about a year ago, and came back pretty jazzed up to move some of our stuff to "the cloud". We did a pretty deep dive on what would have been the heavy hitter $-wise, and looked at moving from Exchange to Google Apps. We had a Google partner come in to discuss it. What I was hoping for was that as we brought up concerns, the partner would say "yep, we saw that at company X, and here is how we addressed it". What we got instead was, "yep, that is an issue". The bottom line was that Google Apps is just not ready to meet the needs of the power Exchange users.

Once I recovered from that blow to my cloud enthusiasm, we looked for other things to do, namely build some of our business applications on the force.com platform. I was considerably more tentative in my approach this time, but unless we tried it, how would we know if it had a place in our application landscape.

One of our challenges in taking these first steps was the cost of doing it. We have a very stable and modern in-house infrastructure. Many of the things the SFDC folks will site as to why app development is less costly on their platform do not apply to us. I have a virtualized environment, so I don't think about buying new hardware every time I need to spin up a new app. I have an established security model. Ditto for our way to approach the database. I already use an agile development methodology. I have a highly-skilled .net development team.

We still wanted to try it, and after struggling through putting a contract together with SFDC that made sense for us, we are finally dipping our toes in the force.com waters. We have two applications in production at this point. One that does very simple high-level project management, and another to manage our IT change management process.

Would I declare success? Sort of. I think I would borrow a term from our friends at Gartner and call our experience to date promising. There are limitations to the data models on the force platform. You can code your way around them, but if I have to write a lot of code, I could just do it in .net, and it would fit in with the rest of our stuff.

On the positive note, for simple stuff like the two apps I mentioned here, it is hard to beat. The change management app I built myself (and you thought VPs were useless). It took about an hour to get the basic shell and a few workflows done, and another couple of hours of tweaking once the users go hold of it. It is hard to argue with that kind of success - built by the business owner without sucking up development resources.

We have several other projects queued up; a test case/execution management app for our QA team, expense reporting, internal purchase requisitions, to name a few. So far we have just built apps for the IT department. I'll be ready to declare success when we have deployed an app (and yes, it actually has to meet a business need) to a much larger audience.