Tuesday, March 31, 2009

What it would take to start a CiviCRM ASP


So we tried to start a CiviCRM/Drupal based ASP to solve the constituent relationship management/website/online donation/ mass email problem that most charities face with CivicSpace.  It failed, but that does not mean that another attempt will also fail.

There are three basic approaches to doing a CiviCRM/Drupal ASP:
  • Technology first
  • Customer first
  • Hamster first
The technology first approach is building out the infrastructure to handle a high volume, self-service ASP.... low monthly price and high customer volume. This requires either piles of money or the super-committed technical geek founder to do the work. It relies on the build it first, then find the market approch. We did that at CivicSpace and we "ran out of runway".

Customer first says lets go out and build a lot of demand. Sure it will be really labor intensive to maintain the technology infrastructure and initially the customer service will not be great, but you avoid solving the technology problem until you have the real problem of too many customers and you need to build automation technology.

Hamster first is buy a VPS, put up a cool web page, market your product and hope for the best. The technology stack (CiviCRM/Drupal)  is actually fine for this approach at the moment, but you'll face bulk mail deliverability, scalability, performance and other issues along the way.

Tech and customer both require a fair amount of capital to pay for the technology development (the ASP platform) or the marketing (making the service known in a very crowded vendor space). Hamster first could financially support a single consultant and once they have a working model, could easily be put in front of investors to attrach "expansion" capital rather than "start up" capital. 

The other trap is the set up fees. We tried to make things self service... life is just too complicated. There has to be a set up service before your customer starts paying their monthly fee. My feeling is copy success... i.e. copy PicNet who have built a similar business on Joomla. People pay a couple thou to get started and then a monthly fee. I think they cracked an important part of the code.


3 comments:

ian said...

I would be interested in hearing the detail behind "ran out of runway". Couldn't you leave the service up while you all took day jobs and give customer support your best effort after hours? Or were the performance issues with CiviCRM such a problem that they were going to require heavy investment to fix?

I'm also interested in how you extended/modified CiviCRM to make it self-service. Did you have to touch CiviCRM code (in which case I think you have to share that code with the community) or did you find a way to do it with the API?

My wife manages a few associations that need this kind of service, so I'd like to use that to start an ASP, but I'm not sure CiviCRM is the way to go. Thx.

David Geilhufe said...

There are lots of aspects to running out of runway. We felt that to effectively serve customers, nights and weekends best efforts were not going to cut it.

There were no performance issues with CiviCRM, Drupal or the ASP delivery infrastructure we created - and it was tested to scale past 10K customers, but when you have to pay cash for your technical staff, the "best effort" model breaks down.

Making Civi available as an ASP doesn't require changes to Civi... you are basically provisioning Civi to each customer, you can do that in a shared model or VPS model. Upgrading Civi also doesn't require Civi modifications.

I'm sure some modification to Civi are always needed and you can get those done yourself and contribute them back or get them done through the core team.

If your starting point is to put together an ASP, then you have three routes... build it from the ground up, start with Civi, or start with a cloud provider like Salesforce or NetSuite.

I still think Civi is the best long term route (lower product development costs). but you are building an ASP based on PHP applications which is not commonplace.

Lee Wright said...

Found you (and this old post) after a friend tipped me off to civicrm.

I'm on the board of a small, non-profit, all-volunteer, member-supported historical society in Massachusetts. Most of the ~1000 local historical societies throughout New England are similar to ours.

We've used hosted solutions for ticket sales--free for free events; a small fee for fee events--and started using VerticalResponse for e-mail when they offered it free to non-profits (@ <10,000 messages/month).

I've been looking for a simple hosted solution for membership management, but have been put off by the high cost (for us) and, in some cases, the clunky functionality.

By comparison, civicrm seems to have the potential to be a perfect solution--if it were offered on a SaaS model at a price we could afford.

In terms of business model, it seems that the now-typical freemium approach, combined with taking a percentage based on financial transactions (memberships and event tickets) would work.

I was disappointed to see that relatively recent efforts to offer this on a hosted basis were no longer up-and-running, at least not that I could find.

Any update on this--either operational now or planned--since your post here (or in the discussion thread last year at civicrm)?