CiviSpace failure and CiviCRM SaaS

[JoeMurray]

I notice on Lobo's blog http://civicrm.org/node/374 that having a SaaS offering similar to some of the big players is one of the holes in the CiviCRM ecosystem. I'm interested in understanding more about what made CiviSpace uneconomic a few years ago, and whether anything has changed. Are conditions better now? Would a different pricing system have a better chance of success? Were there problems with having to provide too much service for the monthly fee? Would the greater maturity of Drupal and CiviCRM now better address some marketing challenges? Was there a need for financing for marketing in order to provide enough scale?

On another note, I was always incredibly impressed with their provisioning software's ability to setup CiviMail automatically. Is that code still around somewhere?

[geilhufe]

My lessons from CivicSpace:

(1) CivicSpace was not uneconomic, but we chose a business model that had two huge investment expenditures: (a) an n-scalable and robust provisioning infrastructure and (b) a large marketing spend to attract target (i.e. unsophisticated/technical) customers. That yielded a situation where you had to attract significant investment capital and then grow very rapidly. To use poker terminology, we had to be "all in."

(2) Conditions have not changed, I think this strategy would still work and you see companies like Acquia doing similar things in the Drupal world. In 20-20 hindsight, the right approach is probably to get a stable business up and running with 1 or 2 employees and then use that as a launching pad for a bigger effort.

(3) If you define success as having a stable service for the CiviCRM community rather than reaching folks that have never heard of CiviCRM, then yes, different pricing is key. 1 - 2 people, couple hundred a month, no big ASP infrastructure, but looking like an ASP from the customer perspective ... perfectly good business, but one that requires continuous customer service, so I don't think people get to thrilled about that model.

(4) CivicSpace wasn't a matter of too much for the monthly fee. It was a matter of needing too many customers to offset investment costs and the inevitable slow ramp in customer growth. There is also a big component of having enough bodies to build sites for people... most folks don't want to have to learn the technology.

(5) I don't see the benefit of greater maturity on the CiviCRM side, other than a more feature complete solution, but riding the tail of Acquia's $7M series A would certainly make things a lot easier from a marketing perspective from the drupal side. Thought how much that helps you in the social sector is debatable.

[JoeMurray]

Thanks, David, I was hoping you might respond. I must say I'm still just speculating based on hearsay about what led CivicSpace on Demand to shut down.

A few questions:

1) Would something like Amazon's EC2 infrastructure now reduce the hardware cost of 1a)?

2) Is the provisioning software you created available in a public repository anywhere, so that it could reduce the software development cost associated with 1a)?

3) Do I understand you correctly in 3) below, one model is to provide a CiviCRM / Drupal service that is locked down but very useful to non-technical organizations with what I term e-advocacy needs, while the other is to provide a hosting service for CiviCRM customizers and configurers. The first would have lower support costs, while the second would have higher ones. Is the difference that the second has significant sysadmin support expertise required, while the first just has user help desk level costs?

4) What sort of improved maturity on CiviCRM's behalf would be useful to a SaaS offering organization? Is it something in the area of release management, or a larger community, or more stable long term funding for the core organization, or an ability to be able to requisition developer resources should the need arise?

5) In my experience, CiviCRM is still a bit complex for intelligent users who have good general computer skills (ie can use office applications and understand website fundamentals and a smidgeon of html). Preconfiguring it and locking its administration down a bit with a simpler set of admin options would help this group. Trying to hit a broader, less technical market may require greater simplification and increased usability. Comments about what worked and didn't in terms of CivicSpace's concentration on a core of maintained modules would be nice. Also, were there issues in terms of the costs of maintaining support for an migrating users from one Drupal version to another.

I take it your own interest if any would only be a well funded effort along the lines of Acquia?

[geilhufe]

(1) Yes, something like EC2 reduces the engineering complexity a lot, allowing you to focus on the Drupal and CiviCRM layers.

(2) Not sure. Much of it is very specific to our situation. There some stuff available at https://customprofiles.civicspaceondemand.org/. Hopefully more will be available.

(3) E-advocacy and hosting are *very* different markets. It is virtually impossible to serve the consultant community because they demand the flexibility of a private server, but don't have the skills to manage one (e.g. CiviMail). I strongly encourage anyone trying to run a business to avoid providing services to consultants... compete with them instead on the low end and take a referral fee on the high end. There are three axis I look at the market on:
(a) Productization. What level of productization is required? A really good product minimizes help desk support costs.
(b) Help desk needs. Do you need a 1-800 number and what depth of knowledge is your customer base?
(c) Configuration/ Set-up. How much set up is needed for your product to be useful (if you can't afford productization, you better charge a set up fee and have a few hamsters under the hood).

(4) I think CiviCRM is perfect for supporting a SaaS vendor. It was mature enough when we started CSOD, but any SaaS on top of open source needs to budget serious engineering resources for every release. I suspect a SaaS vendor on CiviCRM/Drupal would do quarterly releases with a major release (Drupal/CiviCRM upgrade) every year or so. Which is a significant problem since everyone always wants the latest and greatest... a Drupal 6 version of CSOD would probably have been released a year or more after Drupal 6.

(5) That the same as my analysis. What works: focus on a small number of core workflows. What doesn't work: providing CiviCRM/Drupal. The learning curve is too steep and once they have completed the learning curve, they can get a VPS for $20/mo and be fine without you.

I got into CiviCRM and CivicSpace not because of technology but because of my interest in tens of thousands of small groups don't have access to technology that can support their social change goals. I think there are many ways to skin that cat, few of which require venture capital.

I'm very free with my time for anyone that wants to tackle that goal.