Some CiviCRM project and product statistics ...

Veröffentlicht
2009-01-23 14:21
Written by
lobo - member of the CiviCRM community - view blog guidelines
I spent some time analyzing data that we've been collecting over the past few months. In v2.1 we added a ping back feature which lets us know what component folks are using and the size of some of the tables (which gives us an estimate of the size of the site etc). All information is hashed and we do not record any IP addresses. Details of what we collect are in a prior blog post
  • So far we've received more than 8600 unique pingbacks: 5000 (64% drupal, 36% joomla) from v2.0; 3600 from v2.1(56% drupal, 43% joomla). This means that the sites installed CiviCRM successfully and navigated to the admin screen at least once. Interesting to see the large jump in 2.1 with the joomla user base. Is this because more folks have migrated to Joomla v1.5 vs Drupal 6?
  • Of the 3600 2.1 sites, 1050 sites are using CiviCRM actively. This means that they have at least 10 or more contacts. We also exclude sites with 102 contacts (the number of contacts in the sample database). Out of the 1050 active sites, 76% are Drupal and 21% are Joomla. This ratio is significantly different from the ping back ratio.
  • The number of contacts range from 11 - 517K contacts. 90% of the sites have less than 7K contacts. 48 sites have between 7K and 18K contacts and 28 sites have 18K+ contacts
  • 230 sites (22%) have more than 10 contributions with the maximum number of contributions being 75K. 80% have less than 1200 contributions and 5% have more than 7K contributions
  • 190 sites (18%) have more than 10 memberships. 75% have less than 100 memberships and 10% have greater than 4K memberships. The data indicates that approx 2/3rd of these sites have an online membership form.
  • 236 sites (22%) have event registration. 75% have less than 400 participants. From the forum traffic , we expected CiviEvent to have significantly more usage than CiviMember. the data does not seem to substantiate this assumption
  • 143 sites (14%) have sent more than 3 mailings.
  • Our forums get approx 40-80 new posts / day (dec/jan are slow months) and approx 20K page views/day. You can see these stats here
  • Our website traffic has been steadily increasing and we get around 9000 pageviews/day with 3000 visitors (67% of whom are new). This number is a bit off and contradicts the forum traffic pageviews
  • Downloads have been consistent and are approx 200 / weekday and 100 / weekend. 2.1.x has been download approx 15K times. The ratio of drupal to joomla downloads is approx 42% to 57%. You can get the download stats here.
If you have expertise in data analysis and would like to help us mine and better understand and structure this data, please contact me offline
Filed under

Comments

No standalone installs ?

And do you collect stats about numbers of groups & tags ?

I'd

I'd be interested to use that tool for management reporting: how many updates, median "olderness" (date of last modification), % of contacts with email, address, phone...

Beside providing a useful report about each install, could be very useful to benchmark against what the rest of the installs do.

X+

When you say that 10% have greater than 4,000 memberships, I presume that is 4,000 membership records as distinct from membership types?

We are looking at a situation where we would need a little over 250 different membership types, which is likely to be a little bit cumbersome - can't imagine managing 4,000 membership types!

These make interesting reading.

Would be nice to see a couple of graphs of number of contacts in database to get a better idea of the distribution :)

Anonymous (nicht überprüft)
2009-01-27 - 10:47

Don't forget that there are many of us still running CiviCRM 1.9 on Drupal 5 installations due to a lack of support by our hosting companies for InnoDB.

Anonymous (nicht überprüft)
2009-01-27 - 10:52

Don't forget that there are many of us still running CiviCRM 1.9 on Drupal 5 installations due to a lack of support by our hosting companies for InnoDB.