- Average Scores were not reflective of the full breadth of usage, or the value people received from the platform. The previous calibration considered a score of 35 to be good. Adjusting what we measure and how the score is calibrated has seen the average score increase to 45.
- There was some ‘clustering’ of scores. Analysis showed this was due to usage frequency assumptions not being aligned with average activity across our customer base.
This has now been adjusted and the range of scores is now more widely distributed (some user behaviours are very common, so there is still some clustering, but it is less pronounced).
The way the Artesian Score works hasn’t changed for many years. Recently it has undergone some improvements - the calibration parameters have been refined to address some issues raised by customers. For example:
For advanced analysts only:
The above was achieved through dozens of small calibration changes based on events, weightings and frequencies. One of these changes is that the weightings used when combining the individual scores into the Artesian Score has been changed – although not of interest to most, we are letting you know in case you do your own analysis using score components.
The new proportions are Target: 40%, Connect: 50%; and Share: 10%.
This better represents the relative importance of each area.
Note that these weightings may change again, so if you are using these data in your own analytics, please let your CSM know so we can be sure to keep you informed.