Adaptation Coefficient
The number of questions that a user needs to ask so that the presence of answers in their questions does not influence the decision to continue participation. This metric shows a rolling average of the adaptation coefficient calculated based on last year’s data.
Question answered by the metric
How many questions should a user ask to adapt into the community?
Steps to measure
To calculate the adaptation coefficient, one needs to do the following.
- Select questions (topics / discussions) in the community for the last year.
- For each user, number their questions in ascending order.
- Next, iterate by the sequence number of the question, starting with one.
- For the questions with the current sequential number, count the number of questions with and without answers.
- For each group of questions (with and without answers), count how many users continued asking questions and how many left the site.
- Having four groups on hand, create a contingency table and calculate the Chi Square statistics for it.
- If the significance level value is below the threshold level, move on to questions with the next sequential number.
The result of the metric is the last sequential number of the question group with a significance level below the threshold value.
Interpretation
Ultimately, our goal is to reduce the value of this metric. The fact that users continue asking questions even when their previous question remains unanswered means that they believe that publishing a question has high value.
At the same time, if the metric value is zero, then from the first question on the site, the probability of staying in the community does not depend on receiving an answer, which is unnatural for an online community. This happens extremely rarely and rather indicates some problems with the engagement of new users. If the metric value is high (7-10), this means about the same thing - it is quite difficult to become part of your community.