If I may, I would like to suggest that you enable the Value Analytics survey tool in our Community software. It asks just the right questions that align with the classic 'Contact Deflection' formula. Furthermore, it the survey tool can be configured to survey all users or just those that are authenticated, so your pool of respondents can be ever-so-slightly customized in that respect. Oh, and you can also set it up so it only 'pops' on certain sections of the Community, or you can have it 'pop' Community-wide if that is your preference. Last but not least, our Value Analytics survey has the virtue of feeding its data / results directly into the Community metrics engine, so that is pretty convenient also.
But I have not really talked about metrics, have I? There are three metrics that need to be taken into consideration when judging contact deflection, and they are:
Visits
Accepted Solutions
Accepted Solution Views
But my main issue with every single simple or complex formula that uses these metrics is that it does NOT contain a simple thing that the survey DOES contain, and that is a 'Control Question'. To wit, the survey asks, "If you had not received your answer on the Community, where would you have gone next to find it" (or something to that effect). This question thus takes into account what kind of Users typically use your community. If you have a bunch of 16 year old kids that would rather crawl off to crazy parts of the internet to find the solution to their problem rather than calling or emailing you, then that needs to be taken into account. On the other hand, if your community is made up of very nice elderly people that love to hop on the phone at the drop of hat, then it is likely that they would have contacted your Support team if they had not received their answer.
You can use all the metrics math in the world to arrive at an answer, but sometimes the most simple solution is to just ask the person who will give you the answer (i.e. - Occams Razor). It is no small comfort that both Forrester Research and the Technology Services Industry Association (TSIA) have reached the same conclusion:
http😕/tsia.com/documents/Measuring_the_ROI_of_SelfService/
https😕/www.forrester.com/The+ROI+Of+Online+Customer+Service+Communities/fulltext/-/E-RES48002?docid...
But in the absence of running the survey, you can get really clever (and use an arduous path) with an alternative method. This path, however, does require that you at least have Single Sign On (SSO), and becomes even easier if you have CRM connection with the Community. It goes like this:
Get a read-out of all the people that logged into the Community in the last 3 months (their email, customer ID, or something like that) and then reconcile that list with all the people that opened Support tickets during the same time. There is a fair amount of slicing and dicing required with this methodology, and it also rests on the assumption that the unique identifiers for a Community Member are identical to the unique identifiers used in your CRM system. But, it is a pretty darned thorough way of tracking how the Community may be deflecting 'direct deflections' from authenticated users.
A weakness of the above methodology is that it does not take into account the perhaps substantially larger audience that got their answer from the Community but never logged-in to get it / see it. So in that respect, it does not take into account the passive Community Members that got the answer to their question(s) and thus constitute what are called 'indirect deflections'.
I hope this helps.
