Have a solid case deflection method? I wanna pick your brain!


(Alissa Lockwood) #1

Hey guys, we are currently using salesforce community cloud- lightning experience. We are working on putting together metrics around case deflection and I would like to understand how others might be calculating this and how they go about their case deflection process/method. It would be the icing on the cake if you are also using community cloud-lightning experience.

I’d love to hear what works for everyone and if there are any tools you would suggest to utilize that would be good too :slight_smile:


(Richard Millington) #2

Not Salesforce specific, but you can see our methodology here:

Typically people make one of three mistakes here.

  1. They ignore indirect call deflection (i.e. the number of people who see the answers to questions and therefore don’t call customer support).

  2. They assume all visitors above resulted in call deflection (instead of asking/estimating how many would otherwise have called customer support).

  3. The end value they uncover doesn’t show up as a cost saving. This is pretty common. A company might claim they saved $50m+ in support costs, but the customer support budget hasn’t dropped by $50m.

(Nick Emmett) #3

Hey @alockwood - do you use Salesforce’s Case Creation and search features or do you have a plug in? For example we use a third party for both, Coveo - in the back end of Coveo we have dashboards set up that look to handle this on a still fairly rudimentary level but directly addressing our case creation interface. It looks at confirmed case deflection (where someone entered the create a case page, entered some information and then hit cancel after clicking on a suggested solution), assumed case deflection (similar to the above but they don’t click cancel, they just navigate away). @richard_millingtonis right in terms of things to consider, especially around the people that are viewing answers to questions but I think it’s tricky to directly relate that to true case deflection I think it depends on what their query was. I need to reconsider the linked article again to cogitate a little. :slight_smile: