How to get more deflection data

LiselotteP · 2024-09-06T07:50:03+00:00

There was an error rendering this rich post.

LiselotteP

Hi,

I'm struggling with my deflection data, the numbers as so low to get any decent ballpark figure.

Discussions: how to encourage users to mark an answer as accepted? I'm thinking of PMing each user a few days after a discussion has been answered to encourage them to mark it as accepted, but this is a lot of manual work. Does anyone use any automatic email system?
KB articles: we get 30k-ish in page views on the KB's alone every month, but only 1 or 2 votes if it was helpful or not. Our developer has added some colors to add pizzazz, but its not enough. Any way to encourage users to vote here?

Surveys: a pop up shows up after 10-20 seconds a user is on any page, which has increased deflection data, but not by much. We get about 1 response every 2 days, despite the visit traffic over 2 days being just under 700 visits. That's a response rate of 0,3% at best…

Thanks for your insights and feedback!

Liselotte

Find more posts tagged with

Accepted answers

All comments

Genevieve P

Hi hi!

I can help with the first question: "how to encourage users to mark an answer as accepted?"

We recently discovered a setting you can toggle on which auto-emails someone regarding their question if there's no accepted answer.

Here's the documentation: https://success.vanillaforums.com/kb/articles/20-questions-answers-q-a#q-a-follow-up-notifications

BrendanP

That's a good suggestion by @Genevieve P

We do have a ROI calculator dashboard to help to quantify deflection but it's primarily a calculation based on pageviews and accepted answers that is fairly close to being released. @Shauna could share more about that. No one has a perfect science here for a few of the reasons you listed (does everyone say the article was helpful when it was in fact helpful?).

Another angle to consider could be if you could show members vs non-members. Do you currently sync any activity back to your Salesforce CRM? If the members are in SF and you had that activity, you could build a report to compare members to non-members.

ChurnZero did that with the integration they built with their own software in Vanilla and noted that community members submit 67% fewer product/support related questions to support versus non-members.

@stephanierichardson @sophielyons I know your team uses Qualtrics like @LiselotteP 's team - have you found any tricks to get more responses?

sophielyons

On a weekly basis we triage all Q&A without a best answer marked. This is done using CSVs uploaded into a mega spreadsheet. Depending on the last response we take different actions:

Flag any that are waiting for additional help to our Support Team
Ask the author if the query is still open if they haven't responded to follow up questions
Ask the author to mark a best answer

After 6 weeks from the authors last comment, if they haven't responded to us we either close the thread or if there is a definite best answer we mark this ourselves.

So far the volume has been manageable and we've seen a lot of value for the business, customers, and other Community members from triaging them. We're starting to think about how we can tweak the process as we begin to scale. We've recently procured zapier so will explore if we can use this once Q&A status is an available trigger (I think it should be coming very soon).

In terms of Qualtrics, we have a survey that's always available for members to fill out. We've just about got an ideal sample size (it's taken about a year). It pops up for members every month and then is always available from a slider on the side of all pages:

On our roadmap for this year we're going to start looking at CSAT as well, one of the touchpoints will be after a best answer has been marked, this data will also feed into our deflection metrics.

LiselotteP

Thanks for these insights @sophielyons !

Component react.asset.createComment had an error.

Permission Problem You need the session.valid permission to do that.