Customer Effort Score (CES) - Scoping an In-house Solution
As mentioned in #5923 (closed), we are looking to change the way in which we survey customers who work with us in tickets. We want to move to using Customer Effort Score (CES) instead of Support Satisfaction (SSAT). Through our discussions weighing up the pros and cons of using an in-house solution, or using a 3rd party tool, we have ultimately made the decision to begin working on building out our own CES survey tool.
This issue is intended to help us define the scope of the solution, ensuring we accurately capture the requirements for the survey tool before Ops begins building.
The current concept is fairly basic - we send out a link to a HTML form (the content of which will be fully customizable by Support Ops) via a Zendesk automation (with content similar to our existing SSAT automation), asking customers to submit their feedback. Instead of using links that refer customers to Zendesk CSAT survey, the links will direct users to our custom built CES form. I've outlined some of the limitations we'll likely need to work within in my comment here given that this solution will be reliant on Zendesk + Our custom form. Ultimately, we expect the entire experience to be quite similar to our mid-ticket feedback form.
This issue is intended to clearly define the following:
- The conditions under which the survey should be sent (including timing, type of customer, type of ticket)
- The content of the survey
- Details of how the responses should be stored beyond Zendesk (if there is any requirement to do so)
- Reporting requirements (will we just want to report on this in the same way we measure SSAT?)
- Details of any integrations that may be required, and what the expectation is as far as functionality goes (e.g. Slack notifications/GitLab issue creation on negative response/etc.)
- Details of how this is expected to be rolled out - do we want this rolled out to all customers, all at once? Do we want to stagger this rollout (e.g. only >1M ARR customers first)?
I've created an outline of the requirements in the comments section of this issue with some of my own assumptions, but the purpose of this issue is to flesh these out in more detail and provide feedback so that we have a clear idea of what the solution will ultimately look like.