Problem/goal
Customer service level agreement metrics were being consistently missed due to CS staffing shortage and a myriad of other contributing factors. Clients were frustrated with call hold times and time-to-resolution metrics. It was such an issue that clients up for contract renewal leveraged it against the company for discounts. The poor experience also led some customers to reevaluate buying more from us and in some cases, even resulted in customers piloting our competition at a few properties so they could evaluate whether or not they wanted to phase us out and move forward with someone else.
My role
A small team was convened to brainstorm solutions. As the employee responsible for creating and managing self-help knowledge base articles, I was given the opportunity to perform a deep dive with these tools (we were moving from ZenDesk to Intercom at the time) to see how we could improve our ability for clients to self-serve, which would lead to a reduction in call volume.
Strategy/process
At the time, we were transitioning from ZenDesk to Intercom as our CS platform. I dove into Intercom analytics and identified a handful of articles with an lionshare of the views. I set up a meeting with CS leadership to find out if the topics covered in these self-help articles matched up with their most frequent call types and they mostly did.
Knowing this, it was clear that improved articles would better help clients resolve the issues on their own without the need to call a live CS rep. I proposed the plan to leadership and was given the approval to proceed.
From there, I met with tier 2 and tier 3 CS reps (experts in their field) to brainstorm ways to improve the specific articles. Using our collective subject matter expertise, we proceeded to modify each article with what we believed were improvements to the existing articles, then track the satisfaction metric over time. Note: Viewers of the self-help articles could score articles with a thumbs up and a thumbs down, which provided admins a report broken down by article that listed the satisfactory percentage. This metric could be tracked over time to inform us if the changes were making a positive difference or not.
Deliverables
As I no longer have the ability to view these reports and share visuals, there won’t be much from a visual standpoint for this example, though you can view the public-facing non-gated portions of their self-help article content at THIS LINK HERE.
Results
While the platform could not directly attribute improvement to the article satisfaction metric with reduced call volumes, we were able to measure a 10-35% increase in satisfaction score for the articles that were identified and improved. Ultimately, CS metrics did continue to improve to the point that they fell within acceptable service levels and the overall project was considered a resounding success.
