case study - call center

Problem

Too many feature requests, not enough relevance.

Call center management submitted hundreds of feature requests for call representatives' (reps) systems. For instance, "add a widget that lists which articles callers have read on our website." Overwhelmed with prioritizing these requests into jobs, internal product teams asked me to find out what call reps actually needed. What were some typical call rep experiences? What pain points could be eliminated? What meaningful opportunities were there to better meet call rep needs?

Approach

My role: lead qualitative user researcher

I conducted a user-centered investigation to help build knowledge of call reps and their experiences, so that I could generate actionable recommendations about how to improve their systems. Activities included 1. observing call reps 1 on 1 while they worked, 2. gathering written reflections on typical and atypical calls that I couldn't observe directly, 3. reviewing observations and diary entries with call reps (to make sure participants didn't feel misinterpreted), 4. collecting workplace artifacts, 5. digging through call center systems and interfaces, 6. running usability tests for a few call rep and caller tasks, 7. reading journal articles about call center operations, and 8. interviewing call reps, call center managers, and call center subject matter experts.

COVID-19 forced studies to be remote

Observations and interviews happened remotely due to COVID-19. Normally I go on site for contextual inquiries and interviews, but call reps had to work from home. And I couldn't visit them in person. So I observed via zoom calls, 360 degree video shots of their workspaces, and photos of their workstations. Thankfully, Hagerty set call reps up for remote work months prior. This all just meant the study took longer than it would have taken on location.

An image of a zoom call recording during a call rep observation session.

2 other UX specialists provided support

After I collected data and notes, I recruited 2 other specialists on my UX team for help. Together we reviewed my observations and created an affinity diagram to synthesize them. This reduced researcher bias and uncovered salient themes.

An image of a portion of the affinity map for this project.

I provided actionable research assets

I created experience maps for various types of calls that call reps received. Along with the affinity diagram, these allowed me to uncover insights into user needs and pain points. They helped frame user experience narratives to facilitate empathy. And I used them to pinpoint effective, actionable improvements product teams could make for call reps.

After consulting with my team and developers to understand the feasibility of preliminary recommendations, I revised some recommendations to be more feasible. (E.g., many of the changes had to be made in a Salesforce dashboard. Conversely, we had an in-house policy management application that housed features that could not be migrated to Salesforce dashboards.) Then I presented my findings to product owners and call center leadership.

An image of a portion of the affinity map for this project.

An image of one of the journey maps for this project.

Overall impact

Reduced stress, increased efficiency

Not only did I find addressable pain points in technical systems, I uncovered perverse incentives and cultural issues. I helped the call center and internal product teams understand call rep experiences. And, I was able to recommend several things that both could enhance user experiences while increasing sales. My research resulted in relatively short-term low-hanging-fruit-fixes, and important work planned out for the long term.

Highlights

Poorly leveraged single-sign-on
Call reps' suite of tools is a Franken-system with at least 21 applications and 16 logins (and each call rep spends roughly 10-20 minutes a day logging in or opening things). I recommended getting each of the call reps' internal systems onto an existing single-sign-on to reduce the amount of times reps have to manually log in. This is expected to save roughly 5-10 minutes a day for each call rep on duty. That's enough time for them to take a call or write better call notes. That also means some callers won't have to be on hold as long. And most importantly, this saves call reps 5-10 minutes of frustration a day.
Call agents have over 21 (that I found) systems and user interfaces they might have to juggle during a single call.
Cluttered interfaces
Call reps' dashboards are mired with redundancy. I recommended getting rid of useless information on call reps' dashboards to reduce cognitive load. E.g., the same call notes were located in 4 different places on the same dashboard! This recommendation eliminates roughly 1/3 of the crap in call reps' views, allowing for more focus and better problem solving.
An image of an audited call reps customer account dashboard screen. Highlighted areas represent duplicate information (almost everything is highlighted lol). Areas covered with an 'x' indicate areas that go unused by call reps.
Preventable form errors caused staff stress and wasted time
Sometimes input errors sent call reps into stress spirals, or they were forced to spend several minutes troubleshooting while speaking with a caller. I recommended an audit of call reps' forms for preventable input errors. E.g., several input fields allowed entries that caused validation errors, such as a ZIP field that allows space characters.
Can you tell what's invalid about this ZIP code? Neither can the call rep who entered it. They spent almost 5 minutes trying to figure this out and had to put the caller on hold. Then they gave up and submitted a service ticket to IT. (Hint: the input field allows space characters, but doesn't like them.)
Client self-service impacts caller experience.
A lot of calls are about client-facing website issues, because client-facing interfaces for account management are almost entirely unusable. They aren't even mobile responsive (even though 60% of visits are on mobile). Policy management dashboards haven't been meaningfully optimized for over 15 years (or, over 105 internet years, lol, as one of my research colleagues would put it). I recommended making self-service flows for clients mobile friendly. This means a style overhaul, some information architecture updates, and a redesign of flows. It's going to be a years-long project to bring it up to par, but it's necessary. Not just for reducing calls, but to improve things for our customers (this will also increase customer retention).