Have a good read!

Halfway there!

Thanks for reading!

Redesigning a CRM (Customer Relationship Management) experience to improve customer service agents' workflows and transform data into a revenue-generating engine.

Saas
CRM
September - November 2020

My Role

Project Type

Results

In development
70% CSAT increase during Testing

Research (Remote Ethnography)
Redesign Ticket Log
Redesign Ticket History
Data Visualisation
Improving Design System
Prototype
Test

Before After

Gorgias is a B2B CRM solution for e-commerce customer service agents who interact with their customers to provide support.

I was originally asked to redesign Gorgias' Ticket History to improve agents' workflow to serve their customers better.

This is how I helped agents understand their customers from a glance and turned a CRM platform into a revenue-generating engine.

Ticket History

What is a ticket?
Tickets are communications and interactions between a customer and a customer service agent; or between agents.

Background

Click image to enlarge

What is Ticket History?
Ticket History is the archive where all past tickets from a customer, 'open' or 'closed', are stored.


Here, agents can inspect, open/close tickets, add tags, update ticket titles and mention other teammates.

Click image to enlarge

Why is it important for agents?
Ticket History gives agents background information on the interaction with customers. This helps them understand and resolve the issues accordingly, consequently improving the customer experience.

Research

Doing research with YouTube videos and online reviews.

Constraints: despite being used every day by over 1000 online businesses, there wasn't a lot of data and research available on user behaviours, pain-points and needs.

Solution: I looked for real use-cases, pain-points and user needs in YouTube online reviews videos and Shopify reviews. From here, I was able to see the complete picture of how people use the Ticket History.

User Stories
Next, I created the two personas who rely the most on the Ticket History: the customer service agent Annie and Manager Mark, drew up their User Stories to distill the Core User Needs.

Heuristics Evaluation & Cognitive Walkthrough
I also ran Heuristics Evaluation with the design team and Cognitive Walkthrough with the agents on the current User Flows to shed light on usability heuristics issues and user flow problems.

User Stories, Heuristics Evaluation & Cognitive Walkthrough

Core User Needs

Agents Annie:

  • To know when to close the tickets and move on to other issues.

Owners/Managers Mark:
Understand the customers better
Having a global understanding of customers' behaviours and recurring issues to effectively plan future interactions.

Pain-points

User Experience

  • Doesn't show global overview of customer activity.

  • Doesn't allow quick understanding of the issues.

    • What is the conversation about?

    • Which actions have been taken? (Refund, Return, etc.)

    • Has it been resolved? How was it resolved? How long did it take to solve?

    • Which orders have been affected?

  • Doesn't allow deeper understanding of customer's behaviours

    • How often does the customer interacts with support?

    • Has the customer asked that question multiple times?

    • What is the customers behavior?

Interface

  • No hierarchy of tickets.

  • Convoluted, overloading interface, navigation -> easy to get lost

  • Vertical line connecting tickets confuse agents that tickets are related.

  • Ticket design appears outdated and impersonal.

Design

Design Goals

For Agent Annie:

  • Giving them the power to know if they should leave the tickets open or close it.

For owners/Managers Mark:

  • Helping them understand individual issue quickly and efficiently at scale.

  • Helping them understand customers' behavioural patterns and recurring issues.

"Learn from the past, plan for the future." Audrey Farrell

Helping agents close more tickets with Ticket Log.

The Ghost Tickets:
Due to the nature of delayed communications with customers via emails (70% of total tickets), agents work on multiple conversations at once, taking actions on those conversations, move on to other issues, then come back to resolve the previous ones. This creates a loop of interactions that is prone to human errors:

As a result, memory lapse often occurs, resulting in tickets remaining open even when the issue has been resolved, these are The Ghost Tickets. These Ghost Tickets dilute the Ticket History, creates confusion about task priority, affecting the agents' workflow and the business's performance.

💡

My hypothesis: Once an agent revisits an open ticket and sees the history of actions taken, they should be able to determine whether a ticket should be closed or not.

Ticket Log

Like on a ship, the captain keeps the record of the ship's journey is his Captain Log. The Captain Log provides insights into the cause and effects in case of an incident.

The Ticket Log shows the record of activities regarding a ticket to help agents understand the course of actions taken.

Before After

Improvements

User Experience

  • Record of activities: shows what have been done to the tickets, helps agents decide to close the ticket or to leave it open.

  • Bread-crumb navigation helps with orientation within a complex system, diminishing the sense of loss.

  • Tickets are designed to mimic email or chat bubbles, giving it a more personal touch.

Interface

  • Ticket Log CTA is given a text label to help increase discoverability and learnability.

  • Main CTA is placed on a separate place reserved for primary actions.

Results

90% of agents during tests found the Ticket Log to be extremely helpful to see what have been done to the tickets. This gives them the confidence to close the tickets or leave it open and take the next steps. (report to manager, ask for help from teammates, etc.)

With the design of tickets mimicking chat bubbles, agents also felt like responding to their friends or acquaintances, offering a warmer, more authentic interaction with the customers.

This also helped onboard new agents/merchants migrating over from other communication platforms (Gmail, FB Messenger) by following the users' mental model.

Helping agents and managers understand customer's issues and behaviours individually and at scale using tags.

💡

The Ticket History is currently showing tickets in chronological order.

My hypothesis: if tickets were to be organised based on content, in addition to time, that should be able to help agents sort through and understand them faster.

Ticket History

Amongst several options for new Ticket History design, the tags system emerged as a winner due to:

  • The tags provide a snapshot of the content of the ticket without delving into it. This help agents understand the tickets at a glance, hence drastically enhances their workflow.

  • The tags already in used on the Main App screen. Adapting it here would promote system coherence and learnability.

Before After
  • With tags, agents are now able to understand the conversation faster without having to click on individual tickets.

  • tags also help managers identify recurring issues, hence plan a course of actions accordingly.

Improvements

  • Order numbers, order status and action duration are included in each ticket to give a point of reference for the ticket.

  • Declutter information, visually separating open/closed tickets, prioritising ongoing tasks.

Results

9/10 of users during tests found the new Ticket History to be very concise and informative. Contextual information (Order Number, Order Status, Action Duration, etc.) allow them to better understand the past tickets from a customer.

The new Ticket History also helps them draw a global picture of a customer's interactions with the brand through tags

Seeing how the new design helps agents and managers understand their customers' issues and behaviours better, I took a step further and explore how could we help them turn that understanding into a profit-making asset.

Converting data into a revenue-generating engine

Since CRM solutions like Gorgias processes a copious amount of data, I wanted to maximise the potential of these datapoints by linking them through data visualisation.

Exploration

💡

Hypothesis: using 'Time' as a measurement of data – If users could see on which day of the week, month, year, the customers interact with the brand, maybe he could see a pattern and plan for these interactions.

Hence the Calendar View for the Ticket History.

Results

That's why I decided to go with a dashboard design that showcased the summarised Ticket History with two reference points: Time and Content. I also included the interactive tags that customise the charts based on the nature of the tickets. This would enable Mark to make the connections between the time the ticket was created and its content to effectively predict the customer's behaviours and issues.

During Testing, the users found that whilst the Calendar View was a nice addition, it was inadequate to surmise the insights from the interactions with the customers without supporting data (orders, etc.).

Also from a Data Visualisation viewpoint, the calendar shows current events and their contents but not their trends.

In the screen below, Mark can see that the amount of communications with the customer increases in December and that they are usually about shipping. He should then be able to preempt shipping issues before the next holidays season.

Realising the potentials of connecting data, I went a step further and proposed an overview that links Ticket History and Order Summary.

This, I believed would help businesses see the connection between communication (Tickets) and sales (Orders), thus plan their customer relationship strategy not only as a reactive measure but one for growth and adds high-value to the end customers.

Would all this dehumanise the interactions between agents and customers, which is crucial to customer retention?

To answer the question posed by various stakeholders, I conducted Usability Testings to validate my theories but also helped me understand that:

  • Agents are proud of their work. For a dedicated customer service agent/owner/manager, the features offer them additional help that allow them understand the behaviours of the customers better, hence devise a strategy to serve them better.

  • The features reduce the risk of reading a nasty email. This would help them focus on the issue and not the negative feelings that come with it.

  • More importantly, understanding individual issues of customers help agents empathise with them and connect on a more personal level, which is in the end would be most beneficial to build long-term relationships.

Results

However seeing the complexity in development, the Order/Ticket Summary dashboard was put in the backlog for later iterations.

The dashboards were highly appreciated during the Testing phase. The users reported having a much highly connected global view of the customers' interactions with the brand. The dashboards served as an intelligence centre that allows businesses to make forecasts and turn data into revenue-generating assets.

Gorgias had already been an established CRM solution with thousands of active agents every day.

With a complete overhaul of a big part of the design, we planned to deploy the features with an incremental roadmap and additional Onboarding

Onboarding

These major changes will require additional onboarding to bring existing agents up to speed with the new tools.

Tooltips and popovers are handy to familiarise the agents with new features without cluttering the busy workspace.

We also created a series of Newsletter and Video Tutorials on YouTube to onboard new users and teach existing ones new features, thus complete the cycle.

Tooltips, Newsletters & YouTube video tutorials

Takeaways

Learning:
Look outside of the box to find data to support design decisions.

There are many ways to look for information during the Research Phase. The job of the designer is also to find solutions to have real understanding of the users' needs and pain–points.

By tapping into authentic Youtube review videos and online comments, I was able to empathise with the users in a way that gave us fresh perspective on our users - the agents always lack of time, the managers/owners desperate for the intelligence system - to design solutions that help them understand and serve their customers better.