ComputerTalk

2025

Tripled monthly lead generation by redesigning the chat experience

Role

UX Designer

Timeline

12 weeks

Team

Developer, Product Manager, Product Marketing

Platform

Chat application

Role

UX Designer

Timeline

12 weeks

Team

Developer, Product Manager, Product Marketing

Platform

Chat application

Impact

3x increase in monthly sales leads after launch

3x increase in monthly sales leads after launch

More conversations routed correctly, cutting time spent on triage between sales & support

Increased opportunities to engage and nurture leads across every conversation

Overview

ComputerTalk's website chat was one of the company's main conversion tools.

It was the front door for prospects who wanted to talk to Sales, ask about upgrades, or hear about new releases. Yet it was the lowest performing part of the site. 

but… what were the rest doing?

I ran a competitive analysis by looking at Dixa, Zendesk, Intercom and mapped their journey against ours. Across all of them, their chat experience felt conversational: their chat experience opened with a greeting, not a form. They let users ask their questions first, figuring out routing after.

Taking a second look

There were three screens that stood between a visitor and their first message. By mapping their journey, I noticed friction points along the way.

step 1

Users had to fill a form to start

Name, email and company name were all required fields before a conversation could start.

step 2

Then users had to pick a department

They had to choose between Sales and Support before asking their questions.

step 3

Then finally, a chat in a window of its own

The chat had a separate window pop-up browser which was easy to lose behind tabs or close by accident. Here's where I also noticed the interface was crowded with rarely used text editing options.

The biggest issue was the entry experience

Naturally, people don't think "this is a Sales question." They think "I want to know what this costs" or "I need help with my account."

The previous process made users assume their questions were routed to the right department. I restructured the entry point around what they came to do instead.

BEFORE

System asks visitors to select a department.

FIRST REDESIGN

System asks what visitors came to do to instead.

The redesign here wasn't just the visuals and in fact, the more important work was on the content layer.

We couldn't eliminate routing entirely since sales and support were separate teams. But I could change how routing worked. So I changed the entry point, introducing a set of intent options mapped to what people came in asking for typically.

After the first round of tests, things changed!

The first version required the user to read all options sequentially and some users made errors while typing the numbers. So I changed it to buttons, that way, options are scannable at a glance instead of reading line by line.

FIRST VERSION - NUMBERS

Read the list, then type a number

REVISED - BUTTONS

Scan the options, tap one

Keeping it simple by embedding the chat to keep the conversation going

The old chat opened in a separate window that visitors lost behind the website. I advocated for an embedded iframe so that way conversations stays with them as they browse through the site.

BEFORE

Browsing the website hides the chat window, users forget its there.

AFTER

The conversation stays right where it is while browsing.

I also swapped the rich-text editor (though useful for agents but not visitors) for a plain text input and a stripped-down toolbar with their frequently used controls instead.

Reducing friction isn't just about removing steps, it's also about removing decisions.

Asking someone to pick "Sales or Support" may look simple for us, but I learned that I shouldn't assume they'd know how we were organized.

Going forward, I'd try to ask myself: what decisions does the user actually want to make and which are ones we should be handling for them that saves time and reduce errors.