TL;DR
I built an emotional-mapping system to help identify and prioritize the most emotionally severe friction points in our company’s Job Seeker experience.

Using a structured matrix to assess effort, benefit, and intensity — this method allowed us to go beyond volume and sentiment to understand how deeply customers were affected by their experience and what needed fixing first.

This framework was applied to over 36,000 pieces of feedback and helped inform prioritization across product, design, and go-to-market teams.

Imagine you are running a marathon:

Your first mile is great. Maybe there’s a bit of fabric rubbing. It’s nothing major. You ignore it.

A few miles later, the discomfort begins to chafe. Still tolerable. You persevere.

Eventually, it blisters. Now every step is pain. You push through, until suddenly—you can’t.

That’s how friction works in the product experience too. Some discomfort users will manage. But if it intensifies it will become a reason they stop using your product. It is critical to understand what friction in intolerable for your customers.

The Problem: Not All Friction Hurts the Same

One of the most valuable questions a product team can ask: What’s frustrating our users the most? I led an analysis to help answer this question for our Job Seeker experience.

A central feature of our company is our job board and helping professionals find their next opportunity. As part of this experience, we had a feature that showed job seekers a personalized list of jobs that they may be interested in based off of their profile and site interactions.

Alongside each recommendation, users could click a thumbs-up or thumbs-down button to provide feedback. When selecting thumbs down, they were given the option to include a brief open-text response detailing their feedback.

A trending decrease in product feature satisfaction triggered an in-depth review. Thousands of open-text feedback verbatims were received.

Product teams wanted to know:

• What aspects of the job recommendation experience was most frustrating?
• Were these issues driving disengagement with the overall site?
• How urgently did users need us to act?

The problem wasn’t collecting the feedback, it was interpreting it. The collected verbatims were brief and often emotionally charged. We needed a scalable way to determine not just what users were reacting to, but how severely their customer journey was being impacted.

The Solution: Quantifying Emotion from Verbatims

To solve this, I implemented a structured framework for interpreting emotional feedback. Each open-text responses was scored across two dimensions:

• Benefit: Did the user describe receiving high, neutral, or low benefit from the product experience?
• Effort: Did the user describe the product as easy, neutral, or difficult to use?

This created a 9-grid matrix that combined benefit and effort into core emotions we could use to describe reactions (e.g., frustrated, engaged, disinterested).

On top of this we layered on an intensity rating to measure how mild or severe that emotion was (e.g., frustrated -> angry -> infuriated).

To implement this, we used Python to call AI with a prompt to evaluate each individual verbatim for the following scores:

• Benefit score: +1 (high benefit), 0 (neutral), -1 (low benefit)
• Effort score: +1 (low effort), 0 (neutral), -1 (high effort)
• Intensity score: 0-3 based on emotional expression (profanity, tone, emphasis)

Each verbatim was processed on their own row with their corresponding scores. In parallel, we categorized each verbatim into a topical subject (e.g., job mismatch, pay concerns, wrong location, etc). When we combined these two analyses, we could see not only where friction was occurring, but how emotionally impactful each type of friction was to the user.

The Outcome: A Clearer Signal Across +36,000 Responses

The emotional scoring framework revealed patterns that traditional sentiment analysis missed:

• Premium members had more intense emotions when giving negative feedback–especially around perceived value. High-intensity responses correlated with lower future retention, flagging a risk to our subscription business.
  
• Scam job posts triggered the most intense reactions, with 4x the emotional intensity of any other feedback category. Though low in volume, these moments had an outsized impact on trust and the customer journey. This issue was prioritized because of this insight.
  
• Jobs in software, tech, and marketing saw higher emotional intensity due to vague descriptions and unclear remote/location details. This increased the perceived level of effort of users. By contrast, industries like healthcare and education with clearer role/location expectations saw lower intensity.

What Emotion Mapping Enabled

Volume tells you what’s broken. Emotional Severity tells you what’s breaking trust.

By scoring how customers feel—and how intensely they feel it—we uncover what experiences they find mildly annoying and what experiences may turn them away from the product entirely. It demonstrated that even short, in-product feedback can carry weight when you give it the structure to be understood.

By layering emotional severity on top of feedback themes, we gave product teams a clearer path to prioritization: not just what needed improvement, but where the urgency was the highest. That clarity helped teams move faster and focus on what mattered most to the people using the product.

Final Thoughts on Emotion Mapping

This methodology was highly applicable to in-product surveys as feedback gravitates around a specific product feature. Such an approach could also be used to better evaluate support interactions and chat dialogues as they often focused on single issue matters too.

Mild friction is tolerable. But severe friction isn’t. This framework helped us figure out the difference and identify where it was occurring.

Timothy Brown, CCXP