Emotional Volatility Index
Emotional Volatility Index
Not all sentiment is created equal. A conversation that stays neutral the whole time isn’t the same as one that swings from joy to frustration and back again. The Emotional Volatility Index (EVI) captures that difference. It tracks how frequently and dramatically a customer’s mood shifts over the course of a conversation — offering a deeper look into trust, clarity, and conversational stability.
What Is It?
Emotional Volatility Index is a measure of emotional turbulence. It doesn’t just tell you how a customer felt — it tells you how often and how drastically their feelings changed.
Where traditional sentiment analysis stops at “positive,” “neutral,” or “negative,” EVI maps the emotional journey across time, then quantifies the magnitude and frequency of sentiment swings.
Basic Formula
Here’s a simplified way to compute EVI:
1. Capture sentiment scores at consistent time intervals (e.g., every 5 seconds).
2. Calculate the absolute difference between each consecutive score.
3. Sum those differences.
4. Normalize by duration or number of sentiment samples.
EVI = Σ |Sentiment(t) - Sentiment(t-1)| / Total Intervals
Higher values mean more volatility — more mood swings.
Example:
A call with sentiment scores [0.2, 0.5, 0.1, 0.4] would yield:
|0.2-0.5| + |0.5-0.1| + |0.1-0.4| = 0.3 + 0.4 + 0.3 = 1.0 total volatility.
Normalize it across the number of shifts (3), and the EVI is ~0.33.
You can also layer in emotion labels (e.g., from “Calm” to “Irritated”) to measure category-level jumps — such as from “Joy” to “Anger” — which often carry more weight than a slight dip in sentiment score.
Why It Matters
Customers don’t get emotionally volatile for no reason. If their emotional state is bouncing around, they’re reacting to inconsistencies — in answers, in tone, in experience. High volatility is almost always a proxy for broken trust or unmet expectations.
Key reasons volatility spikes:
- The rep is contradicting earlier statements.
- The customer isn’t being heard or understood.
- Process feels unpredictable or unfair.
- The resolution path keeps shifting.
What It Tells You
- Low volatility, negative sentiment: You failed them, but they expected it. That’s resignation.
- High volatility, mixed sentiment: You confused them, gave hope, took it away, gave it back. That’s chaos.
- Low volatility, positive sentiment: Trust. Stability. Flow.
A spike in EVI mid-call can point to a trigger moment — a misunderstanding, an unclear policy, a handoff that didn’t land. These are the moments that most shape memory.
How to Use It
- Identify call types or agents with high average EVI.
- Drill into call segments where volatility surges.
- Pair with CSAT drops or escalation markers — you’ll often see alignment.
- Design training or scripting changes to reduce unnecessary emotional swings.
EVI is especially powerful when paired with:
- Turn-by-turn sentiment trajectory
- Resolution confidence scores
- Agent emotion mirroring
Self-Check: Are You Tracking Turbulence?
Most dashboards don’t flag emotional instability — they just average out the sentiment. That’s like grading a rollercoaster by its height. The question is: can you spot the swings? Because customers feel them.
And if they feel like they’re on a ride they didn’t sign up for, you’ve already lost the trust.
Further Reading
- IBM Watson: Measuring Customer Sentiment Over Time
- Qualtrics XM Institute: Emotion in Customer Experience
- Harvard Business Review: “The Most Important Emotions in Customer Experience”