See emotions and spot brands across your video interviews

A participant says the new protein bar tastes fine, and her face at the first bite says something else. Across forty interview recordings, moments like that are easy to miss. So is the competitor's jar sitting next to your brand in someone's fridge. Finding them used to mean watching every hour of footage yourself.

Emotion and brand detection, how it works and what it changes, no sound needed.

How it works

• Three signals per moment. Conveo reads facial expression, tone of voice and the words in the transcript, and tags six emotions: happiness, surprise, sadness, anger, fear and disgust.

• Every cue links to the recording. In an interview's emotion view, filter by emotion, category and intensity, then jump to the second the cue happened.

• Patterns across the study. Question coding charts which emotions came up per question, with a coverage caption showing how many interviews had a usable reading.

• Brands, spoken or shown. Conveo flags when participants mention a brand or show a product on camera, competitors included.

• Ask for the moment you need. Emotions and brands feed Talk to your data, so you can ask for "a participant expressing disgust about the taste of this protein bar".

Research teams asked for it in four situations

These come from teams we work with in consumer goods, food and drink, media, home appliances, hospitality and insurance, and from the agencies that run research for them. We kept them anonymous.

• Ad and commercial tests. Teams wanted the moments that made viewers smile, laugh or lose interest, next to what they said about the ad afterwards.

• Concept and taste tests. Teams asked to catch the gap when someone calls a product fine and their face or voice suggests otherwise.

• Peaks across a whole study. Teams wanted to know which topics got people smiling and which emotions came up most, without rewatching every interview.

• Brands at home and on the shelf. Teams asked to see what sits in the fridge, the pantry or the shopping cart, competitors included, without asking participants to name every brand.

What this means for your research

The non-verbal side of an interview becomes something you can filter, chart and search next to the transcript. You go straight to the moments worth a closer look and can turn them into a highlight reel for stakeholders.

Some limits to know. These are automated readings of recorded signals. They can't tell you for certain how someone felt, so treat them as leads and replay the exchange around them. Readings need usable video: a participant with the camera off, or a typed answer, gives no facial or vocal reading, and a missing reading does not mean neutral. People also express emotion differently, so check coverage and sample size before comparing groups.

Read the emotional analysis guide, or see Talk to your data to search your interviews by emotion or brand.