top of page

Hearing doesn't scale.

If it did, people would love employee engagement surveys, and CFO’s would base the forecast on the social listening report or the brand tracker.  

Lots of charts, little actionable learning: Why?

Most quant research, NLP, and text analytics take the views people express, then chop them into bits that have little relevance to how people think. The result is the illusion of empirical rigor, and 'insight' that seldom correlates with real-world outcomes. If you notice any of the approaches below, you have a hearing problem

Sentiment Trap

Keyword Trap

Sentiment is sometimes helpful, but by itself, sentiment is hard to action     without the 'why'.


People don't think in keywords, yet most tools run based on keywords or word vectors - which miss experiential drivers.

Screen Shot 2021-01-29 at 4_edited.jpg

Topic model trap

Topic models capture one narrow aspect of what people say. They can't measure the wider context. 


Complexity Trap

You don't have time for  ‘paralysis by

analysis’ decks or for a new data

science initiaitve.


One-size-fits-all trap

Force-fitting the views people express

into standard categories or models? 

Expect blindspots!


Mismatch Trap

Tracking studies seldom correlate with performance, and research findings are often hard to action.


By leveraging the power of large language models, narrative analytics captures the practical and experiential drivers that other approaches miss


Phrasia is easy to use - even for non-technical people.  It works in up to 60 languages in a single analysis - making multi-market analysis dramatically clearer and more actionable.

Contact us
bottom of page