Synthetic scoring in Thematic

Apart from themes and sentiment, Thematic can generate a number of custom synthetic scores and categories from unstructured data. Thematic derives this synthetic metadata by interpreting text and capturing the nuance. This helps companies easily measure specific metrics that matter to them, without having to explicitly ask people.

1. What are synthetic scores and why do they matter?

When you're working with feedback, it’s easy to focus on simple numbers like NPS or CSAT. But they have several limitations:

  • You need to ask people to provide a rating (not everyone does)
  • They reflect a mix of different things and figuring drivers is difficult
  • They aren't useful in day-to-day reporting on success of specific initiatives

Synthetic scores solve these problem allowing you to calculate specific metrics from conversations, reviews, and complaints directly.

2. Examples of synthetic scores and metadata Thematic

Thematic helps companies fully customize metrics they are trying to extract from unstructured text.

This eliminates the need for asking people to rate something by sending them a survey. This also helps to derive from reviews, conversations and responses more granular metrics.


Here are some examples:


⭐ NPS (Net Promoter Score)

Predict how likely a customer is to recommend your company, based on their conversation or review. The score is expressed on the familiar 0–10 scale — no survey required.


⚙️ CES (Customer Effort Score)

Estimate how much effort a customer had to put into resolving their issue, using a 7-point scale. Lower scores suggest smoother support experiences.


😊 CSAT (Customer Satisfaction Score)

Turn open-ended feedback into a familiar 5-point satisfaction rating, helping you track how happy customers are without needing them to fill out a survey.


😤 Customer Frustration Score

Measure emotional friction in an interaction on a scale from +10 (delighted) to –10 (frustrated) — perfect for spotting pain points in real time.


🔁 Repeated Follow-Up Detection

Classify conversations as "First Contact," "Repeated Follow-Up," or "Unclear" — helping you track cases that needed more than one attempt to resolve.


📱 App Reliability Score

Rate how reliable your product or app seems to the customer, based on review text. A targeted way to separate technical feedback from general sentiment.


🚚 Delivery Experience Score

Focus only on the delivery-related part of customer feedback (ignoring unrelated comments like food quality or service) for precise monitoring.


😡 Rudeness Detection

Rate the perceived rudeness in customer service conversations, helping flag negative interactions for review or follow-up.


🎬 Entertainment Value Score

When customers react to content on platforms like social media, this score tells you how entertaining or engaging they found it.


✨ ...and many more!


Synthetic scores can be fully customized to reflect the specific experiences you care about — whether that’s product usability, service friendliness, pricing perception, or something entirely unique.


👉 Combined with Thematic themes, you can not only track these scores but also uncover what’s driving them, giving you both the "what" and the "why" behind customer experience.


3. How synthetic scores can be used in Thematic

Synthetic scores and metadata provide additional dimension in Thematic. This means that they can be used as filters, within impact charts and pivot tables. They can also be pushed in real-time into data warehousing and trigger various alerts or workflows.

Here's an example of how you can see which themes drive a synthetic NPS from chat spport:

Here's an example of evidence that a when people mention "Update requests" in their feedback, they often have to follow up repeatedly. It's an example of gap in the process that needs to be addressed.