Skip to content
Back to Headspace
  • There are no suggestions because the search field is empty.

How are Digital Twins created and updated?

Learn how digital twins are trained, validated, and updated to provide accurate, up-to-date insights for targeted audiences.

Training the Digital Twins

The collected consumer data is used to train the AI models (digital twins) alongside data from a comprehensive knowledge graph. You can find more information on our knowledge graph in another article:

'What is a Knowledge graph?'

The combination of consumer insights and scientific data, makes the digital twins highly accurate and uniquely tailored to specific audiences.

 

Ensuring Accuracy

Liking scores predicted by digital twins are compared with actual consumer panel scores, showing a strong correlation of 75% to 95% and an error margin below 6%.

 

Regular Updates for Relevance

Both digital twins and the knowledge graph are regularly updated to reflect changes in consumer behavior and product formulations:

  • Digital twins: Updated every 6 months using fresh data from social listening, recipes, and surveys.
  • Product data: Reviewed every 6-12 months or sooner if triggered by new formulations or recipes.