Digital Twin

project image

“A digital twin is a digital replica of a living or non-living physical entity.”


A digital twin in marketing is a user that is similar to a user that has already been on your website.
This is done because you are already knowing what products or pages the user viewed and can offer a similar using the same experience.
But how is this done?


Porsche needed another way to reduce marketing cost and increase targeting.


Big Data aspects of a digital twin

First, we saved all the web events of our websites in a No-SQL database called Elasticsearch. Elasticsearch is a great application for text-based search and quick queries on huge databases using Lucene.

Graph search in Elasticsearch

Simplified, we used a graph search that is integrated into Elasticsearch (Link) to find users that are similar to the current user. Via REST-API calls with the matching Lucene queries, similar products from digital twins were extracted, and presented as recommendations to the current user.

Challenges using Elasticsearch for Digital Twins

A huge challenge in this scenario has been speed. Users usually leave a website if it loads longer than 250ms. This expands drastically, with 11% of the users leaving after one second loading time, and around 90% with more than four seconds of loading time (Source). Therefore we had to optimize the Elasticsearch cluster and partly simplify our queries to achieve a loading time of around 200ms. How? Leave a message below or contact me to find out more. is your partner for high quality, state-of-the-art solutions.