How are 'Similar Companies' identified?

For a selected company, EXPLORE and ANALYSE show the 5 most similar companies. What does this mean, how should they be used?

Our 'similar companies' feature has been re-developed, leveraging cutting-edge LLM-based methods to identifying similarity between companies based on their webtext. More detail can be found here.

This feature might be helpful for identifying a company's key competitors or even to gain insights into whether the company’s website is capturing the intended message and thus reaching the desired audience.

Our process evaluates the web text of the more than 1.6million companies in our database to assess each company’s likeness with each other, and where similarities are identified, up to 5 of the most similar companies for each are reported in our EXPLORE and ANALYSE platforms.

Brewdog_Similarity

Improvements to the final output are ongoing as we seek to achieve the very best results in this and future iterations.

You can currently find similar companies on company pages in the 'Similar Companies' data tab.

Please note: This methodology is not the same as our classification engine, which we use to build our RTICs. It does not generate a comprehensive list of a sector, instead it only identifies the companies which are most similar to the given company.