Product

Refining outlier detection for more accurate company data

You may have seen our work on outliers before. You can see it in our platform, as shown below. You can also read more about outliers here and the reason for identifying them here. 

 We are working on improving our outlier data further. To do this we’re providing our model with additional examples where company accounts are misreported, or are unlikely to be true.  

Over the next couple of releases, you may see the number of outliers increase or decrease. Despite the changes, outliers will continue to affect a small proportion of companies – we’re just finding additional examples to draw your attention to.   

We have specific goals in updating the outliers data. One of the assumptions that drives many of our estimates is that a company’s employees are full time employees. Updates to the outliers data will try and identify companies where they are likely to have part time, or agency-based, employment. Examples of these companies are more likely to be found in the hospitality or care sector, where employment contracts may be more flexible. 

If you have any questions about this process, feel free to get in touch.  

About the author