So the reason for this is that they are not necessarily being assessed against the same groups.
Firm A could be taking your test and assessing it against GCSE level educated candidates while Firm B could be taking the same test results and assessing you against partners within a law firm. Being assessed against different norm groups could result in less than the 20th percentile in one and over the 80th percentile in the other.
So why do firms use different norm groups? Well, it depends on how they use the test. Some firms will have a "low" norm group as they may use the assessment for apprentices, interns and grads. Other firms may use the "higher" norm group because they are trying to recruit long-term talent and therefore want to compare you against long term talent within the firm.
There are a few other considerations why percentile rankings can be different. The key one is diversity concerns. Sometimes these assessments disproportionately filter out certain groups more so than others. Having a lower percentile benchmark can remove this though. So some firms wish they could put their benchmark up to help with the selection process, but know if they want to recruit diverse talent, they will be hampering their efforts if they do so.
