They say 16.6% of people who changed jobs last year used these services, but only 23.2% of companies report having any employees use them. If 16.6% is correct, the % of companies number should be much higher, supposing companies have multiple resignations per year.
The method for the first number on how they found their respondents is described just as "internet survey," with no further info. There are a lot of ways to do this that would over-sample people who use these services.