Automated scraping markup+CSS from a list of relevant URLs, using a variety of user-agent strings. Provides reporting on usage of CSS properties and apparent user-agent sniffing.
The systems underlying thresholds for what is a "good" versus a "bad" site should be reviewed - it is a percentage difference that it reports so it's a bit gray and not black and white. Can probably use std deviations here over an entire site scan to say something like this site is 21% different between mobile sites, which puts it in the lower 10th percentile of matching sites so it isn't so bad. However, this other site is 60% different which puts it in the 90th percentile of different sites and therefore it needs to be investigated.
The systems underlying thresholds for what is a "good" versus a "bad" site should be reviewed - it is a percentage difference that it reports so it's a bit gray and not black and white. Can probably use std deviations here over an entire site scan to say something like this site is 21% different between mobile sites, which puts it in the lower 10th percentile of matching sites so it isn't so bad. However, this other site is 60% different which puts it in the 90th percentile of different sites and therefore it needs to be investigated.