We use a database of millions of URLs that are relevant to education and filtering, including Alexa top sites to determine what sites to log/block/allow per the policy setup. We intentionally make an effort to not log sites that could be a nuisance to observe in the audit-report (e.g. advertisements, CDNs, embedded content urls, etc). An A/B test between a full-logging and selective-logging enabled audit report shows that at even a single-user level, the selective logging is much more usable. However, we are not perfect and sometimes miss sites that should really be logged. These are usually added to our cloud database as soon as they are reported. We are also in the process of including such sites more proactively, even before admins report them. Please provide your feedback to email@example.com as we help achieve the right balance between over logging and missing important information.