Last month, yet another YouTube controversy erupted when several internal metrics, including a content “throttling” metric, were leaked via its codebase. At the time, YouTube didn’t address the controversy and instead quietly scrubbed these metrics from its code.
Now, more than two weeks after the controversy, YouTube has issued a partial response and given an update on P-Scores – one of the metrics that was leaked via its codebase. However, as with a lot of YouTube’s communications to creators, the update is vague and alludes to why the P-Score was removed but provides little in the way of specifics.
Before the P-Score was discovered in YouTube’s code, it was believed to be a proprietary internal metric that YouTube uses to surface the most “engaging and brand-appropriate content” for Google Preferred – YouTube’s brand-safe advertising category. YouTube had listed five broad signals that impact the P-Score but has provided little information on how YouTubers can raise their P-Scores.
The discovery of the P-Score in YouTube’s code and subsequent research based on this data revealed that:
- The P-Score gets applied to every video on YouTube, regardless of whether it’s in Google Preferred, which suggests that the P-Score is used to rate how marketable every video on the platform is to prospective advertisers
- The P-Score is region-based and YouTube channels generally have a higher P-Score in their home countries
- The P-Score is influenced by the age rating YouTube applies to videos internally
- Corporate brands, Hollywood stars, and YouTube’s “authoritative sources” (legacy media outlets that YouTube has deemed to be trustworthy) have the highest P-Scores
In an update on P-Scores, YouTube’s director of Project Management Tom Leung, alluded to the metric being removed from the codebase because YouTube only wants to share data externally when that data is useful for creators.
Leung claimed that the P-Score is “not very useful to creators” because it’s a “relative score.” According to Leung, this means “there’s not much you can do about it, it’s not very actionable,” and “it doesn’t even explain your Google Preferred eligibility very well.”
Leung added that YouTube wants to make sure external data isn’t abused by bad actors and used to affect advertisers negatively because this may result in the advertisers taking action that impacts everyone on YouTube.
“So we’ve got to be really careful about only providing stuff that’s useful and that is unlikely to be abused,” Leung said.
Leung advised creators to focus on YouTube’s advertiser-friendly guidelines instead and said: “If you meet the advertiser-friendly guidelines in letter and spirit, that is your best chance of having maximum monetization including eventually eligibility for Google Preferred if the conditions warrant it.”
While Leung’s update provides a vague explanation of why the P-Score was removed, it leaves a lot of questions unanswered. For example, if there isn't much creators can do about their P-Score, how is it used and what does it actually mean?
Leung’s update also doesn’t contain any reference to the other internal metrics that were leaked via the codebase – content label rating and content “throttling” rating. For now, creators are still being left in the dark about what these metrics mean and how they affect video performance on YouTube.