Netflix’s metrics, which are not verified by a third party and are not comparable to Nielsen’s average total audience measurement, previously counted an account that watched 70 percent of an episode or film as a viewer. Now, it will count anyone who “chose to watch and did watch at least 2 minutes” of a title as a viewer. Per Netflix, it is “long enough to indicate the choice was intentional” but more accurately shows popularity than the 70 percent threshold, which may have negatively impacted longer projects.
Netflix said that viewership numbers are now about 35 percent higher on average than under the previous measurement system. For example, the documentary Our Planet had 45 million member household viewers under the new measurement system, compared with 33 million under the previous metric.
Surprise: when you dramatically lower what counts as a viewer, you have more viewers.
It may turn out that this change more accurately captures viewing patterns on Netflix. Longer projects, as the said, will also benfit (but I wonder how they’ll count people watching something over multiple sittings?). And they’ll no doubt continue to track completion rates for their recommendation engine.
But changing what you consider a view – and dramatically inflating viewership numbers in the process – right after the successful launch of a new competitor doesn’t exactly seem like the move of a confident market leader.
Megan Mackay summed it up well on Twitter:
remember when facebook’s video metrics were set up to be so badly skewed that an entire industry pivoted to prioritize facebook video, only to have the floor collapse underneath them because their strategies were based off of bad numbers? anyway, everything looks fine here
My thoughts exactly.