Comment by confused_boner
7 months ago
could that be solved by finding and using the 'most replayed' mark timestamps for each video and using that to find the best start points for each video
7 months ago
could that be solved by finding and using the 'most replayed' mark timestamps for each video and using that to find the best start points for each video
Depends on the content and how it tends to be consumed. That might work for YouTube but maybe not, say, Netflix or services that stream movies.
We did a mix of human curation and random selection, noting which random selections were successful and which were not. It was effective enough. The thing about an experience like this is that it doesn't need to be perfect. Like channel surfing, if it doesn't catch you then just switch the channel.
Uhm, only if this mark doesn't get affect by it being streamed here or it was be less relevant later.