91–100 of 313 posts
YouTube is pretty much like this but it's the creators selecting the thumbnails. The most common thumbnail I see is a picture of someone looks like they are blown away with their hands on their head and their mouth wide open.
Would Netflix have to do this if it actually had content people wanted to watch? It's just sad watching it double down on this pulp because every other copyright holder took their ball and went home.
Netflix REALLY went for quantity over quality with their library. I'm certain that there are some great shows in there, but I have to start ten different series just to find one I like. On the other end of the spectrum, the Disney+ library is rather small, but I find at least half the shows to be enjoyable, so that's where I turn to first.
I've found a much better model is friends or critics' recommendations cross-checked against Rotten Tomatoes and IMDb ratings. (Which implies you need some people to use the scattergun approach, but I'd rather not be the guinea pig).
This way it doesn't matter if good Netflix shows are "needle in a haystack".
Would Netflix have to do this if it actually had content people wanted to watch? It's just sad watching it double down on this pulp because every other copyright holder took their ball and went home.
very well said. that line is spot-on.
for my taste / prefs, I could notice a clear state of uselessness around late 2019. I stopped netflix subscription starting Jan 2020.
I am not sure I understand what clickbait actually is. It seems to be a term a lot of people use, but in many different ways, and often in a 'I know it when I see it' sort of way. The author claims that basically any sexiness constitutes clickbait, that a hammer is clickbait, and wield is a gun is clickbait. I think only the last of those is plausible. Title images really give very little information, and usually sug…
It’s clickbait because it is simply picking the image most likely to elicit a reaction - and misrepresenting the nature of the movie in the process. I think the article did a decent job of explaining the psychology behind it.
Another simple example: imagine the title of this post was “Netflix is dying - here are three reasons why”. No actual connection to the content, just something to grab your attention. That’s clickbait.
I worked at Facebook for some time, and did a bunch of data work. We had this culture of building something, looking at the results between the experiment groups, and then choosing the statistically more successful one -- i.e. a newsfeed algorithm that had better engagement. This sounds great at first, and certainly is straightforward if you want a promotion. But behind the scenes some of us had this thought that our…
I think you can maintain the feedback loop without degrading the user experience if you actually ask them what they want to be shown. The problem lies in that "engagement" definition : if the user clicked on a thumb, then they must have been attracted to it. In reality, there is considerably more explanation for that click : They could be looking for a specific title, mistakenly thought _this_ actor was playing in th…
I would think a bigger threat to this approach would be being stuck in a local maxima.
I worked at Facebook for some time, and did a bunch of data work. We had this culture of building something, looking at the results between the experiment groups, and then choosing the statistically more successful one -- i.e. a newsfeed algorithm that had better engagement. This sounds great at first, and certainly is straightforward if you want a promotion. But behind the scenes some of us had this thought that our…
The directors would ask about things like normalization and if things were measured over a ncertain time period, but if you had some excuse or answer prepared, it was enough to get a nod and some recognition for your presentation. Heck, sometimes your team manager would do the presentation on your behalf to lay claim to some of that precious impact.
Then, after the new review cycle starts, priorities change and you never have to follow up on the holes in your data. Or, you can even admit it has holes, you already got credit.
Earlier quoted context omitted.
I do get a lot results that way on Netflix on failed searches… It doesn’t present only lexically similar results (“streetcar”) but thematically related films (classics, of the same era and related directors or cast) My main issue with Netflix is that where I live their movie catalog (esp. classics), is very short, seemingly in favor of Netflix’s own produced “shovelware”. Also, regarding a search for “the office”, ho…
There are a million nitpicks in the city, this is just one of them. Netflix and Amazon both have a strong interior belief this smart referral thing is good and who knows: A/B testing may even validate it. I just know that recommendations in books film and music creep me out, maybe it's an analogue of "uncanny valley" but being told "you may also like" kills the mood for me. Dammit, I want to like what I like, not wha…
I think most recommendations for video services is done through collaborative filtering and similar ML techniques, meaning you get recommended what other profiles similar to you have watched. It may be possible to infer a profile of "people who like the Office (UK) but will not like the Office (US)" but I imagine it may present challenges, especially if the intersection of people who like both is large.
> users consider each title for a whopping 1.8 seconds Users are most likely not “considering” them at all. We’re not machines going over one tile at a time and generating a score. Most of the time you’re looking for something specific and just trying to find it. This type of metric is the worst. I really hope this data-driven fad dies down and we start designing with human factors in mind again.
If by human factors you mean basic heuristics based on human intuition, that fails too. It might satisfy 50% of viewers but there's too much diversity in taste for this to be a good model.
I'm not defending the metric as described, but there's obviously good data driven approaches and bad ones. The trick is to find a solid proxy for the business problem and to try to weed out any dark side effects by applying it.
Earlier quoted context omitted.
> The data/engineering/product loops at tech companies favor boosting short-term metrics I think this hits the nail on the head. Leadership is hard, and executives can be many layers removed from the frontline. It seems that A/B tests and family have proven a really effective vehicle to communicate across leadership levels what has happened, why it happened, and the impact it had on the business. I'm definitely guilt…
How might a streaming service be designed to increase user happiness and wellbeing? From a positive design perspective, could certain streaming features give users a greater sense of purpose or goal accomplishment (e.g., mastering a genre) or help users meaningfully connect to others (e.g., seeing what friends have recommended lately)? While measuring the effects of a feature on wellbeing outcomes is substantially ha…
I have complicated feelings about the roles of business in our lives. I'm not entirely certain that's a metric I'm concerned with them optimizing on, nor do I believe we can even remotely quantifiably capture it in any meaningful with with human-computer interactions.
Earlier quoted context omitted.
This is the consequence of lack of vision. What is Netflix's long term goal? Ideally, it would be to fund and broadcast fantastic cinema. Yet, I really don't get the impression that contributing to the art form of film is very high up on the metrics chart.
I'm of the opposite opinion. They're executing their vision perfectly. Netflix is very aggressively pursuing the bottom line, and they're not doing it differently from a drug dealer. It's not random that the term "Netflix binge" exists. Netflix interface is designed to shove content at the users' face at all costs, and in a very disrespectful way. Here there are some disgusting dark patterns: - it took them a long ti…