Hacker News Reader

Comment by fredrikholm | Hacker News Reader

This general trend of deteriorating UX in widespread services is extremely exhausting and depressing.

The accuracy you could get with Google searches in 2005-2010-ish was amazing. Knowing how to Google was a secret weapon of computer literacy, you could find anything. Now I save links on a small server (I work on many machines) unless something is posted here on HN, as it's one of increasingly few places where I can search and actually find things.

Same story with YouTube, searching for a video there has become an exercise in self flagellation.

I don't know if it was Netflix that started the ball on nonsensical, avant garde categories followed by thousands of auto-playing calls to action, but combined with intrusive ads and obfuscation the day-to-day experience on mainstream internet really sucks the joy out of whatever amazing content that lurks in the cracks of what is actually a set of amazingly marvelous technology.

View on HN

I’ve been using Kagi ($10/month user-oriented search engine) for a few months now and it’s what finally scratched my Googling itch. The quality of results is almost like Google a few years ago, plus you can buff, nerf or even block certain sites! Goodbye Quora, Pinterest, Yahoo Answers and other crap, hello Wikipedia and StackOverflow results always at the top. And of course, no ads.

> Same story with YouTube, searching for a video there has become an exercise in self flagellation.

This so much, I don't know what's wrong with Youtube search, but on desktop it gives me like 10/20 genuine results then a block of suggestions and no other results.

I came to the conclusion that some PO and data guy figured out that if instead of me seeing what I want and leaving they can hook me on some Youtube journey they'll make more money.

> instead of me seeing what I want and leaving they can hook me on some Youtube journey they'll make more money

This sounds so stupid, but it makes a lot of sense and explains so many issues. Why make you click on the video X that you were looking for instead of sending you on a 1 hour rabbit-hole, bonus points if at the end you even had forgotten what you were looking for.

I firmly believe (as in assuming, no proof) that it was both a cost saving measure and what you describe as increasing revenue - think about it in terms of statistics, if you give the user the most relevant videos as the top 5 suggestions, you can cut down the cost of a query substantially and so cheaply so that you spend micro-cents per query, whereas before you'd arbitrarily cut off search results related to the user query after ~50 pages or potentially more, costing maybe thousanths-cents per query

The new scenario gets you the user conversion to a watched video and subsequently to a relevant ad, all the way to the 95th percentile, and maybe even 99th percentile conversion, at a fraction of the cost, and doubled with the fact that the extremely relevant result got the user to a video, got more monetization as a result. Your cost is then simply the query time to the cheap edge model, and then the search that provides the most relevant videos to search top-n (in this case 5)

It's not working for some people - but who cares, you hit 99th percentile of all users and you've cut down the spending of a black hole that YouTube is, in fact you might've even made the experience better for 99% of users. I have nothing to substantiate my claims, but knowing how I'd reason about it, this would be my assumption for the state YouTube search is in

I also hate this and it has made me move away from YouTube search, unfortunately YouTube recommended is quite good at giving me relevant videos to myself

> Knowing how to Google was a secret weapon of computer literacy

I think it is still a matter of knowing the right techniques. You’ll hear people recommend features like `site:` and such (which are great), but really it’s a matter of learning and adapting to how to phrase queries for Google. It’s still a skill.

The default experience is far worse than it used to be. A common situation is I'll search for a 3-4 word phrase and the top results contain one of those words. If I'm really lucky Google will have attempted to match some terrible synonym of the other words, but frequently it doesn't even do that!

If you put everything in quotes, it's better, but there's so much SEO spam in the first page of results. I can't remember the last time I found anything really interesting with Google, for software development it's basically just a mediocre way to search Stack Overflow, GitHub, and MSDN. It's depressing.

Most of the old tricks don't work anymore. Instead of using operators and syntax to refine results, now one needs to rephrase their question, which is far more subjective and imprecise. It's no longer tailored for the searcher.

Google Search is completely dysfunctional now. I use Github own search (as bad as it is) to search for code. Searching for an exact code text (with quotes) yields no result despite having multiple results in Github; and these repositories are public.

The old adage that if your not paying, your definitely not the customer still holds true. If you are paying, but at an affordable consumer rate, your likely still not entirely the main customer, especially for anything as complex as search over the entire internet.

I'm gonna go out and assert that the statement "if you're not paying, you're the product" has never been true on a large scale. Rather, the two are unrelated. In this day and age, you are always the product, whether you're paying or not. And back in the 2000s when I first started hearing that adage, it was only true on the most transparently awful platforms. It was really just a thought-terminating cliche meant to shush anyone who bothered to complain about a free website doing something bad. The difference between then and now is that we now know how to monetize somebody's data in a way that solo startup founders in the early 2000s didn't. Sure the science probably existed, but it wasn't obvious to somebody making a website that that was how you make money.

Nowadays, the process of converting daily active users into monetizable data is trivial, so that's what everybody does, again regardless of whether or not you paid. When you think about the economics of it, your value as a commodity does not decrease just because you paid for something. No matter how much you pay for your subscription, your data is always valuable, and the only reason Netflix doesn't force ads on all their subscribers is because it'd look bad for them. But behind the scenes, do you really think they're not selling your interaction data just because you subscribe to the most expensive package? Monetizing users has always been a function of ease of implementation, not the out-of-pocket expense of the user.

Okay?

I think this was implicit in my second sentence: "If you are paying, but at an affordable consumer rate, your likely still not entirely the main customer, especially for anything as complex as search over the entire internet."

I didn't mean to contradict you, just elaborating on it a little more and taking it in a slightly different direction. I think even with your caveat, there's still an implicit belief that there's still a relationship, however distorted, between what the user pays and what the seller does with your data, and my point is that the incentives just don't seem to work that way.

> if your not paying, your definitely not the customer

This is correct enough. "p → q" does not imply "~p → ~q". In other words, if you are not paying, you are the product. If you are paying, you may or may not be the product. The adage does not say anything about this scenario.

Reply on news.ycombinator.com