Live data from Hacker News

Bay Area woman is on a crusade to prove Yelp reviews can’t be trusted

sfgate.com

191–200 of 331 posts

Re: Bay Area woman is on a crusade to prove Yelp reviews can’t be trusted

#191
post #13

Good for her. Yelp's business model seems to work best if they can pick winners, which is unfortunate. If I search for "dinner" on Yelp for "current location" and look at the map, it leaves out an astonishing number of places that I can see within a mile if I drive down the street. And if your business model is centered around hiding all but the winners - who might pay for placement + reviews both - then you have to…

> It's a broader internet problem in general - "trending" algorithmic reinforcement stuff will push everyone towards the ONE TRUE BEST HOTTEST MOST POPULAR bbq/ramen/whatver place, when probably there's half a dozen other options within 10% of taste and quality, that some people very well could legitimately like better if they heard about them too. Instagram doesn't help -- gotta get a shot of me eating noodles at th…

Back in my bar-tending days we always told the door staff (bouncers) to slow down when checking IDs early in the evening, just to get a line going. It works surprisingly well for drawing in more customers.

Totally tangential but follow-ups on that psychology:

* some people would walk in and see the still mostly-empty bar and get fed up that they waited in line for "nothing". This was surprisingly few people though.

* some people would ask and buy just about any explanation - "oh that guy is new and still learning", "we had someone get in on a fake ID and got a fine the other day so the guys are just being careful" or whatever and the people would be glad they got in early to get a good spot.

* sometimes we'd slow down the bar staff a bit too so that there was a line to buy drinks and people were happier that they had to wait in line again - confirming their bias that they must have gotten in at the start of the rush.

* some people were wise to the scam and everyone who called us on it thought it was great. They often offered to have their buddies come through the line a few times in exchange for a free drink.

* for some reason when you let your buddies, regulars and off-shift staff cut the line and get in quick, the rest of the folks seemed more willing to wait in the line.

tl; dr - group psychology of humans is wierd.

Re: Bay Area woman is on a crusade to prove Yelp reviews can’t be trusted

#192
post #177

Earlier quoted context omitted.

Google Maps is terrible for hiding results. I'll use the filter to show restaurants with a 4.5+ star rating, and it hides dozens of places 4.5+ stars while instead showing results 4.0 - 4.4 stars and even some that have no reviews at all. I find the arrogance insulting. You want to see 4.5+ star Japanese food restaurants? Well, I don't think that's best for you, and I'm going to replace some of those results with 4.1…

I hesitate to suggest that this may indicate there aren't that many 4.5+ star Japanese food restaurants. Even great resaurants in my area struggle to break the 4.5 barrier.

Perhaps. But then why not just say “no results found”? Because that might give the impression to the “average” user that Google is broken and isn’t solving their problem. To the more sophisticated user, Google is just meddling where they shouldn’t.

Google is proving the central limit theorem by catering to the average.

Re: Bay Area woman is on a crusade to prove Yelp reviews can’t be trusted

#193

It's a protection racket. Total scam. If you have 20 minutes I did a show about it 2 years ago https://youtube.com/watch?v=e9rmh8O_ksM . I guess you can 2.0 speed that and get it done in 10.

https://www.youtube.com/watch?v=dechvhb0Meo

The original model was as a long-form instagram live stream which I understand nobody did but I was being experimental. Basically a single shot, single take show on alternative platforms. The portrait format was to accommodate that media expectation.

I made dozens of them ... they failed to gain traction. A few are over 100 views a couple years later (like this: https://www.youtube.com/watch?v=3byGVjPq468) but I have still failed to crack the world of social media. It's not zero ... that's nice I guess.

I'm planning to do a reboot sometime soon. I took a couple weeklong trips around the country to some research libraries in the past year or so, but I'd like to execute differently to try to be more effective. This was my third show run attempt.

I think I may need a new foil for my host. The guy I did the shows with is a longtime TV comedy writer (https://www.imdb.com/name/nm1104065/) but he's not exactly a second city graduate.

Re: Bay Area woman is on a crusade to prove Yelp reviews can’t be trusted

#194
post #34

Edit: I would have loved to add to the conversation, but people do love their pitchforks and I should be working instead of commenting on HN. :-) Admitting you worked for Yelp on HN: not even once

> It's amazing to me how many people have conspiracy theories about Yelp reviews, when in reality it's just really hard to determine good vs bad actors in review systems and it's possible to do a bad job as an engineer. When cards are cloned, it's exceptionally easy for a bank to determine which merchant is compromised by looking for common denominators. Take three actors. What are the odds that two of them have revi…

> When cards are cloned, it's exceptionally easy for a bank to determine which merchant is compromised by looking for common denominators.

Is it actually? My bank calls me to ask which transactions were mine and which ones werent, thus requiring outside feedback

Re: Bay Area woman is on a crusade to prove Yelp reviews can’t be trusted

#195

Good for her. Yelp's business model seems to work best if they can pick winners, which is unfortunate. If I search for "dinner" on Yelp for "current location" and look at the map, it leaves out an astonishing number of places that I can see within a mile if I drive down the street. And if your business model is centered around hiding all but the winners - who might pay for placement + reviews both - then you have to…

Google Maps is terrible for hiding results. I'll use the filter to show restaurants with a 4.5+ star rating, and it hides dozens of places 4.5+ stars while instead showing results 4.0 - 4.4 stars and even some that have no reviews at all. I find the arrogance insulting. You want to see 4.5+ star Japanese food restaurants? Well, I don't think that's best for you, and I'm going to replace some of those results with 4.1…

Google maps is also terrible at showing _only_ relevant info. I searched for “coffee” in a small area earlier today (small because I wanted to walk); it zoomed out to a much larger area and showed me coffee shops along with taquerias, oil change shops, banks, etc. both in the results pane and scattered all over the map. It was such a mess as to be unusable.

Re: Bay Area woman is on a crusade to prove Yelp reviews can’t be trusted

#196

Good for her. Yelp's business model seems to work best if they can pick winners, which is unfortunate. If I search for "dinner" on Yelp for "current location" and look at the map, it leaves out an astonishing number of places that I can see within a mile if I drive down the street. And if your business model is centered around hiding all but the winners - who might pay for placement + reviews both - then you have to…

Google Maps is terrible for hiding results. I'll use the filter to show restaurants with a 4.5+ star rating, and it hides dozens of places 4.5+ stars while instead showing results 4.0 - 4.4 stars and even some that have no reviews at all. I find the arrogance insulting. You want to see 4.5+ star Japanese food restaurants? Well, I don't think that's best for you, and I'm going to replace some of those results with 4.1…

It's so frustrating to find the perfect restaurant by chance and find it on the map directly when it didn't show up in a very targeted search of the same map.

There's also a lot of brigading on Maps. You'll see 5 1-star reviews in a row, and realize they are either duplicate accounts or friends who are all reviewing the same experience.

Re: Bay Area woman is on a crusade to prove Yelp reviews can’t be trusted

#197

It is such an easy thing to solve with a blockchain. We do not need a 3rd part like Yelp. Just a chain with 2 actors, restaurants and reviewers.

And that solves fake reviews... how?

I am pretty sure that we can use clustering for that. As they said there was a clear pattern the restaurant recognized. If they had all the data it would be easy. The reviewer could also ask for a token from the restaurant before submitting the review or a signature, there are plenty of options.

Re: Bay Area woman is on a crusade to prove Yelp reviews can’t be trusted

#198

Good for her. Yelp's business model seems to work best if they can pick winners, which is unfortunate. If I search for "dinner" on Yelp for "current location" and look at the map, it leaves out an astonishing number of places that I can see within a mile if I drive down the street. And if your business model is centered around hiding all but the winners - who might pay for placement + reviews both - then you have to…

First time I used OpenTable I went to a highly rated and expensive place, had horrible food, and then my wife spent most of mothers day in bed with food poisoning. 0/10.

Re: Bay Area woman is on a crusade to prove Yelp reviews can’t be trusted

#199
post #177

Earlier quoted context omitted.

I hesitate to suggest that this may indicate there aren't that many 4.5+ star Japanese food restaurants. Even great resaurants in my area struggle to break the 4.5 barrier.

Perhaps. But then why not just say “no results found”? Because that might give the impression to the “average” user that Google is broken and isn’t solving their problem. To the more sophisticated user, Google is just meddling where they shouldn’t. Google is proving the central limit theorem by catering to the average.

Google long ago decided that "No results found" is never an acceptable response. Same with regular Google search. It will just choose arbitrary things to show you instead.

It's one of the things that has reduced the usefulness of these tools.

Re: Bay Area woman is on a crusade to prove Yelp reviews can’t be trusted

#200
post #194

Earlier quoted context omitted.

> It's amazing to me how many people have conspiracy theories about Yelp reviews, when in reality it's just really hard to determine good vs bad actors in review systems and it's possible to do a bad job as an engineer. When cards are cloned, it's exceptionally easy for a bank to determine which merchant is compromised by looking for common denominators. Take three actors. What are the odds that two of them have revi…

> When cards are cloned, it's exceptionally easy for a bank to determine which merchant is compromised by looking for common denominators. Is it actually? My bank calls me to ask which transactions were mine and which ones werent, thus requiring outside feedback

That's your perspective. That's when -your- behavior doesn't match your previous behavior.

What they are looking for in a different context is "Customer A reported unauthorized transactions, and so did Customers B through D. Each of these customers, prior to these reports, all had a transaction occur at XYZ gas station, so is there a skimming problem at that gas station?"

Post reply on HN