Hacker News Reader

Comment by ahallock | Hacker News Reader

What does a useful local business review site look like? Not being snarky--genuinely interested to see if there's an opportunity here.

View on HN

I think Yelp is missing proof of identity and proof of patronage. It was created before social media took off, and before you could reliably know if someone had been somewhere.

I don't know if you need both, but proof of patronage alone might not be enough to trust a review (See Amazon verified purchases).

Aren't Yelp "Check-ins" factored into the trustworthiness rating of a review, or whatever it's called?

That's kind of an ersatz proof of patronage.

Maybe. It still wouldn't be foolproof - after all, Yelp Check-ins are powered by an API that trusts the coordinates it receives.

It might be possible to craft a request for that API. And even using a legitimate copy of the Yelp app to make requests, a jailbroken iPhone can change its coordinates to whatever you want. An Android doesn't even need that - just set it to developer mode and you're ready to fake check-ins all day.

Would you ever trust an app that was linked to your identity and stored a list of places you visited? Be careful what you wish for.

Is that much different than what Facebook or Google Maps are already doing behind the scenes? They might not make that data accessible to you, but it's undoubtedly being (or able to be) compiled behind the scenes. See this update:

http://techcrunch.com/2015/07/28/google-search-now-shows-you...

The average HN user might not care for such a system, but I think many people would find it reassuring for better or worse.

It's a great question. I think a lot of the problem is reviews get stale AND you don't know/trust the reviewer.

Yeah. I don't live in the boonies but I also don't live in a big city. Pretty much any restaurant or small business around where I live has, at best, a handful of reviews over the past few years. It's just not very useful.

Biases baked into the platform are one problem for a review site. But you also just need critical mass.

I think the biggest things working against Yelp are the reports from businesses that Yelp offered to censor bad reviews and/or promote/shill positive reviews in exchange for money.

I've heard about it all N-th hand, so I'd imagine that Yelp's team is very careful about wording it such that it's some sort of legit promotion that they're offering.

...and yet, despite what must be hundreds of thousands of calls or emails made by Yelp sales folk to businesses, nobody, to my knowledge, has ever produced a voicemail, telephone recording, or message containing a promise or threat to do anything other than what Yelp publicly offers to its paying customers[1].

I'm actually sure this has happened, since it's essentially impossible to prevent thousands of salespeople from deviating from an approved script when pushing hard for a sale, but surely it can't be commonplace.

[1] https://biz.yelp.com/support/full_service_advertising

Exactly. I made the same comment above before reading yours. I also find it hard to believe that after all these years, not ONE person has solid proof of shenanigans.

Zagat?

Zagat is (still) pretty decent if you're interested in restaurants in one of the metropolitan areas that it covers and you're in their foddie-ish demographic. However, I don't think Zagat, which had a pretty long pre-Web history in what we now call crowdsourced restaurant reviews, offers a lot of insight for how we might do a small business review site today.

The question didn't qualify "useful" as being crowdsourced or not. And that gets to my point, a review site that has its own reviewers is pretty much the only thing that is going to be useful. Otherwise the system will be gamed by vendetta seekers and the business itself. And I believe it is inevitable that a crowdsourced site will ultimately end up gaming the system to try to offset those effects.

The bigger problem with crowdsourcing is that the reviews doesn't have a calibrated baseline. As a business traveller at 150+nights/year, my demands on a hotel are different from a leisure traveller's. Having a nightclub in the hotel is cool if you're going to that nightclub, but it's a special kind of hell to get back from a long day and have to fight your way through a nightclub line to get into your $600/night hotel (never mind having to deal with the noise from the nightclub). Another pet peeve of mine is to deal with clever/funky/cool "concepts", like that one time I tried to order a drink in a all-organic bar (according to the bartender things dry and bitter can't be sourced organically, so they only did sweet drinks).

Trip Advisor (most relevant for hotels) and Yelp is supremely bad for surfacing the aspects I care about from the depths of things I don't. It's better with restaurants, as most people seem to agree to a much higher degree on the baseline.

>pretty much the only thing that is going to be useful

IMO that's an overstatement. I find Zagat, Yelp, and Amazon reviews all useful--however imperfect. That said, however unpopular this opinion may be in our crowdsourced world, professional reviews/ratings tend to be more reliable than random ones on the Internet. The problem is that professional reviews don't scale, especially for local businesses. The New York Times restaurant reviews aren't bad but that's NYC and, even in that case, a lot of the reviews are going to be quite old.

So, in an ideal world, there might be armies of professional evaluators of hotels, restaurants, and other businesses--but that's not the world that we live in. (Films are one of the few areas where you can get a pretty good pool of professional and semi-professional reviewers.)

I think smart pay systems might help here. I'm not an avid reviewer - I just don't care enough for it to be worth my time. However, uber makes it part of the workflow. To get another ride I either need to make a rating or skip making a rating - so I make a rating. And it only takes one click. I suspect if Apple Pay or Google's equivalent had a similar process they would be able to generate some very sold scores for local businesses since they have both identify and proof of purchase.

Yeah but with Uber, the expectation isn't that you're going to leave a nuanced review about the comfort of the car seat and the smoothness of the ride, but to basically leave 5 stars so long as there was nothing horribly wrong. Which is fine. It's more or less a commodity service.

With restaurants, on the other hand, there is a wide range of quality. Even if the average shopping mall restaurant delivers a hot meal in a reasonable timeframe, I'm probably not going to give it 5 stars.

If reports are to be believed, then clearly Uber expected that it would be 5 stars or nothing, and their low tolerance for 4 stars makes reality more consistent with that belief. But as a user from the early days, nothing in the app clued me into this expectation and accordingly I gave 3 stars for "meets expectations", 4 stars when I was offered refreshments and USB connections, 5 stars when I couldn't imagine a better ride...

It's only lately and due to the prevailing rumor of Uber's mercilessness that I give 5 stars to meet "don't fire the driver on my account" and less than 5 stars to mean "I demand a refund and the head of the driver."

Back in the day when it was primarily an "Internet flea market", although it didn't use a star system, eBay was similar. Both sellers and buyers expected positive feedback accompanied with a comment like "A++++++++++ turned water into wine" so long as the reality wasn't "sent me a box full of bricks and axe murdered my extended family." Neutrals for a transaction that didn't quite go right even if issues were resolved were NOT appreciated.

Reply on news.ycombinator.com