Hacker News Reader

Comment by bnralt | Hacker News Reader

One of the first things I do when I'm considering a piece of media is I read the negative reviews. Finding out the reasons why someone doesn't like something is a good way to see if you'd like it yourself. For instance, most of the time I'm going to enjoy a film that gets dinged for being "confusing" than I am for one that gets dinged for being "bland." Others might find the opposite to be true.

It's not that I'm necessarily going to agree with the negative reviews. But I definitely find them much more useful.

It's a sad state of affairs when we're told that there's only one correct opinion about a particular piece of media and dissenting views are "wrong." We lose the deep insight that comes from looking at who likes and dislikes a piece and why. And it's hard to truly like something when liking it is presented as the only correct choice from the start.

View on HN

I do this for everything. Negative or low score reviews for products on Amazon are the most useful for deciding whether something might not meet my needs.

Though it’s very easy to fall into the opposite trap and get disillusioned with everything. Some people just have bad luck or like to complain a lot.

It's important to evaluate the negative reviews for yourself. Often enough they'll complain about things that won't bother me or aren't relevant to the product itself or what I need it for. The frequency is also an important data point. If a single review out of 1000 says that it broke after a day, it has far less weight than if 10 or 50 reviews complain about quality.

> Some people just have bad luck or like to complain a lot.

But this is exactly what you look for. How many of the bad reviews are people who seem like they whine a lot, how many are just unlucky, how many have seemingly reasonable complaints, etc.

How dumb are the bad reviews? If most are just stupid, it’s probably a good product, if you can find at least one review that really seems like it was by a real person with a real issue, it’s an indicator that some of the reviews are at least real.

Yeah, but my issue is I find products, almost all having quite a few bad reviews that all seem real and about serious issues, and then I end up only getting a cheap item because if it sucks at least I didn’t spend much.

But maybe that’s just a "me" issue ;)

I've had this issue too, actually. I think choosing a cheap version is a perfectly valid solution. Either the cheap item meets your needs or you notice things that bother you and that experience can inform your next purchase.

In some cases, I've instead decided that a completely different kind of product can fill the same need in a better way.

> notice things that bother you and that experience can inform your next purchase.

That’s what often ends up happening, but then I also don’t want to be wasteful and buy something new before the old thing breaks.

I've also personally seen a pattern in low score reviews on Amazon or other shopping sites where the person leaving the low score just simply didn't understand how to use the item properly, in some cases this is a majority of the bad reviews.

Another is when they had a problem with something external to the item itself such as shipping.

One time I was reading reviews of teapots on another company's (not Amazon) website. They seemed like really well-made beautiful porcelain teapots, and I couldn't figure out why there were so many people leaving disappointed reviews about the teapots cracking. Eventually after noticing certain details and reading certain reviews, I realized the reviews were all from people who had been using the teapots to boil water, as if they were kettles. Suddenly these comments went from reviews of the teapots to an index of how many people don't understand the difference between a teapot and kettle.

I feel like there's a big gap in retail websites for reviews that encourage more back and forth between the reviewer and reviewee and maybe some independent arbitration to draw this stuff out.

Something like a cross between an issue tracker and a lightweight court of law.

If Amazon were whacked around the head with an antitrust stick like Microsoft was in 2001, this is the kind of innovation that might actually flourish from newer startups.

I've tried this for books and ended up not reading any books. Well articulated reviewers always seem to find pretty damn good reasons to dislike any book. Buying on recommendation has worked ok, I still end up buying a few that I can't get into, but at least I find some to read.

This works for technology or products as well. The first thing I do when evaluating a framework is googling the phrase "technologyXY is terrible" and see what comes up.

For me as as well. Same for dislike buttons. I d like to see them. Shame they dont have it anymore at youtube.

Interestingly, I don't think I really found dislike counts to be very useful, though I do go for negative reviews on books first.

Negative reviews tend to be a bit more articulate and specific about what they didn't like, and why they didn't like it. Pain points tend to stand out a lot more than positive things, especially if you go into an experience with the expectation that you'll enjoy it.

That said, there's also plenty of occasions where I've found nothing useful in the negative reviews, but found some positive reviews compelling enough to try a book.

Movies, TV shows and music I tend to ignore specific reviews altogether, and just go by the description and overall positive/negative average from ratings. There are so many more dimensions that short form reviews generally aren't useful at all.

I wonder how useful it would be for a dislike button to require a brief statement explaining why. I'm not a fan of dislike/downvote buttons in general because there are so many different reasons why something could be perceived as negative, and what might be considered negative by one person might be a positive for somebody else. It just feels like noise to me.

Eh, tbh dislikes on YouTube were more of a popularity contest for the outrage du jour or a sign that some streamer told his followers to brigade something.

I’m still disappointed that YouTube/Google didn’t find it worthwhile to figure out brigading and render it useless.

Wouldn’t that be an entirely solvable problem? Or am I expecting too much from machine learning?

Given the gigabytes of data that just viewing the content generates for them, I'm sure they could correlate those actions. But given it took them many months to deal with extremely obvious comment spam... I have low expectations.

This makes a lot of sense since quite often many positive reviews are just fake and paid.

Reply on news.ycombinator.com