New Presentation of Reviews on Product Pages? Like It or No?

Contributor: P'Gell P'Gell
Quote:
Originally posted by Selective Sensualist
I've noticed this, too.

For example, on one product page, the featured "critical" review rated the product as 4 stars, and this review has 11 Extremely Useful votes. But there is another review that also rated the product 4 ... more
I hope this doesn't lead to people overwriting and only using either 5 stars or only 1 star, simply to get on the review page.
05/21/2011
Contributor: CAKES CAKES
I like it!
05/21/2011
Contributor: namelesschaos namelesschaos
Quote:
Originally posted by Selective Sensualist
I've noticed this, too.

For example, on one product page, the featured "critical" review rated the product as 4 stars, and this review has 11 Extremely Useful votes. But there is another review that also rated the product 4 ... more
Considering looking through I notice some of my reviews where featured I hope I'm not biting myself on the ass by saying this but...

What you said seems to suggest that the ranking for the most helpful positive/critical is using a simple mean, which is problematic a review could have a 1000 EU and one useful and it would have a lower mean then a review with only 1 EU vote and no others. If it is a simple mean, a weighted average would be more indicative of usefulness. There is a lot of way of doing depending on how much wight EF, what to give number of votes v. their ratings.

But a simple unsophisticated way for illustration purposes is to multiple the mean rating by n/n+1 and use this weighted average for the ranking. Assuming the 5 ranks correspond to 5,4,3,2,1. The math would be something like like this

1 Extremely Useful (n=1)
Mean= 5/1 =5.0
Weighted average= (5/1)*(1/2)= 2.5

11 Extremely Useful (n=11)
Mean = 55/11= 5.0
Weighted average= (55/11)*(11/12)= 4.583

31 Extremely Useful, 5 useful (n=36)
Mean= ((31*5) + (5*4))/36= 4.86
Weighted average= 4.86* (36/37)= 4.730

This seem like a much more accurate view of usefulness...unless of course it bumps my review off the product page then it is perfectly fine as is, please disregard this message
05/21/2011
Contributor: Selective Sensualist Selective Sensualist
Quote:
Originally posted by namelesschaos
Considering looking through I notice some of my reviews where featured I hope I'm not biting myself on the ass by saying this but...

What you said seems to suggest that the ranking for the most helpful positive/critical is using a simple ... more
I think you're right about the system using a simple mean. The n/n+1 would be a good solution, I think, to highlighting the reviews the community voted as the most useful on the product page.
05/21/2011
Contributor: PiratePrincess PiratePrincess
It's not as easy for me to find things, but it may be easier for people just visiting the site for the first time. Although I agree that the better reviews should be the ones highlighted and not just due to star rating.
05/21/2011
Contributor: namelesschaos namelesschaos
Quote:
Originally posted by Selective Sensualist
I think you're right about the system using a simple mean. The n/n+1 would be a good solution, I think, to highlighting the reviews the community voted as the most useful on the product page.
I just realized this creates another problem if does just use a simple mean you have created a new incentive for down-voting another persons review. If you both have all extremely useful vote on your review what the best way to ensure your review get featured, give the other person a useful vote. People before where bothered about down-voting but, honestly for most people a single bad vote isn't going to have much of an impact but here a single useful vote does make a big impact.

People could easily do this intentionally, post review, have your buddy rate you EU ,down-vote current featured all EU review, bask in glory until you get a useful vote yourself. However it doesn't even have to be intentionally, if you know in the back of your head rating this vote useful is to your benefit you have an incentive to be more critical then you would other wise.

Also I've thought of another reason a simple mean is problematic ,it aggravates the frustration us who review off site have with people not clicking "Read the Entire Review". Getting down-voted by someone who didn't read your review is annoying enough now it pretty much torpedoes your chance of getting featured, with the weighted average even if you get a down vote your still stand a chance.
05/21/2011
Contributor: Kindred Kindred
Quote:
Originally posted by namelesschaos
Considering looking through I notice some of my reviews where featured I hope I'm not biting myself on the ass by saying this but...

What you said seems to suggest that the ranking for the most helpful positive/critical is using a simple ... more
If they decided to change to this type of equation, I would also love it if the votes were weighted based on rating of the person casting the vote.
05/21/2011
Contributor: Selective Sensualist Selective Sensualist
Quote:
Originally posted by Kindred
If they decided to change to this type of equation, I would also love it if the votes were weighted based on rating of the person casting the vote.
That would be awesome if they could do that.
05/21/2011
Contributor: MeliPixie MeliPixie
Wow! I've had to have limited access to other sites besides my class for a few days because it's finals time, but now that I've got time I see you all made some very valid points! I really hope someone who can do something, sees this thread, I like to think our constructive criticism can go over well and no fall on deaf ears. Thanks, everyone! ^_^
05/22/2011
Contributor: ShadowKitten ShadowKitten
I hate the new system, it makes viewing opinions and finding previously viewed reviews a major pain in the ass and will likely make me take peoples reviews less into consideration when looking at toys because I don't feel like fudging about through numerous pages just to find bloody factoids...EF shouldn't become Facebook in the sense they don't need to change these things. They were all fine before and now are just a mess :\
05/23/2011
Contributor: Naughty Student Naughty Student
Quote:
Originally posted by Selective Sensualist
I've noticed this, too.

For example, on one product page, the featured "critical" review rated the product as 4 stars, and this review has 11 Extremely Useful votes. But there is another review that also rated the product 4 ... more
I find it kind of awesome, I like change but I agree with both your and Kindred's criticism's as I have noticed them myself.

I agree that some people might do something like play with the rating to get featured, but I am sure these type of people will be obvious to pinpoint and the rating of their review would probably not be high enough to allow them to be featured. People will notice their intentions and will probably rate the review accordingly.

We love EF and we stick together so when we see something off we notice and things get done. The admins work hard like hell to keep tumultous waters calm, and to keep abuse of the system as low as possible.

Their will always be rotten apples.

But overall I try to have faith in people's good will, and I love the new system. You can still get the full list of reviews on the left side of the screen while looking at one of the reivews on a product. So I don't think it's a big problem, I feel like part of the old system is still there in that sense.

And I love the idea of the weighted averages, it's so simple and will be more representative.
05/23/2011
Contributor: P'Gell P'Gell
Quote:
Originally posted by Kindred
If they decided to change to this type of equation, I would also love it if the votes were weighted based on rating of the person casting the vote.
That would be great. It always irks me when someone who has NO reviews or only a couple of poorly written reviews doesn't understand the work so many of us puts into our reviews and downvotes due to not understanding the WORK we all did.

This new system will make even a "useful" review a "bad" vote, as namelesschaos said. When it's hard to see ALL the reviews, then there is more of an incentive to want to be "the" chosen review. Also, my fear is people will not want to give 4 stars or 2 stars or 3 stars, as it appears the highest and lowest rating given a product is more likely to be chosen.
05/23/2011
Contributor: Sir Sir
Quote:
Originally posted by Red Vinyl Kitty
I agree with this, also, there are some reviewers who I prefer to read if I see they've reviewed a toy because I trust their opinion and/or their body is similar to mine. This new system makes it harder to find the reviews by the reviewers I ... more
Agreed. I haven't been so active lately (in fact, I listed myself as inactive), but I saw that and did not find it too pleasing. The reviews need to be easy to find - people who are new to Eden will not have an easy time finding them.
05/23/2011
Contributor: Sir Sir
Quote:
Originally posted by P'Gell
That would be great. It always irks me when someone who has NO reviews or only a couple of poorly written reviews doesn't understand the work so many of us puts into our reviews and downvotes due to not understanding the WORK we all did. ... more
Yeah, I saw this problem too. I saw a review get on the front page that had six extremely useful votes, when one of mine had about nineteen or so and one useful. There was another review, I believe, that had about twenty-six extremely usefuls and NO useful votes, and the rating was 5-stars too, but it was not on the front page. Don't understand how that works out!
05/23/2011
Contributor: Sir Sir
Quote:
Originally posted by P'Gell
I hope this doesn't lead to people overwriting and only using either 5 stars or only 1 star, simply to get on the review page.
I do not think that it would get that bad just based on a page layout, really. That's a little over the top. However, there are some kinks that need to be worked out.
05/23/2011
Contributor: Antipova Antipova
I've not been able to be online in the evenings lately so I wasn't able to come to the review workshop. Were any of these constructive criticisms brought up?

You guys have made some really good points---

(First, having the histogram is a great idea!)

I agree with namelesschaos's weighted averages instead of pure averages for review ranking.
------
**as an aside, I'm not even sure if pure averages are used in all the cases. In one product, when all reviews are sorted by "most helpful first," each of the top four reviews (each giving 5 stars) has only EU votes, and no lower votes. But the highest-ranked review has 17 EU votes, the second has 23 EU votes, the third has 16 EU votes, and the fourth has 18 EU votes. Maybe the date the review was written factors in, too?

------

Kindred makes a good point about a vote having more or less weight depending on a contributor's rank. However, there is some merit to having it be perfectly democratic---there are lots of people who use the site who are not reviewers, and really, their opinion matters, too. I know we get very involved in our community, and it's easy to think "I'm writing this for my friends on EF," but we are also writing this for people who never make their presence known on the site. I try to write mine for a someone new to sex toys who just stumbled in. Before our names were recorded with our votes (what's that, 2 years ago now? Guess I spend too much time reading EF history...), any anonymous user to the site could vote. That seems fair on one hand, because those are the people we really want to help make their decisions. At the same time, though, anonymous people don't understand that the review "grades" have an effect on those of use who write the reviews, and might vote carelessly. So it makes sense to have to register to vote.

And so looking at that, I totally understand why it would be attractive to say "higher contributer rank, means their vote counts more..." I know that over the course of a week, one contributer rated about half of my reviews a grade lower than *every other voter to date*, and when that contributer wrote their first review a little while later, the writing quality was terrible (imo). I was frustrated. (And I refrained from voting on the review.) But at the same time, we want to be friendly to newcomers. And we don't want voting to be a popularity contest. So I think it's necessary to take the hits that come from unfamiliar voters, because their thoughts and voices matter too. We can't just say "well that frustrates me so let's change the system to only benefit the established reviewers."

-----

Along with Selective Sensualist and P'Gell, I think it makes more sense to weight pure usefulness (by whichever algorithm you decide) instead of showcasing the most useful 5 star and the most useful 1 star reviews, because a) we'll start to see more 5 and 1 star reviews as reviewers subconsciously vie to have the showcased slot, but more importantly b) when I'm looking to buy a product, I usually learn a lot more from reviews in the middle: 4 star, 3 star, and 2 star reviews, where people have honestly weighted pros and cons, and not fallen into the (to borrow a phrase from UP!) "I just met you and I love you!" fallacy, or into the "this didn't work for me so there is nothing good about it" fallacy.


I think this is important both for the community at large and for the rest of the shoppers of EF who are just browsing through.
-----

I also miss being able to check at a glance on the product page who the author of the review is, because I know some reviewers have similar responses to me. At the same time, I see how that matters more to those of us in the community than it does to random shoppers on the site.

-----
I do worry that showcasing only two reviews will lead to voting that jiggers for placement. At best, people will be totally magnanimous and it won't be a problem. In the middle scenario, if a reviewer knows (s)he will be reviewing a product, (s)he may not cast votes on the reviews (s)he reads as she makes her purchase decision. At worst, people might vindictively give lower grades to good reviews just to gain position.

I believe that most people on Eden are above that worst case scenario. But the middle scenario really doesn't seem that far-fetched to me.

-----

I think that MeliPixie is looking in a good direction when she hybridized the old and new systems: You keep the histogram, and the featured positive and negative reviews (with the weighted average upgrade, and the "all star types have equal opportunity" upgrades, hopefully), and you also list all the other reviews in the same one-line format that they had been in, taking up less space but still visible, with the author's name. Hopefully that will take some pressure off the "grading other reviews just to be noticed" effect.

-----

SS said this well, so I'm going to quote her:
"As with anything new, I realize there are glitches to be worked out. The above is given in the spirit of feedback and is not meant to be an attack. I know the tech team works hard; therefore, I wanted to also give feedback on the nice, neat appearance of the new product page. I am just hoping that some of the good features about the older system could somehow be incorporated with the new."
05/30/2011
Contributor: Dusk Dusk
A lot of people have written a lot of great criticisms above that I don't feel I need to re-hash. But I will say that they are valid criticisms, I DO NOT like the new system, and I hope we can revert back to the old or a hybrid version before I lose my mind.
05/30/2011
Contributor: VanillaFreeSex VanillaFreeSex
here is the community of reviewers reviewing the reviews layout...find it kinda cool. I love our group of opinionated opinions! the positive, the negative, the suggestions. we do rock!

anyways...I LIKE that you can sort reviews by either newest first or most helpful first. Now it is like shopping for a review.
06/04/2011
Contributor: Shellz31 Shellz31
At first I liked the idea that I could see how many reviewers gave the item 5 stars etc.

I still like that, but why couldn't that just have been added and leave the way the list of reviews where before!
It was much easier to go into reviews. Made finding a toy to free review easy too - now it's a bloody friggin nightmare!
I dislike going through reviews now.
06/05/2011
Contributor: That Man from Mars That Man from Mars
Quote:
Originally posted by Red Vinyl Kitty
I agree with this, also, there are some reviewers who I prefer to read if I see they've reviewed a toy because I trust their opinion and/or their body is similar to mine. This new system makes it harder to find the reviews by the reviewers I ... more
I agree as well, with what you and Alyjazz have said.
06/06/2011
Contributor: daniel and frances daniel and frances
Quote:
Originally posted by MeliPixie
Hi everyone, is it just me or do some toys have a new way of presenting the reviews on their pages? Like, it used to be just a list of reviews, but just today I have noticed on the product pages there is now a chart-type thing displaying what is ... more
I'm all about easily classified info.
06/13/2011
Contributor: pixieluv pixieluv
I hate the way that reviews are showcased right now as reviews that were over two years old when the product was new and exciting get 21 votes and NEW reviews on the product get 2-4 votes. If the 21 vote old review is placed as the most helpful, only it will be read and then voted on and made "more" helpful whereas other reviews that have new information (especially on products that have been updated) are never scene in the front lines. This means they will not be looked at and will not be voted on.

BAD BAD BAD system.

If a review is going to front line (be on the front product page) then it needs to be a random choice, changing every time the page is loaded with the link to all the other reviews.
06/18/2011
Contributor: pixieluv pixieluv
Another thing I just noticed, it doesn't even matter if another better review comes along and gets more extremely helpful reviews, it still does not show up as the most helpful review.

The example I saw was one where the first 5 star review was in January 2010 and currently has 21 extremely useful votes and 1 useful vote. It is listed as the "most helpful favorable review" on the product page. However, when you go to the page with all the reviews there is a 5 star review dated in January 2011 (one year later) that has 31 extremely useful votes and 3 useful votes. You would think that this one would be the one that qualifies for the "most helpful favorable review" but it does not.

So the "most useful" review system doesn't even work! (BTW if you are interested the product is: Touche Ice Big
06/18/2011
Contributor: MeliPixie MeliPixie
Wow, that's crazy! I just assumed there was a sort of algorithm or something so it could instantly identify which was "most useful" and update automatically. Maybe the webmasters have to go in and update things by hand?! That seems like a huge waste of company resources... Hmmm....
06/18/2011
Contributor: namelesschaos namelesschaos
Quote:
Originally posted by MeliPixie
Wow, that's crazy! I just assumed there was a sort of algorithm or something so it could instantly identify which was "most useful" and update automatically. Maybe the webmasters have to go in and update things by hand?! That seems like ... more
Actually that not crazy that is exactly in line with what I already out lined before, it does have an algorithm: a simple mean. It seen "crazy" because we keep (rightfully, which is why I advocate for a weighted mean) the review with expecting the most EU votes to comes out on top but aren't actually doing the math.

Here is the math:

Review one (the one in the voted most helpful slot)
n= 22 ( 1 useful, 21 EU)

(21x5)+ (1x4)/22= 109/22 =4.955

Review 2 (the one that seems like it should be the most helpful)
n= 34 (3 useful, 31 Eu)

(31x5)+(3x4)/34= 167/34 =4.912

Mathematically the first review has a higher mean. Again it seems crazy because it does line up with what we intuitively perceive as usefulness.

To compare a weighted mean, using the method I described above:
Review 1: 4.739
Review 2: 4.771

If we where to adapt this system then things would line up with out intuitive perceptions.
06/18/2011
Contributor: Tuesday Tuesday
Quote:
Originally posted by P'Gell
That would be great. It always irks me when someone who has NO reviews or only a couple of poorly written reviews doesn't understand the work so many of us puts into our reviews and downvotes due to not understanding the WORK we all did. ... more
Kindred, P'Gell, I enthusiastically support this idea. I too have noticed people with no reviews themselves rapidly rating reviews, spending less than a minute on each one.

I also really like namelesschaos's weighted averages idea.
06/18/2011
Contributor: namelesschaos namelesschaos
Quote:
Originally posted by namelesschaos
Actually that not crazy that is exactly in line with what I already out lined before, it does have an algorithm: a simple mean. It seen "crazy" because we keep (rightfully, which is why I advocate for a weighted mean) the review with ... more
I made some typos here the biggest is that it should read: "it seems crazy because is does NOT line up with what we intuitively perceive as usefulness". The math is still right.
06/18/2011
Contributor: Antipova Antipova
Quote:
Originally posted by namelesschaos
Actually that not crazy that is exactly in line with what I already out lined before, it does have an algorithm: a simple mean. It seen "crazy" because we keep (rightfully, which is why I advocate for a weighted mean) the review with ... more
I wonder if they're even using the simple mean right now, though. Sort these by usefulness:

link

The "first-most-helpful" review has 18 EU votes and no others, and the "second-most-helpful" has 23 EU votes and no others. Both are five star reviews, so why do those sort backward?
06/18/2011
Contributor: namelesschaos namelesschaos
Quote:
Originally posted by Antipova
I wonder if they're even using the simple mean right now, though. Sort these by usefulness:

link

The "first-most-helpful" review has 18 EU votes and no others, and the "second-most-helpful" has 23 EU votes and no ... more
Ok now that one IS crazy.

The majority of examples line up with a simple mean yet, that one clearly doesn't. I can't figure out what the algorithm behind that one is.

I'm stumped, anyone got any ideas? Math Geeks we need you!
06/18/2011
Contributor: Tuesday Tuesday
Quote:
Originally posted by Antipova
I wonder if they're even using the simple mean right now, though. Sort these by usefulness:

link

The "first-most-helpful" review has 18 EU votes and no others, and the "second-most-helpful" has 23 EU votes and no ... more
I've noticed that new reviews get a boost for a while just for being new.
06/18/2011