Customer Discussions > top reviewers discussion forum

Who decides where your review goes in review listing?


Sort: Oldest first | Newest first
Showing 1-23 of 23 posts in this discussion
Initial post: 10 May 2010 04:22:00 BDT
I do not have a particularly high rating, but I only started in March 2010. At first I wondered why people cared what their rating was, now I am addicted, although my reviews are getting worse.

My question is really about who, and what a nice job, decides to put a review you have just posted in a high position as most helpful. I just posted a very brief review for a film and they were 44 previous reviews, to my pleasant surprise, and this has happened quite a few times, Amazon placed in on page 2 out of 5 pages of reviews.

So who is it, like Barry Norman or someone, and it is done so quickly, makes the mind boggle about who is sitting at a screen just waiting for you to write a review or I notice also, reading your "fix recommendations". It is all so IMMEDIATE!

Also I notice I mentioned Vincent Price's The Tingler and the next day I found it on my recommendations list.

Feels a bit like BIG BROTHER IS WATCHING YOU. I'm not complaining as it is very useful. Be careful what you write though, as AMAZON ARE EVERYWHERE!!!
Your reply to Scrumdiddleyumptious's post:
To insert a product link use the format: [[ASIN:ASIN product-title]] (What's this?)
 

Posted on 10 May 2010 07:09:02 BDT
Last edited by the author on 10 May 2010 07:19:56 BDT
Ethereal says:
Having just done one review I'm a novice and no doubt will be corrected if I'm wrong, but it seems fairly clear from where mine is in relation to only 6 others: largest number of helpful votes at 100% first; largest number of helpful votes with highest percentage but not 100% next; then date order with most recent first.
As to who decides, I imagine it's programmed into their software because it's so immediate. Obviously someone at Amazon does read stuff when they get round to it, which probably explains finding something you mentioned on your recommendations list.
Edit: I forgot there's also a button you can click to sort reviews by most helpful or date order first.

Posted on 10 May 2010 08:34:50 BDT
Danny says:
Its all calculated by a computer. There is nobody sat there manually adjusting the rankings. Perhaps when you die, if you have given too many negative votes, the Gods that govern us may gather your soul up and inject it into that sad little algorithm, there to spend the rest of eternity raising and lowering reviews on a whim, but thats the only extent of human intervention you will get.

In reply to an earlier post on 10 May 2010 13:36:26 BDT
Well it was very funny, immediately after posting this discussion and just after the said review being placed on page 2 out of 5 previous reviews, I wrote another one, which was actually better than my earlier one. Much more informed and with lots of helpful information about the product, and strangely enough, my review was on the last page of quite a few pages. SO YOU ARE TELLING ME IT IS DONE BY A COMPUTER?
I can understand the recommendations thing being done by a computer, but what a computer! You are not telling me a computer can be spiteful or even judge where to put a review????????????????????????????????????
(Computer says, "we'll show you, you little ratfink, critising Amazon!!!!!!!!!!!! YOU ARE NOW ON BAD LIST!)The Haunting [DVD] [1963]
Classic Sci-Fi Collection : Invasion Of The Bodysnatchers / Thing From Another World / Incredible Shrinking Man / This Island Earth / Creature From The Black Lagoon / It Came From Outer Space [DVD] (The computer might have changed my listing after reading this?)
Just because I'm paranoid doesn't mean they're not out to get me!

Posted on 10 May 2010 13:43:41 BDT
OEJ says:
It's actually very simple. Your review of the Sci-Fi collection is where it is because it hasn't had any votes. All the ones above it have had votes, and the only one below yours is there because it has more negs than pos. Nothing massively technical, and done with no human input at all.

Posted on 10 May 2010 15:06:57 BDT
Damaskcat says:
I have no idea . . . . - you are making wild assumptions I'm afraid as Pundit has explained. Don't forget correlation does not equal causation

In reply to an earlier post on 11 May 2010 03:28:54 BDT
I don't think you understand. The earlier post had no votes either but was put on page 2 of 5.
Doesn't really matter I suppose, Pundit and Damaskcat, correlation does not equal causation, what the hell has that got to do with anything!

In reply to an earlier post on 11 May 2010 06:21:13 BDT
Last edited by the author on 11 May 2010 06:41:41 BDT
Ethereal says:
Not going over or under their suggested word limit for reviews, perhaps, and maybe star ratings are taken into account too? Amazon prioritising highest star ratings.....if you're concerned you could email customer services and ask. I believe the system isn't foolproof and there are errors made which can be corrected on investigation.

Posted on 11 May 2010 08:44:29 BDT
Danny says:
Call of Duty: Modern Warfare 2 Signature Series Strategy Guide

If you look at the above product:
The top review is 2 stars with 7 out of 7.
The second review is 2 stars with 2 out of 2.
The third review is 5 stars with 9 out of 11.

This tells us:
1. Number of stars do not have any affect on a review's position.
2. Negative votes affect your position more than positive votes.

I would guess that they use a simple percentage. IE 80% positive would be higher than 50%, so 8 out of 10 would go above 2 out of 4. However, the system is obviously very basic, as it doesn't take into account the sheer volume of votes. Should 2 out of 2 really be better than 9 out of 11?

But in answer to the question, there is no human intervention. Its a silly notion when you think about it. Look at the number of review products there are on Amazon. To maintain them all manually would be a monumental task.

Posted on 11 May 2010 09:46:35 BDT
JJG says:
I assumed (though I have nothing to base this on) that a reviewer's score played a role in where they were placed on a product's review pages. Obviously, once voted on, this would no longer apply unless two reviews tied.

Posted on 11 May 2010 10:43:00 BDT
Last edited by the author on 14 May 2010 16:20:08 BDT
Damaskcat says:
I have no idea what's going on!. you suggested that immediately after posting a comment on here you posted another review which went on the last page of reviews for that product - suggesting that someone at Amazon dfdn't like what you've written on here. I merely pointed out it was coincidence and had nothing at all to do with your comments on here - i.e. no one is punishing you for your opinions - hence correlation does not equal casusation - it was coincidence

Posted on 11 May 2010 11:06:43 BDT
Last edited by the author on 11 May 2010 11:10:09 BDT
OEJ says:
Last year I did some analysis of this topic - in the days when I cared a lot more about it. I'll copy and paste it here but being on Excel I have a feeling it will look messy and difficult to see clearly. It's in reverse order, in other words the worst figures are at the top. It shows an exact pecking order for different review 'scores', with 100% at the bottom and many variations of less than 100% above it. The purpose of this set of figures is to demonstrate how Amazon allocate positioning, and AFAIK it's mainly (but not exclusively) down to percentage of positive votes relative to the total.

They may have changed the formula since then, but I doubt it.

To help you understand the data, the first entry is 35 out of 90, which is calculated to be not as good as the one below it, which is 1 out of 6. And so on. The figure on the right is the percentage figure of positives relative to total, e.g. 35 out of 90 is 39% positive. As you will soon see, the percentages alone don't determine the order, because 39% is obviously better than 17%, but in this algorithm it falls below. All of these numbers were taken, if I remember correctly, from reviews of THE LOST SYMBOL by Dan Brown, and are shown here in ascending order of positioning at the time of my analysis (last November I think, not sure)

35 90 0.39
1 6 0.17
1 5 0.20
3 15 0.20
1 4 0.25
2 8 0.25
1 3 0.33
2 7 0.29
4 14 0.29
13 45 0.29
4 13 0.31
2 6 0.33
5 15 0.33
1 2 0.50
2 5 0.40
3 8 0.38
5 13 0.38
42 116 0.36
3 7 0.43
2 4 0.50
13 31 0.42
5 11 0.45
3 6 0.50
6 12 0.50
26 54 0.48
67 133 0.50
2 3 0.67
1 1 1.00
7 13 0.54
6 11 0.55
3 5 0.60
4 7 0.57
12 23 0.52
8 15 0.53
10 19 0.53
9 16 0.56
7 12 0.58
6 10 0.60
12 21 0.57
10 17 0.59
9 15 0.60
8 13 0.62
12 20 0.60
7 11 0.64
19 31 0.61
18 29 0.62
3 4 0.75
17 27 0.63
8 12 0.67
36 57 0.63
14 21 0.67
12 17 0.71
2 2 1.00
6 8 0.75
8 11 0.73
9 12 0.75
26 36 0.72
12 16 0.75
37 51 0.73
20 27 0.74
69 95 0.73
23 31 0.74
95 131 0.725
41 56 0.732
119 162 0.735
45 60 0.75
3 3 1.00
25 32 0.78

In reply to an earlier post on 13 May 2010 05:32:17 BDT
Pundit, your last post really explains why I have no idea what's going on. And I am not upset at my rating or placing in a review list. I haven't done any empirical research just a single case study, MINE!

I still don't think you understand my earlier posting. I am not angry about my reviews but was interested as to how and why your new review might go well above much older and higher rated reviews. At that time of posting a new review you obviously have no votes at all as no-one has actually read them. Surely with your logic a new review should go to the last position until it receives some votes either way. That has not been the case with my reviews, that is all I was trying to say.

Posted on 13 May 2010 08:25:14 BDT
Danny says:
Yes but negative votes have an effect too. So a review with 1 positive and 2 negatives may be placed below a review with no votes. That is an issue, indeed. I think the system only really gets effective when a few votes have been cast. Before that, it all seems a bit of a lottery.

Posted on 13 May 2010 09:02:58 BDT
Damaskcat says:
Why would it matter to people where their reviews were placed on the product page? If people want to read reviews they will read reviews not just the ones which are immediately visible. When someone writes a good quality review - whether positive or negative - it will find its way to the top of the pile. If no one wants to vote on your review it will remain in oblivion - so what?

Posted on 13 May 2010 13:37:50 BDT
Danny says:
Indeed. I am amazed at how important some people find all this.

In reply to an earlier post on 13 May 2010 15:57:57 BDT
OEJ says:
IHNIWGO said "I still don't think you understand my earlier posting"

I do, and I did. You wanted to know how and why reviews are positioned where they are, and I provided a rather OTT answer, but one that should leave no room for confusion. It's possible that in the first few hours a review is 'wrongly' positioned, possibly because the pecking-order filter may not operate every minute of every day. But once a review IS placed in accordance with the algorithm, it goes where it goes in keeping with the numerical formulae I listed. Generally speaking, a review with no votes at all gets placed higher than those with negative weighting.

I think it's an imperfect system, as everything else is, but what the hell. You asked, I answered.

Posted on 13 May 2010 16:31:10 BDT
Danny says:
Incidently, if you publish reviews such as this, you are liable to get negative feedback for them:

http://www.amazon.co.uk/product-reviews/003013661X/ref=dp_top_cm_cr_acr_txt?ie=UTF8&showViewpoints=1

-------------------------------------------------------------------------------------------------

See my review for this book in paperback, have upgraded as old one wearing out from re-reading. The War Against the Jews, 1933-45

-------------------------------------------------------------------------------------------------

Posted on 13 May 2010 18:28:36 BDT
Last edited by the author on 13 May 2010 18:31:15 BDT
A review with no rating often gets placed higher than those with positive weighting as well. Under the new system, 0 votes goes above, say, 4 from 6 positive.

'Why would it matter to people where their reviews were placed on the product page? If people want to read reviews they will read reviews not just the ones which are immediately visible. '
No, they'll read the immediately visible ones first, then maybe go onto the next page.

Re: War Against the Jews. Why not just copy and paste your old review, rather than tell people to seek it out? I agree with Mark Twain, that review is obviously going to attract negative votes, for good reason.

In reply to an earlier post on 13 May 2010 23:24:13 BDT
Mark Twain, I am amazed too. After all it was just a question, not the Spanish Inquisition, think I'll stop tracking this pointless discussion, it has become pointless now, and personal comments about my reviews are pointless too. I think I do have some positive votes, around 40% so far. I will go back to the discussion regarding the negative voting system. Ta Ta!! God some people really take this seriously don't they.

Posted on 19 May 2010 00:36:52 BDT
Last edited by the author on 19 May 2010 00:38:25 BDT
Geek Girl says:
The most helpful review for The Thousand Autumns of Jacob De Zoet (bestseller #17) has no votes at all. Why is that considered more helpful than a review which has picked up 10 helpful votes already? I don't get it.

Posted on 19 May 2010 06:42:59 BDT
Danny says:
The one with 10 helpful votes also had 6 negative votes. Similarly the one with two helpful votes also had a negative vote. However, I gave that one another helpful vote, as it was a good review, and it is now placed top. So amazon treats nothing better than 2 / 3, but not 3 / 4.

The reviews for that book do seem to be polarised somewhat, which probably skews the results a bit.

Incidently, if you check out the other reviews by the Red on Black chap, he's done an hilarious one on some Lily Allen clone.

In reply to an earlier post on 28 May 2014 18:16:47 BDT
[Deleted by the author on 28 May 2014 18:16:58 BDT]
‹ Previous 1 Next ›
[Add comment]
Add your own message to the discussion
To insert a product link use the format: [[ASIN:ASIN product-title]] (What's this?)
Prompts for sign-in
 


Recent discussions in the top reviewers discussion forum (310 discussions)

 

This discussion

Participants:  9
Total posts:  23
Initial post:  10 May 2010
Latest post:  28 May 2014

New! Receive e-mail when new posts are made.
Tracked by 1 customer

Search Customer Discussions