I'm in the process of developing a review-driven site, and I'm wondering how so many reviews get written on a site like Amazon when there's no incentive to do so (from what I can tell.) The incentive on a site like Yelp seems to be badges and awards given by other users to your reviews. How do you get people writing reviews on your site?
I think Amazon reviewers are people with opinions, who relish the idea of sharing their opinion with the world. I don't think these people need any more motivation than that-- no badges or flair required.
However: I think that the reviewer-to-reader ratio is tiny, and this is masked a bit by Amazon's absurd popularity.
And, to make matters worse, this absurd popularity is one of the driving motivators (cf above, "share one's opinion with the world", not "share one's opinion with the 200 people who happen to follow my micro-site.")
If anyone knows any way out of this besides increasing the number of users by several orders of magnitude, I'd love to hear it-- I, too, have a site that allows reviews, and only a tiny handful of reviewers (so far.)
I think the main thing would be to build a community where people feel some need to belong. I don't know if Amazon really does this, and what you're seeing in terms of number of reviews on their site is probably just another reflection of their popularity. Some % of people will write a review, Amazon just has enough people that it ends up being a rather large number.
The problem then becomes how to encourage a greater % of your audience to contribute. Some people might be more willing to contribute if they could see that their contributions were valued. Can users rate a review (like Amazons X of Y people found this review helpful)? Some might be more willing if it increased their standing in the community - some people enjoy being seen as an expert - do you recognize people who take the time to leave reviews (perhaps magnified by review quality)?
Most of this will only really work if it is within a community, however. If lots of people just come and go then those tactics are fairly meaningless. If you have a niche site with a lot of repeat visitors that you can build a community around, this might be a good way to start that process.
Currently either a reviewer has 2 choices: 1. put a review on Amazon, but hand over ownership of the review. 2. put a review up on a personal blog, and maintain ownership but get no readers. Micro-formats offered a way to publish reviews and have them more easily aggregated, but haven't delivered. It is clearly a non-trivial problem.
"If anyone knows any way out of this besides increasing the number of users by several orders of magnitude, I'd love to hear it"
Disclaimer: I have never run a reviews site so this might just be a wild idea.
What if you implemented a karma system like HN/Reddit/Digg and rewarded users for their relative karma score for particular comments to other users commenting on the same thing. For those who achieve a certain relative score, you can give them prizes, discounts, etc. I mention relative score since this would curb the incentive to game the system.
Don't forget that there are also a lot of Amazon reviewers who are nothing but trolls. They are making it harder to post reviews like that, but I still see plenty of cases where people have hijacked Amazon comments or reviews for the lulz.
> I'm wondering how so many reviews get written on a site like Amazon
I'm sorry to burst your bubble. The well-respected consultancy User Interface Engineering has come up with the magic number of 1 review for every 1,300 purchases.
"While only one in 1,300 purchasers of the product writes a review, the number who indicate a review was helpful is even fewer. For the Harry Potter volume, which is Amazon's best selling product ever, it was about 0.0014% or about one in 7,300 purchasers of the product. The most helpful review garnered only 566 votes, even though it was written on the first day the book was released and Amazon has sold more than 2,000,000 copies since."
Not everyone who buys a copy even bothers to read the review. If it's a book that the customer already knows they want (which is probably often the case with the Harry Potter books), why read the review? Just buy the book.
And I know that I for one usually don't upvote a review, even if I did find it helpful. I don't mean to be intentionally rude, I just somehow haven't gotten in the habit of doing that. So the number of votes on a review isn't an indication of the number of people who found the review helpful, but rather, the number of people who both found the review helpful and took the time to vote it up.
That said, yes, Amazon's immense customer base makes even tiny percentages of users contributing out to be useful. But they weren't always so large. I've heard that the main way they became the prominant online bookseller was precisely because they offered a review mechanism, at a time when other online bookstores did not, or offered a more feeble review system. Since it's so easy to add a nice review system to an online store today, that feature alone won't be enough anymore.
It's even worse than that. It's not the Harry Potter books (plural) they were talking about--it was just one particular Harry Potter book, book 7, which was the last book in the series. How many people are going to read 6 books in a series, and then need a review to decide if they want to read the 7th?
About the only reason I can think of one might check reviews before buying is to see if there are any goofs, like chapters printed out of order or things like that, that might make you want to delay purchase until the second printing.
The Amazon review system is kind of annoying. There are three kinds of things I've seen people cover in reviews: content, presentation of content, and purchasing issues. For instance, I did a short review of a math book. The content was excellent. However, the conversion to ebook for the Kindle was unusable. Various mathematical symbols were not recognized during the OCR and were replaced with blank space.
Amazon provides no way for me the rate the content at 5 stars and the presentation at 1 star for the Kindle edition. (And it provides no way to review just the Kindle edition. The reviews for all the editions are collected together).
> Since it's so easy to add a nice review system to an online store today, that feature alone won't be enough anymore.
It's not the feature. I'm convinced it is is the sheer volume. Much like Arn from MacRumors says to monetize a website - first you need traffic. Traffic then monetization. Not monetization then traffic.
Target has the exact same software Amazon.com has - as the result of a 2000-era licensing agreement.
I had to sort by highest rated to even get a Harry Potter item with reviews:
"The secret to success on the internet can be boiled down to one simple accomplishment: building traffic.
That’s it. If you have a site that attracts a lot of visitors, you will be able to make money. On the internet, traffic equals power, which subsequently equals money."
I find Amazon reviews fascinating for precisely the reason you indicate: they're incredibly widely used, despite not appearing to use many of the psychological tricks you see elsewhere.
I think there are a bunch of reasons Amazon reviews work so well:
1. The audience. If you want your opinion to be heard, posting your review on Amazon is the best way to go - just like selling your stuff on eBay makes sense because that's where the buyers are.
2. The high quality of the existing reviews. They set a standard which new reviews then attempt to match. That's the result of years of careful management of community expectations.
3. The "Top 1000 reviewer" badges. They're relatively subtle as badges go, but offer a powerful incentive for the best Amazon reviewers to keep on reviewing.
I've written reviews on Amazon; not lots, maybe in the dozens, but far more than I do for other shopping sites.
I think most fundamentally, I find a lot of value in reading Amazon reviews written by others, and I feel like it would be useful for me to write my own reviews. I don't tend to review every product I possibly can, but rather, I usually just write a review if I have something interesting to say, or if I feel very strongly about an item, good or bad.
I like to think by doing this I can help other people make more informed decisions about what to buy.
As others pointed out only about 1 in 1000 are posting reviews on Amazon so they aren't doing anything particularly great, they just have volume.
I will share one trick that I use on my site UniversityTutor.com, to get more reviews. I saw this in an email that Google Checkout sent out:
1. Firstly, of course you should send an email asking the person to leave a review a week or two later
2. Do NOT require the person to login to leave the review (use a "one time login link"). Removing this hurdle can help.
3. Embed the actual review form RIGHT IN the email. With a submit button.
Basically all this is about reducing the hurdle to leaving a review as much as possible. The "archive" button is just to easy so if they have to leave the email or login you've already lost people.
Submitting forms doesn't work in all email clients so you'll have to also include the one time login link for anyone who can't submit the form right from their email. But it does work much of the time. To generate the HTML for the email form that works across most email clients, the easiest way I've found is to create a new form in Google Docs, and email it to yourself as a starting point to copy from. The HTML they generate is pretty well optimized from what I've seen.
I still don't have tons of reviews on that site though. So it is definitely tough to get people to engage.
Also - the simpler the review the better.
A pulldown will get filled out more often than a free form text field.
Radio buttons will get filled out more than a pulldown.
A simple thumbs up thumbs down will get filled out more often than 5 stars.
And only a thumbs up (or like) will get more than thumbs up or down.
When I moved from a 5 star thing to just up/down on BuyersVote.com it helped a bit with engagement.
I've written a single Amazon review. I spent hours finding an external USB disk enclosure that would support S.M.A.R.T. and I wanted google to shorten other people's search process.
I find myself motivated by the extremes. Either a product was so fantastic that I felt the need to reward the manufacturer with something more than my money or it was so terrible that I felt the need to get back at the manufacturer for taking my money.
I do not think people participate on Yelp to get badges. That is ridiculous. I have a friend who recently fell in love with Yelp. She uses it to figure out what restaurants she wants to go to in New York City. She likes reading the reviews, good and bad. She feels like she's become good at figuring out what places she will like, even when they have bad reviews. Reading between the lines on Yelp has become a kind of sport for her. Even more recently, she began contributing reviews. I think for her it is a way of sharing with others what she likes, and participating in the Yelp community, which is a community that she likes. The question is sort of similar to "Why do you have the friends that you have?" Why are any of us drawn to one group of people, and not some other group of people? Or why do we go to some show, some party, follow some musicians? We get something out of it, but it clearly isn't badges or money, or merit points, or anything ridiculous like that.
A large number of reviewers on Amazon are rewarded for good reviews by publishers sending them free stuff.
That's what an Amazon engineer told me: the real name program and the top 1000 badges were designed to encourage this, even though it creates a reward for very shallow, positive reviews.
I've thought EXTREMELY long and hard about this - also because I run a review driven site. I wrote up a blog post talking about some of the more abstract/theoretical concepts revolving around motivation:
http://mysimplemindedworld.wordpress.com/2009/04/16/a-framew...
So under my framework I'd say the major motivation is that the people contributing reviews are those that are passionate about the Content/Topic. And the simple reason why they get so many is because 1) ridiculous Amazon volume that somebody already mentioned, and 2) Amazon is the place where you voice CAN have the most impact (if you're one of the top).
Reasons people will post reviews of a product on Amazon:
1) That's where they bought it -- sometimes when I buy something at Amazon I get an email from them asking me to review it
2) Sharing one's opinion is its own reward
3) Lots of people visit Amazon, so your opinion is going to be meaningful. What's more, your review will appear on the page for the product itself.
4) If you use Amazon a lot, you look at the reviews before buying. This means you are aware that they exist -- you trust them, maybe. Having the chance to add to a corpus of reviews you use and respect is a powerful incentive.
I'm not sure a plain-vanilla reviews site will be able to compete successfully with Amazon reviews. Maybe find some niche, like IMDB (movies) or Harmony Central (musical instruments) or KVR Audio (music software).
I just want to clear this up: I'm not developing a competitor to Amazon - I already have a niche for the reviews, but without the reviews the site loses much of its value.
I write reviews under a pseudonym because I either really like the product I purchased (rare) or someone really pissed me off (common, half my textbooks) and I want the entire world to know about it but I don't want it coming back to haunt me. :)
I have a review-driven site as well, RateMyStudentRental.com, so here are a few observations...
-People review for all sorts of reasons, but the largest reason students write reviews on RMSR seems to be simply to "give back to the community." In other words, they just like helping. However, keep in mind that my site helps students find suitable housing, so compared to product review sites (like Amazon), there's probably much more of a charitable motivation for RMSR.
-People like sharing and helping, but not that much. If the review form requires too much info, they'll just say "ahh, fuck it," and leave. RMSR's review form is still WAY too long if you ask me, but I just haven't had time to change it. That actually leads me to a piece of advice: try to get your review form nearly perfect from the outset. If you have a semantic site that arranges data in a meaningful way from the reviews, the review form becomes difficult to change, because every change requires some data migration of all your existing data to the new common format. Not to mention the extra logic you have to add to the site. E.g. if you add a criteria on which to rate, do all the existing reviews get a zero? Obviously not, so now you have to add in some if-statement for "N/A" and get fancy with your AVERAGE calculations.
-When we started, there was a lot of fear (some from us, mostly from landlords and school officials that were on board) that the site would just be a landlord-bashing playground for tenants. Thankfully, rather than site around and try to invent rules, incentives, policies, etc, we decided to simply address it as it became a problem. This was one of the best decisions we ever made, because here we are 3 years later, and it still hasn't become a problem. Out of a 4.0 scale (like report card grades), the average rating for rentals on the site is like a 3.5 or something. This brings up another interesting point: they say reviews tend to display the extreme ends of the spectrum (either really satisfied or really dissatisfied). FWIW RMSR probably still has way more A reviews than C, D, and F reviews combined.
-At one point, I tried integrating some game mechanics, by giving badges and stuff for the number of reviews a student has written. But we ended up getting rid of it mainly because it added complexity and no one seemed to care. No one cared most likely because RMSR is a site that students use only a couple times a year. If you only use a site once every 6 months, why does it really matter? Note that I could be totally off on this, because our "badge" implementation was half-assed at best, so perhaps I just didn't try hard enough.
-To motivate students to write reviews when we first started, I wrote a promotion engine, where if students from a certain area of schools signed up and wrote a review, they'd get a print-out coupon for a free coffee from a local coffee shop. In return, the review would get the coffee shop's logo placed on the review for all eternity, which would link to the coffee shop's website. I wrote this such that it could work with any local business. It worked really well, but by far the hardest part is getting businesses to give stuff away that is actually worth the students' time.
-Some students leave reviews because their landlords asked them to, so that's perhaps a motivator specific to us. I've even heard reports of landlords waiving rent late fees in return for writing reviews (though I haven't seen any indication that it affects the quality of the review or rating, which I was afraid of when I first heard of it).
As for the game mechanics, I would imagine that the low frequency of visits really was the major factor in the apathy towards them. Our startup (IActionable) is creating a game-mechanics as a service system and one thing we noticed is that it's only really useful when employed on a site that users typically visit fairly frequently (like once a week or better, ideally). There are other factors too - like number of integration points - or how many things can a user do on your site. The "game" is hard to make interesting if the user can only do one or two things since there isn't a whole lot of interesting stuff you can do around that usually.
Quick Question: was there any fallout when your removed your "badge" implementation (from people who'd earnt a badge)? Fear of this makes me reluctant to try new things, but I think it's good to try them. Did you do something to handle this, or was it just a non-problem?
It wasn't really a problem. When I said no one really cared, I'm not exaggerating, they literally didn't care. I guess it could have been a problem, but I don't think there was a single user who had actually tried to earn a badge.
I write reviews for things I purchased occasionally if I think there's not enough good information about the product in the reviews section, there was something I couldn't figure out until I'd purchased the product, or I really like (or hate it). I'm not expecting much, although the idea that other people might read it and find it useful makes it worth it.
Many Amazon product pages are the de-facto location for product data in a variety of verticals. When was the last time you read a great book and sent someone a Borders/BN link? If that page page becomes the default home for that product, posting a review at Amazon becomes natural. This also leads to weakness in other areas, and I'd argue that for product categories where Amazon is not the natural choice, the review quality isn't fantastic (e.g. outdoor gear, etc.) If the goal of the review is to inform as many people as possible about your product experience, Amazon is a natural outlet.
I've posted maybe two. The one I'm sure about was for a very handy O'Reilly book that got a bad review that was undeserved on the merits and made a passing reference to the author's politics. I thought the book better than that and the politics (confined to the dedication) not relevant to the book's worth.
Amazon and Newegg both email me after I make a purchase requesting that I review the product. They do this a little after the product is delivered--to give me a chance to use it.
I almost always respond to these requests, because I find the reviews helpful in my buying process and that motivates me to contribute back.
In some cases, reviews are a matter of Internet Forum Argument. This happened a lot before they added actual forums--the reviews for anything by Ayn Rand dissolved into Ayn Rand arguments.
I suspect that most reviews are pissed off customers that are simply looking for an emotional outlet. I am actually amazed that so many products on Amazon have a high rating.
Amazon could probably increase reviews 10 fold by creating an incentive program (e.g., coupon codes for reviewing products).
Or, though it's a little spammy, I wouldn't be opposed to receiving a Netflix/Ebay style email a few weeks after I order a product asking for me to give a star rating on it.
I suspect that most reviews are pissed off customers that are simply looking for an emotional outlet. I am actually amazed that so many products on Amazon have a high rating.
The fact that you're amazed at the outcome is a good indicator that your hypothesis is wrong.
I think incenting people to review products defeats the whole point of what Amazon has to offer--genuine opinions about products from actual users. As mentioned above, no other e-commerce site has volume of reviews that Amazon has and few look poised to overtake them. Epinions took the incentive route and I think the difference in value is pretty apparent. Once you start compensating users for submitting "something", the value of that "something" is going to be pretty minimal.
Yeah - but putting an incentive behind the process would remove the obvious self-selection biases. Right now, there are two reasons that people would post a review:
1) They had an exceptional experience with the product and feel the need to react to that.
2) They are the type of person that likes to tell the world about their experiences.
Unfortunately, both of those are a very small subset of total purchases and, most likely, highly overlapping. Which means you won't get the proportionally accurate number of reviews in the middle range. And you won't get reviews from Luddites or online recluses.
If you put an incentive behind the review system, perhaps you'd have people gaming the system, but at least you'd have a better shot at hitting a wider demographic of product reviewers.
You are probably right though. Amazon, like Google, is very, very good at analyzing piles of data to make market decisions. They may well have already piloted review incentives. Regardless, I suspect they've already got numbers that back up their current process.
I think the incentive compounds the self-selection bias. For someone who is an expert in say audio equipment, a small monetary reward isn't going to encourage him/her to post a review (assuming they have a decent amount of disposable income). However, for those who could greatly benefit from that incentive, they'll post reviews like there's no tomorrow.
Another great example of how Amazon has succeeded with user generated content is Listmania--genuinely helpful product lists, all provided for free. I don't completely understand the psychology of why these are created, but I value the content greatly because I know that contributors are not compensated for their suggestions.
However: I think that the reviewer-to-reader ratio is tiny, and this is masked a bit by Amazon's absurd popularity.
And, to make matters worse, this absurd popularity is one of the driving motivators (cf above, "share one's opinion with the world", not "share one's opinion with the 200 people who happen to follow my micro-site.")
If anyone knows any way out of this besides increasing the number of users by several orders of magnitude, I'd love to hear it-- I, too, have a site that allows reviews, and only a tiny handful of reviewers (so far.)