Do Hostels Cheat on Online Reviews? (Survey Results)
Hostel Reviews, Booking Engines, and Directories
Here are the results on the survey about whether hostels are cheating on hostel reviews.
33 people replied to the survey. Most of them work at hostels.
- 2 people said: I don't know
- 4 people said: Not a problem
- 12 people said: Only a minor problem
- 15 people said: It's a widespread problem
Here's a chart:
I've edited the following anonymous comments for spelling and grammar, and possibly slightly modified the wording while keeping the meaning.
Here's a comment from someone who thinks it's not a problem:
I've heard of some hostel owners encouraging guests to leave reviews on the hostel computer before checking out, but i think it backfires because the guests resent being browbeaten and talk badly about the experience in their later travels.
Here are some comments from people who think it's a minor problem:
There's a bigger problem in the hostel industry. Hostelworld does not regulate whether a hostel signing up with them has a license to operate legally or not. This will affect the image of hostelling as some of these illegal hostels are very dodgy and do not meet the safety requirements (e.g., fire safety). Hosteling is a growing industry and a huge part of the world have yet to try out the real hostel experience, and not just some run down, dodgy place with no hostel culture. These people might be turned off staying in hostels forever.
An easy way to fix the problem, is if they are caught once, they have their listing removed for a day, caught twice, one week, caught three times, two weeks and so on. If you don't cheat you stay on, if you do cheat and are caught you are penalized by making you invisible on the site.
They pressure guests into making reviews.
Some hostels are heard to have pressed (hard) its guests to make positive reviews. They get guests' emails, then a couple days after they leave they send emails begging for a positive review.
Here are some comments from people who think it's a widespread problem:
Properties, should be categorized by age and facilities. With in each category they should then be ranked.
They offer free nights in exchange of a good review.
They give free stuff to guests in exchange of a good review.
They make fake bookings with prepaid gift credit cards.
Capitals with a lot of competition have a cheating problem, especially in Eastern Europe.
It is easy to know who is cheating and who is not. Just look what is written in comments in the low season in [Eastern Europe] and you will have such a laugh. The problem is not that the supposed to be top 10 hostels are buying their comments to get listed better; the problem is booking engine staff doesn't want to do anything about it! You talk to them. You show them and then they don't even check the fake comments regularly, not even warning the obvious cases! I'm not talking about canceling them, but just telling the owners "eeeeei mate watch out cause ..." NOTHING !!!!
[Hostelworld should] not to give the possibility to cancel the reservations after people have showed up.
Canceling the booking is not a way out to avoid the comment at all. In Hostelbookers you can't even really cancel the booking, you just reallocate beds. In Hostelworld the guest always receives the link after showing up in the hostel, even if you cancel his booking asking for deleting ?!?!?!?! hehehehehe well its soooooo obvious that few of the hostel owners have friend on the other side of the computer, its not hard to delete the review, everybody knows that!!! Problems is when you receive a comment done by a guest that wanted to threaten you - for example saying that of he gets free accommodation or he's going to comment badly. And you know what the booking engines do if you let them know about something like that to them????? NOTHING!!!!
Some people who thought hostels are cheating, gave some examples of hostels they thought were cheating. I looked at them and it looked like there might be some attempts at fake reviews, but nothing that had a serious effect.
Several examples:
A hostel with ~80% and somewhere in the range of 50-100 reviews was thought to be cheating, but there are no 100% reviews. I think it's unlikely.
Another hostel with ~80% and 50-100 reviews was accused of cheating, but there are no 100% reviews.
Another hostel with ~80% and 50-100 reviews was thought to be cheating and they do have a long string of 100% reviews. It's possible that they are trying to cheat, but they are still only at 80%. If they weren't cheating it would probably drop them into the 70s%.
Here's a screenshot of a long string of 100% reviews for the hostel above that only has about 80% rating -- highly suspicious:
[ATTACH=CONFIG]1092[/ATTACH]
Another hostel with ~95% rating was accused of cheating, but I've been there and I'm not convinced that they are cheating, unless it is in the form of them previously having a significant number of bad reviews and then asking HW to delete them all in order to get rid of the bad ones. I don't think they are writing fake reviews. It's a very high quality hostel and the guests were loving it.
The deleting of all reviews on Hostelworld is something I've seen happen recently with a hostel which is why I mention it. This particular hostel was (deservedly) the lowest rated hostel in their city, but somehow got HW to delete all their reviews. Now they are the highest rated hostel in their city, but have fewer than 10 reviews.
So, there are the results. What do you think?
- Comments
14 years
My personal opinion -- which could be wrong -- is that hostels try to do it, but it's typically not effective. You need too many reviews to make a significant difference in ratings numbers.
The example I posted above looked like the hostel made a string of 100% ratings for themselves, but they are still only at ~80% rating.
If a hostel pressures guests to make good reviews, I think it will eventually backfire.
I can't picture hostels asking a lot of guests to make a fake booking with their credit cards. It would backfire.
Hostels could ask friends to make fake reviews, but there are only so many friends that one knows well enough to ask, and I don't think it's usually high enough to take an 85% hostel to 95%.
I've been searching around in Google many times to find evidence that backpackers are talking about crappy hostels that asked them to make a review and I haven't found any. Sample query:
http://www.google.com/search?q=hostel+asked+me+to+make+a+review
I'm not saying hostels don't try, but I don't think it's highly effective -- unless there is a technical loophole.
I think a lot of the current problems with reviews could be solved by a few things:
- The most popular response on the survey is that cheating is a major problem. Whether it is or not, the booking engines could respond to it more publicly -- for example, when a hostel complains about a competitor making fake reviews, investigate the reviews of the hostel. An idea for an easy way to do it would be to make a simple admin interface accessible to booking engine employees on each hostel's page. You click the tab and it pulls up a table of the hostel's reviews with IP address, country, city, last 4 digits of credit card number, billing name and address, and other hostels booked with the same credit card billing address (to detect if people with multiple hostels are using the same info to make reviews at more than one property). Sort the list by the billing address city and country, and highlight rows where the rating is over 90%. That would make a visual interface where booking engine employees could quickly check for fake reviews and then send a reply to the complaining hostel, honestly saying, "we could not find any evidence of fake reviews" or "we have addressed the issue." Run all the complaints through CRM cases to make sure the hostel is happy that their complaint was addressed. That could help clear up the perception that fake reviews are a major widespread issue, and let hostels know that the booking engines care enough to even remove a few ineffective fake reviews here and there.
- Create a better algorithm to automatically flag suspicious reviews for manual review.
- Improve hostel reviews systems so that travelers can filter "the long tail." The long tail refers to a huge amount of information (or products) that can't be filtered by a single top 10 list. There are virtually endless top 10 lists inside of the long tail. Applied to hostels: "Top 10 party hostels in Buenos Aires as rated by men in your age group who are single, like beaches, alcohol, night clubs and party until 6am" or "Top 10 small hostels in Buenos Aires as rated by women in your age group who like like history, hiking, beaches and go to sleep at 10pm".
See also:
14 years
First off I would have to say the pool data is not even close to being big enough to provide any clear indication in any direction.
You click the tab and it pulls up a table of the hostel's reviews with IP address, country, city, last 4 digits of credit card number, billing name and address, and other hostels booked with the same credit card billing address (to detect if people with multiple hostels are using the same info to make reviews at more than one property). Sort the list by the billing address city and country, and highlight rows where the rating is over 90%. That would make a visual interface where booking engine employees could quickly check for fake reviews and then send a reply to the complaining hostel, honestly saying, "we could not find any evidence of fake reviews" or "we have addressed the issue." Run all the complaints through CRM cases to make sure the hostel is happy that their complaint was addressed. That could help clear up the perception that fake reviews are a major widespread issue, and let hostels know that the booking engines care enough to even remove a few ineffective fake reviews here and there.
You are talking some major functionality and dev time to pull this off to fix as you put it 'the perception' that fake reviews are an issue. That is allot of money to combat a perception.
Improve hostel reviews systems so that travelers can filter "the long tail."
That is a great idea, it is something we are looking at in RC3 or 4 of our booking engine, the issue is it has to be tag or similar powered or profile and tag powered through a members profile and free tagging. You cannot use a string, well you could but it would bog down the search function chewing so many resources that it just would not be feasible on a large scale.
14 years
You are talking some major functionality and dev time to pull this off to fix as you put it 'the perception' that fake reviews are an issue. That is allot of money to combat a perception.
I don't think it would necessarily cost a lot. The tab or button itself would be one small block of code:
Forgive my pseudo code:
Then the $review_verifier_link button links to something like:
example.com/review_verifier.php?hostel=1234
Then the code on the review_verifier.php page is something like:
It might not be more than a couple of afternoon's of work for companies that have inhouse programmers. Small price for customer satisfaction.
Just an idea though :)
That is a great idea, it is something we are looking at in RC3 or 4 of our booking engine, the issue is it has to be tag or similar powered or profile and tag powered through a members profile and free tagging. You cannot use a string, well you could but it would bog down the search function chewing so many resources that it just would not be feasible on a large scale.
More comprehensive (though more expensive) than filtering by tags would be something like this:
I was thinking that each user has some attributes:
I like:
- Party hostels
- Quiet hostels
- Small hostels
- Large hostels
- City hostels
- Rural hostels
- etc.
(multiple choices allowed)
Check all that apply:
- I like to go to bed early
- I like to stay up late
- I like things to be quiet
- I like to socialize with upbeat music
- I don't drink
- I like social drinking
- I love to party!
Make a bunch of questions like that. Then you have a list of attributes for the person.
Compare the results between two users to come up with a percentage. E.g., 78% similar. There might be more than one percentage and different factors might have to have their weights adjusted. E.g., age might be weighted less than whether someone prefers a city or rural hostel.
The exact math is beyond me, but I could draw the concept...
The end result is they get more accurate reviews and the site could better match the hostels to the guests. There wouldn't necessarily be something like five "top 10 lists" -- there could be a near infinite number of top 10 lists, each one tailored to the viewer:
[INDENT]"Top 10 party hostels in Buenos Aires as rated by men in your age group who are single, like beaches, alcohol, night clubs and party until 6am"[/INDENT]
14 years
The issue I see with the fake review issue, is it is so dependant on the site in question. Like where do you start or finish, so for instance we have HW and HB being currently the two major players BUT Lonely planet do a fair volume, boots n all would not be too shabby either then of course there is trip adviser and many many other sites that would have a pretty high volume of bookings. (we seem to forget just how many bookings affiliate sites do) then each major brand also has many sub brands and support sites.
So suddenly it does not sound so easy just to satisfy some hostel owners that you probably are never going to satisfy because they are always going to be saying Jack up the road cheats. (even though this huge network of booking sites has exposed them to a market they could never of dreamed of)
The long tail search or profiling as I like to think of it as, we are on the same page just calling things by different names, your example above is exactly how we have done it in the reviews system. (although guarantee when we finally show our review criteria here we will hear some say it is too involved)
14 years
So suddenly it does not sound so easy just to satisfy some hostel owners that you probably are never going to satisfy...
True, some people will never be happy with it :)
The issue I see with the fake review issue, is it is so dependant on the site in question.
I think the next generation of affiliate sites will include reviews from the booking engines. Hostelz.com already does it on hostel listings like this that aggregate the reviews of multiple booking engines:
[ATTACH=CONFIG]1098[/ATTACH]
In the near future, a fake review on a site like Hostelbookers will be a fake review on many affiliate sites too.
I don't know if Hostelworld is letting sites use their reviews yet. I thought their reviews were on Hostelz, but they're missing now.
The future of reviews and affiliates is going to be interesting. The Semantic Web is coming too... :)
14 years
Yes - hostelz basically just went ahead and used reviews from HB and HW it is a bit of contention at the moment, and the jury is still out on what is going to happen.
From a new brand perspective I believe it is not a good idea, it promotes further investigation something a booking site does not want, although we will be introducing some interesting functionality for people that have visited other sites before ours.
Reviews are always going to play a major role in the booking process, although we have a massive long road ahead of us we hope that our approach to the review system will offer the user a better rating system, and better insight from the editor reviews.
Furthermore we have to credit this forum for some of the ideas that have been included in the review system, proof more brains make for better results.
14 years
Reviews are always going to play a major role in the booking process, although we have a massive long road ahead of us we hope that our approach to the review system will offer the user a better rating system, and better insight from the editor reviews.
Looking forward to seeing it :)
14 years
I have to say that while I agree that cheating on reviews is rampant, but if the major booking engines started flagging reviews based in ip addresses, it would be a serious hassle for my staff.
We get a lot of bookings from HW, HB, minihostels, etc. and the vast majority of them are for one night. That doesn't mean that the guests we get from the booking engines only stay one night though. They often, very often actually, extend their stay, they just don't book through a booking engine. And we have a very high return rate (we're located in the entrance and exit capital of the country, so people stay with us at the beginning and the end of their trip, or while criss-crossing the country) for our guests, above 85% for guests that stay with us their first night in the country.
This means that a lot of guests write reviews while they are still staying at the hostel. So, I think you can see how this can cause some problems. I also don't see why a cheating manager/owner wouldn't just use a proxy server.
I'm almost to the point where, while I think this is rampant among some hostels, that it's almost a non-issue. I think most travelers these days are smart enough to read through the reviews, and that people who have only left one review probably don't carry too much weight, even if it is a legit review.
Does sending a Christmas gift basket to your HW rep constitute cheating?
14 years
This means that a lot of guests write reviews while they are still staying at the hostel. So, I think you can see how this can cause some problems. I also don't see why a cheating manager/owner wouldn't just use a proxy server.
I don't think that IP addresses can be used to find fake reviews for that reason. Also, some hostels have dynamic IPs. They are other things they could do also (though I don't want to mention them :)).
It would have to be done with other factors, like credit card billing address, or billing name and city. It would be good to have the IP address listed on the table though. If someone complains about a fake review and the IP address on most of the fake reviews match the hostel front desk, then it's just another clue. Not everyone is technically savvy enough to know...
There are a lot of other things that could be done with the algorithm too...
Pages
Related Pages
Log in to join discussion