Blackberry BBM App and Suspicious Google Play Ratings 67
sl4shd0rk writes "In what could be an act of desparation of a company in it's death throes, Blackberry has submitted their BBM messaging application to Google Play for download. While this may seem like a logical path for a company on life-support, what wasn't expected is the sheer number of identical 5-star reviews the application has received since being posted. In what appears to be review 'ballot stuffing,' it poses the questions of just how Google is going to handle the subject of manufactured reviews as well as how many other entities have engaged in the same behavior. The same problems have plagued Amazon's review system as well bringing into question the validity of 'crowd based review' and whether it's possible to legitimize this type of system." The linked article points out that the suspicious posts may be the result of ballot stuffing intended to hype one of the unofficial Blackberry apps, rather than RIM's own.
Really (Score:5, Insightful)
5 stars! (Score:5, Funny)
(I'll wait patiently for my check now)
Thanks blackberry (Score:3, Funny)
Re: (Score:2)
Re: (Score:2)
(I'll wait patiently for my check now)
Cash it quickly when you get it...
or maybe they just made a good app (Score:1, Interesting)
Maybe they just made a good app that people like?
Comment removed (Score:5, Funny)
Re: (Score:1)
Re: (Score:1)
I downloaded it. I can't use it - they put a really long waiting list in place in order to try it; which makes these reviews even MORE suspect.
I can tell you that they failed to use Retina assets in 2013. TWENTY THIRTEEN USING LOW RES GRAPHICS.
Re: (Score:1)
They had SIX-MILLION pre-registered users at the moment of launch (who could activate and use the app immediately after downloading it). In addition to that, they've cleared over 5 million more since then and are working on ANOTHER 5 million. So I'm pretty sure between the 11+ million NEW, ACTIVE users and the 65-million existing Blackberry users, there's plenty enough people already using the app in order to review it...
Also, wtf do you care about "retina assets" on a text-based IM application?! wtf?! is t
Re: (Score:3)
The waiting list is about 3 days, mine came through this morning. Maybe that's really long to kids these days, but I'm old enough to remember the waiting lists for Google Wave, and Facebook when it first went from invitation only to open to the public.
I suspect fanboys. (Score:1)
There are plenty of BlackBerry fanboys around, especially here in Canada. I believe it's far more likely that some idiot script-kiddie fanboys are behind this than actually BlackBerry employees... Or, if it IS BlackBerry employees, it's people acting alone who are afraid for their careers.
They're stupid either way, because of course it just makes BlackBerry look ridiculous (not that they need any more help with that nowadays).
Re: (Score:3)
Re: (Score:2)
Perhaps this was done to purposely embarrass BB?
No, Blackberry doesn't need any help doing that themselves.
Re: (Score:2)
Re: (Score:2)
RIM is indeed dead because they don't exist anymore after they changed their name.
Unprecedented (Score:1, Offtopic)
I mean Samsung would never stoop to such levels [androidcentral.com]
Re: (Score:2)
I am shocked! Nothing like this has ever happened on the Internet before.
FTFY.
Nothing compared to Kik Messenger (Score:1)
Re: (Score:1)
Thanks Blackberry (Score:1)
Softly now (Score:1)
Reminds me of the negative review wars, too. Antivirus reviews frequently exclaim it installs virus, even of legitimate products.
This on top of normal reviews where people are more likely to go bitch over minor problems than praise -- the one in a thousand guy wins.
Re: (Score:2)
polarizing reviews (Score:3)
Crowd-sourced reviews suffer from at least the following issues:
- reviews by people who have no business reviewing (Amazon.com)
- reviews only by people who feel strongly about it (Amazon.com, app stores)
- aggregate ratings based on averages, not presented as histograms (amazon.com and app stores have started adding this in the "details", but it's still gameable)
- changes to reviews over time are not obvious
I'd like to share how, despite its many problems, Bricklink does a fairly good job on this particular topic. As a buyer or a seller, you are heavily encouraged (it's part of the workflow) to rate every single transaction. There are no reviews that are not based on experience, and each experience is rated only once. While a total count of reviews is shown, there's no other aggregate value shown that could be misleading -- by the time you see the reviews, they're already broken into a simple histogram (good, bad, neutral) for comparison. They also do a sort of log(t) rating system on the reviews: they're broken up into current-month, current-year, and all previous years combined. So you can tell if things have recently taken a turn for the worse, or someone's tried to fix an image problem by actually improving. History is not lost, but for a potential buyer, recent history is highlighted.
I'd like for reviews (Bricklink, Amazon, etc.) to be broken up into aspects -- the product itself, customer service, shipping, etc. But I recognize that by asking more questions, you raise the barrier to entry, and you'll get less (and much more biased) data. I see far too many 1-star reviews on Amazon not as a result of the product itself, but of the shipping or customer service.
I kind of feel sorry for app developers who embed a "rate my app" feature directly in the app. It feels gimmicky, it feels like they're trolling for 5-star reviews, and yet it kind of makes sense -- try to hit up every user with the question, even if they wouldn't have naturally thought to bother, and do so after they've started using the app, so you get a fairer opinion. But mixed with in-store reviews, and the ugliness of "rate me 5 stars, get bonus stuff for free" offers... ugh.
Re: (Score:2)
Bricklink does a fairly good job on this particular topic.
what is this, like a lego thing?
Re: (Score:2)
Yes, it's a site that allows you to buy pieces by the brick/specific minifig/part you're looking for. I was trying to get into the whole Lego Arbitrage thing a few years back and got quite familiar with Bricklink.
http://www.bricklink.com/ [bricklink.com]
Re: (Score:2)
The fact that it's a peer-to-peer LEGO marketplace didn't seem particularly important to point out, except this:
The product (almost entirely new and used LEGO bricks) is well known to both buyers and sellers, and the reviews are therefore only about the service provided (accuracy of parts, quality of shipping, timeliness, correct representation of the state of the product). There's little risk of someone posting a review about a seller complaining that a given piece is, in a generic sense, good or bad. Keep
Re: (Score:2)
We should be careful in building review systems, particularly "find stuff I will like" pages, to:
- give low or zero-review items a chance, until it's fairly certain they are unwanted (wide margin of error, benefit of the doubt)
- give irrelevant items a chance, purposefully breaking the relevant-items algorithm
People may have selected categories they like (or we may have determined them automatically) but as Daniel Tiger says, "you gotta try new foods, they might taste good!". It's far too easy to lock custo
Re: (Score:2)
That would be a better way than the way most stores work now.
I think a lot of apps ask a minority of their users, for example 10%, for ratings. That way most users aren't annoyed by it. You've probably used lots of apps that ask for ratings unbeknownst to you because you weren't one of the 10%. That percentage could then be lowered over time as the rating stabilizes at a high level, so for a mature app it may only be 5% who get asked to rate.
It can be mitigated by simple actions... (Score:2)
Google has no clue about social spam. (Score:5, Informative)
Back in 2011, I wrote a paper, "Social is bad for search, and search is bad for social" [sitetruth.com] There, I described the social spam ecosystem, from the SEO firms to the phony account generators to the proxy sellers. I named some of the big social spammers.
Most of the same companies are still social spamming. In the paper, I mentioned "Google Plus1 Supply". [googleplus1supply.com] They're still active. They're still selling "+1"s. Their site looks almost exactly the same as in 2011. But their prices have gone down, and their number of fake "+1"s sold has increased from 4 million to 33 million. BuyPlus1Fans.com [slashdot.org] is still up.
Where do they get the accounts? BulkAccounts.com [bulkaccounts.com] is still up, just like they were two years ago. They're an outsourcing firm, using low wage labor to create new accounts. For an automated approach, there's JetBots [jetbots.com], which claims to be able to create 250,000 new accounts per day on a fast connection. They offer "CAPTCHA Bypasser", which runs CAPTCHA's through OCR, and when that doesn't work, ships them to an outsourcing firm for manual recognition. Once the account is established, their "voter bots" add any desired number of stars to reviewed items.
Facebook is no better. BulkLikes.com [bulklikes.com] is still up. In 2011, they charged $260 for 500 Facebook fans. Now, it's only $70 for 1000 fans.
Old-style link spamming was expensive - spammers had to set up content farms, run servers, refresh them with interesting content, and worry about their farm being blacklisted. Social spamming is cheap - Google, Facebook, and Yelp host the spam for free. Yelp tries to push back against social spam; they've sued some spammers. But Google and Facebook don't seem to be trying at all. The fact that the big spammers of two years ago are still big spammers clearly show this.
social scoring (Score:2)
I am the first one to hate social networks for their privacy concerns. But what we actually need is social scoring for app reviews and all. You could build a recommendation based on the friendship relationships. Since I am not going to be friend with "spamer1234", the impact of their score whould be very low. Of course, you need something more than just take the average of your friends. You need something closer to "personalized page rank" or a graph based inference system. But anything in that matter would
Re: (Score:2)
I'm subjected to a bit of this through that "facebook" thing. Suggested posts based on friends or friends-of-friends liking it? I've yet to see that return relevant stuff. That could be because I have too few friends (!), or because each person only "likes" a small fraction of what they actually find interesting, or because there's so much out there, that there's no reason to expect even two close friends, with similar interests, to both like the item (especially close in time to each other.) For something
Re: (Score:1)
Re: (Score:2)
And besides, just because I'm friends with someone, doesn't mean we have similar tastes.
Pretty much this. I care and would be influenced by what brand one of my friends might go with for a gaming rig power supply... but couldn't care about his selection in cars, or wine, or clothes.
Another friend I might care about his tastes in wine, but would probably slap him upside the head if I even KNEW what hardware he was putting in his PC.
But I don't need (or use) facebook. When I'm buying wine I'll ask the friend
Re: (Score:2)
I don't think that would work out particularly well, without a lot more granularity than is really feasible. My best friend and I have similar taste in novels, for example, so I'd want to weight his opinion on them pretty highly. On the other hand, I absolutely loathe 99% of his music, and when it comes to tech... he's not exactly a power user (didn't even realize that his STB was network-enabled, much less used the function).
"Social" selling is bad enough with the silly tribe-mindedness that surrounds it n
More than just suspicious of (Score:1)