3 ms·

First of all, I think it's very likely that one of your own (ex-)employees will either go public with the information or blackmail you in order to buy your sile

First of all, I think it's very likely that one of your own (ex-)employees will either go public with the information or blackmail you in order to buy your silence. Secondly, a sudden increase in the number of positive reviews should be fairly easy to spot by a competitor who has already hired a developer to watch for such things and do automated sentiment analysis on the reviews of competitors. Thirdly, it's entirely possible that a broader-spectrum monitoring of reviews (perhaps provided by a third-party service) could identify common authors and/or text that is likely to be fake by analysing the content of the reviews themselves. This would help pick up paid reviews that were placed over a long period of time.

gnosis 15y ago
"First of all, I think it's very likely that one of your own (ex-)employees will either go public with the information or blackmail you in order to buy your silence." Why would they go public? They're presumably making (what they consider to be) good money writing fake reviews. By going public they'd be 1 - losing their revenue from you, 2 - undermining the public's confidence in reviews, which would translate in to a less lucrative business for the provider of fake reviews. As for blackmail, I'm sure it happens from time to time. But I don't know how to estimate the frequency with which fake reviewers blackmail their clients. What is your basis for estimating the probability of such blackmail as "very likely"? Furthermore, blackmail is an obvious risk for any secret/illegal undertaking, and yet people attempt try to do things in secret and break the law all the time. Clearly, for them the risk seems worth it (at the time). "Secondly, a sudden increase in the number of positive reviews should be fairly easy to spot by a competitor who has already hired a developer to watch for such things and do automated sentiment analysis on the reviews of competitors." So, the fake positive reviews might be staggered over a reasonable period of time. But even the business receives many positive reviews in a short period of time, that's not proof the reviews were fake. "Thirdly, it's entirely possible that a broader-spectrum monitoring of reviews (perhaps provided by a third-party service) could identify common authors and/or text that is likely to be fake by analyzing the content of the reviews themselves. This would help pick up paid reviews that were placed over a long period of time." Are there any such services which claim to be able to detect fake reviews? Or is this mere speculation on your part? Regardless, just because some service or algorithm claims a review is fake doesn't mean it is. Once again, their guess/estimate is not proof.
- arctangent 15y ago
  > Why would they go public? Sorry, my bracketed "(ex-)employees" wasn't really clear enough. I was trying to suggest a vengeful motive after they had been fired. > What is your basis for estimating the probability of such blackmail as "very likely"? Basic human nature. If you're directly employing people whose job is clearly unethical then you're painting yourself into a corner. Either they do in fact actively blackmail you or else you will feel unable to fire them in case they start telling people what you're up to - I'm regarding this implicit threat as passive blackmail here. I agree with you that there are lots of people who seem perfectly happy to do unethical or illegal things in order to make a living. They calculate (if not always very accurately) that the potential gains outweigh the potential costs. However, a lot of these people make such a decision in light of particular circumstances e.g. their low intelligence, or their lack of employability (due in part perhaps to an existing criminal record), or their unwillingness to work. Someone who has established a business that is making enough money to pay the bills should clearly be reaching a different conclusion to these sorts of people (at least most of the time). Their potential losses are much higher (i.e. loss of the business) and the potential gains are actually not all that favourable compared to doing things the "right way" - for example, by hiring a developer who can boost your business in other ways. > But even the business receives many positive reviews in a short period of time, that's not proof the reviews were fake. It's evidence. What a business owner might do if they suspected a competitor had hired someone to "opinion spam" is up to them. If they though a competitor was paying someone to post negative reviews of their own business then they could very well sue, and request access logs etc. from the review site in order to further the investigation. > Are there any such services which claim to be able to detect fake reviews? I haven't looked. But it would certainly be possible to create such a service without too much difficulty.

gnosis 15y ago

"First of all, I think it's very likely that one of your own (ex-)employees will either go public with the information or blackmail you in order to buy your silence." Why would they go public? They're presumably making (what they consider to be) good money writing fake reviews. By going public they'd be 1 - losing their revenue from you, 2 - undermining the public's confidence in reviews, which would translate in to a less lucrative business for the provider of fake reviews. As for blackmail, I'm sure it happens from time to time. But I don't know how to estimate the frequency with which fake reviewers blackmail their clients. What is your basis for estimating the probability of such blackmail as "very likely"? Furthermore, blackmail is an obvious risk for any secret/illegal undertaking, and yet people attempt try to do things in secret and break the law all the time. Clearly, for them the risk seems worth it (at the time). "Secondly, a sudden increase in the number of positive reviews should be fairly easy to spot by a competitor who has already hired a developer to watch for such things and do automated sentiment analysis on the reviews of competitors." So, the fake positive reviews might be staggered over a reasonable period of time. But even the business receives many positive reviews in a short period of time, that's not proof the reviews were fake. "Thirdly, it's entirely possible that a broader-spectrum monitoring of reviews (perhaps provided by a third-party service) could identify common authors and/or text that is likely to be fake by analyzing the content of the reviews themselves. This would help pick up paid reviews that were placed over a long period of time." Are there any such services which claim to be able to detect fake reviews? Or is this mere speculation on your part? Regardless, just because some service or algorithm claims a review is fake doesn't mean it is. Once again, their guess/estimate is not proof.

arctangent 15y ago
> Why would they go public? Sorry, my bracketed "(ex-)employees" wasn't really clear enough. I was trying to suggest a vengeful motive after they had been fired. > What is your basis for estimating the probability of such blackmail as "very likely"? Basic human nature. If you're directly employing people whose job is clearly unethical then you're painting yourself into a corner. Either they do in fact actively blackmail you or else you will feel unable to fire them in case they start telling people what you're up to - I'm regarding this implicit threat as passive blackmail here. I agree with you that there are lots of people who seem perfectly happy to do unethical or illegal things in order to make a living. They calculate (if not always very accurately) that the potential gains outweigh the potential costs. However, a lot of these people make such a decision in light of particular circumstances e.g. their low intelligence, or their lack of employability (due in part perhaps to an existing criminal record), or their unwillingness to work. Someone who has established a business that is making enough money to pay the bills should clearly be reaching a different conclusion to these sorts of people (at least most of the time). Their potential losses are much higher (i.e. loss of the business) and the potential gains are actually not all that favourable compared to doing things the "right way" - for example, by hiring a developer who can boost your business in other ways. > But even the business receives many positive reviews in a short period of time, that's not proof the reviews were fake. It's evidence. What a business owner might do if they suspected a competitor had hired someone to "opinion spam" is up to them. If they though a competitor was paying someone to post negative reviews of their own business then they could very well sue, and request access logs etc. from the review site in order to further the investigation. > Are there any such services which claim to be able to detect fake reviews? I haven't looked. But it would certainly be possible to create such a service without too much difficulty.