Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

It's the natural cycle of the internet.

1. Be a nobody company

2. Utilize generous user content to build up an important knowledge-base/economy

3. People begin to trust the information in the system and come to rely on it

4. Sell-out entirely and make a lot of money letting people buy influence

5. People abandon the site realizing it has no signal anymore

(Digg, Reddit, Amazon, Glassdoor, experts-exchange, facebook, google search ~ish)

Notable exceptions: Craigslist, HN, stackoverflow, wikipedia, youtube



You're missing step 3.5, where as a company gains trust and notoriety, more and more 3rd parties will attempt to game the system to their favor. That's where most trust can end up lost. The moment it becomes worthwhile to "buy upvotes" there will be trouble.


Nicely summarized. I see this pattern playing out so frequently that I'm at the point where I'm starting to think no one bothers creating content or communities without an eventual commercial angle that ruins the intent of the first iteration.

In terms of your notable exception, I'd say Indiehackers (now owned by Stripe) is still quite good. But perhaps that's because Stripe appears to be pretty hands-off at this point, and of course that could change.


> without an eventual commercial angle

If you're not planning on selling your personal projects, then they're hobby projects.

Once a hobby project no longer becomes a hobby project, it becomes a commercial endeavour.


What do you think is the difference in the companies that lose trust and that keep it?

Would you agree with this idea: the companies that keep trust figure out a business incentive that aligns with keeping it?

for example - HackerNews wants the best content to attract smart people to apply to YCombinator, which is where they make money - Wikipedia is a nonprofit, relying on happy users and foundations to donate - Craigslist is effectively a Public Benefit Corporation; They have a noncorporate noncommercial ethos as a closely held entity, and choose not to profit maximize

What do you think Glassdoor should do to align their business with having accurate and trusted reviews?


I feel that youtube is on the edge here; still lots of good information, but under heavy pressure.


And the noise is quickly overwhelming the signal with the recommendations. It's hard to see at a glance what's new with the channels I'm subscribed to and actually want to see.


I mostly agree with your analysis, but I think your examples don't actually fit. Please show me all the people abandoning google and amazon.


You forgot Yelp. :-)


Re: the exceptions, I think they're all on that cycle, just sitting at step 3 (or borderline 4 in the case of YouTube).


HN as an exception? Astroturfing seems pretty common here.


It's not exception, I've been trying to ask some questions about Coinbase (YC company) in "Who Is Hiring" post, they all got removed. I reworded these comments multiple times, after ~10 attempts I gave up.


You reckon?

HN gives off almost a courteous Victorian gentlemen's club vibe it seems to me. Blatant astroturfers are given the silent treatment or shown the door.


Whatever astroturfing happens comes from uninvited guests, HN itself isn't trying to sell anything, aside from providing a community that is useful for YC startups, and in general I think the mods and the community do a good job keeping away the spammy stuff.


Hmm. How do you buy influence on Reddit or Amazon?


Do you trust the reviews on Amazon? Especially from cheap Chinese suppliers? Those reviews come from somewhere.


Again - yes, the reviews can be manipulated and they definitely ARE. But that is not in any way encouraged by Amazon as it doesn't get anything from that. "Sell-out entirely and make a lot of money letting people buy influence" - doesn't apply here


I am not sure about this, but could imagine it being the case that Amazon has done some analysis and found some fake reviews for counterfeit or legal knock-off items to be in their interest.

If the average customer believes that a lot of positive reviews means the quality of the product is high, they are more likely to buy the product. As long as the product is a least decent, they will probably be happy with it. If they are unhappy, they can complain and Amazon will refund them to make them happy (this is probably rare so not a huge impact on profit). Amazon gets to keep lots of the margin on the lower quality knock-off items, instead of the original IP owners. If the item really sucks and they get a lot of returns, Amazon can then use that as a signal to bury it.

If Amazon produced millions of fake reviews themselves, it would be a huge scandal. But if thousands intrepid sellers invest in making thousands of fake reviews, it’s not as much of a scandal for them. Amazon makes a show of fighting the most obvious fake reviews, so people will continue trusting the platform. But are they really using their top talent to fight the fake reviews? I work on a ML / data analysis team and I feel like we could do a much better job at that task than Amazon seems to be doing, and they ought to have more resources than we do.

I’m not accusing Amazon of doing this strategy intentionally, but perhaps there is an understanding that the ROI of engineering talent spent fighting fake reviews and counterfeit items is low or negative.


Not sure about Amazon, but this website for Reddit and Quora looks pretty easy. https://upvotes.io


Right, but this doesn't look like Reddit or Quora selling-out, these are just users exploiting the platforms


On Reddit, sub mods can be bought, and many are.


As I said above - that's not actually item 4


Well, I suppose you could apply the same to Admins, that would count. I don't know of this happening, but then again I don't think we would have any way to discover it.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: