Crowdholding

Part 1: How do we moderate content?

Part 1: How do we moderate content?

task

EXPIRED


YUP icon 2000


Part 1: How do we moderate content?

Rewards

No YUP/tokens awarded

This is a two part task. Part 1 focuses on moderation rules, while part 2 focuses on how to eliminate/reduce spam all together.

 

What’s amazing by CH is that content produced on our app focuses on high quality. That’s what we are shooting for. Unlike FB or Reddit where people troll with content and can in fact decrease quality through pointless noise of unwanted or worthless content. Here at CH, the content must be of high quality because it’s validating concepts for startups.

 

First we want to show you our moderation rules and ideas we are finalizing with the product team. The idea is to have two types of users for the start. Super Admin and Startup Admin:

 

Startup Admin:

  • 1. Can block a user straight away

    • The User can Raise an Appeal to CH for Review (eg. in case he was targeted by a botnet)

  • 2. Or send warnings:

    • Reporting a user triggers a notification/warning sent to the offending user

    • Warning is also triggered if he /she  is reported based on the Crowd's consensus

  • The Warning

    • Sent as a Notification, potentially also as an email

    • Tells the user that his / her comment was not appreciated and if that happens two more times they will be blocked from the project

    • The votes for that comment are set to 0

  • If reported 3 times the user is blocked from the Project

  • If Blocked by 3 Projects the User is Blocked from the platform

  • The User can Raise an Appeal to CH for Review (eg. in case he was targeted by a botnet)

 

Super Admin (for CH internal moderation):

  • Super can add role of Moderator to anybody….

  • Super admin contact and arrange conditions with Moderator manually

  • Then set up conditions

  • Their Report is as strong as of a Startup Admin

 

So what do you think? Do you see these moderation features being effective? Is there anything we missed or need to change?

We look forward to your responses.

 





30 COMMENTS 72 VOTES

Most recent Most voted Most verbose

Cool

7 months ago

nurul adam s

finance staff of KREDIT PLUS

I think that the new rules in this moderation feature are good, but their implementation should be consistently monitored so that there is no gap for someone to search for and exploit weaknesses in the new moderation feature

8 months ago

This comment is not available, it has been removed for violating the code of conduct

8 months ago

The bitcointalk type moderation may be implemented

8 months ago

joe Šovčík

Experimenter, Goal to understand humans and delight them (I'm human, too )

Can you give me more details please? How does it work?

7 months ago

This comment is not available, it has been removed for violating the code of conduct

8 months ago

This comment is not available, it has been removed for violating the code of conduct

8 months ago

Please do not spam the platform. If you continue to do so, then we have to ban you from the platform.

This platform is built for knowledge-sharing and co-creation with businesses. Getting reward without provide value is not welcomed here.

8 months ago

This comment is not available, it has been removed for violating the code of conduct

8 months ago

Due your comment's short length, we have considered it as a spam, therefore removed it.

If you ignore our warning, and keep spamming, then we will ban your account permanently.

This platform is built for knowledge-sharing and co-creation with businesses. Getting reward without provide value is not welcomed here.

Please read our code of conduct before publishing: https://intercom.help/crowdholding/general-support/code-of-conduc...

8 months ago

jbm386

Signification time spent with enterprise software companies

I would see about adding a block for inappropriate language or links from task givers that could be scams or viruses.

I think adding automation to your moderating with human oversight will simplify things.

I am not sure if you are familiar with discord chat but they have solid admin bots in their platform that I think would be good to replicate.

One other thing you will want to think about as you grow is how do you moderate so the best content / replies get the most tokens. Right now the voting system could easily be manipulated. As YUP grows in value more people try to cheat the voting / task system (similar to how people cheated the referral promotion a little while ago) which will ruin the integrity of the platform and turn off your customers / project owners who you sell to.

I love the platform have been contributing for a while now and want to see it succeed. Please make sure you are thinking about how to address these types of issues because a few bad apples could easily ruin this for everyone.

8 months ago

LeeKe Mac

If I make a comment - its usually honest and thoughtful. #dontfeedspamtrolls

I do think these moderator rules are good. But they could be more robust.

For example - would moderators be responsible for checking comments for plagiarism, duplicate accounts, duplicate responses or just be used for checking on comment quality?

As long as the admins are consistent in their approach to the reporting and the system is visible to all users then those users who respond without spamming, irrelevant one liner style responses or content farming posts will not need to worry too much about the moderation.

Also providing users with warnings that give them feedback on their comments is a good idea. Again - these should also be a quality feedback comment to avoid confusion.
ie:
Warning for low quality post
vs
You post has been considered of low quality because you did not provide an answer or discussion regarding the topic as asked.

Another option for the reporting system:
It may be a risk as bots could use this to give potential false positives - but why not use a community report section as a "downvote system" where a typed reason must be given for the downvote (ie, not answering the questions, content farming, short answer response with no relevance etc). The downvotes could be checked and if they are valid the "downvoter" could receive a YUP for helping to verify the system, and the "downvotee" could be stripped of their votes. If the downvote is not upheld (via gaming the system), then the "downvoter" could lose YUPs?

8 months ago

You missed the whole point of incentives. You incentives system sucks. If you set the system up so it ACTUALLY rewards good content and the people who vote for that content you won't have ANY problem with spam, because nobody will spam for nothing.

8 months ago

Ethan Clime

CEO of Crowdholding.com

Yes, agree rewarding voters as well for doing good content. We started planning moderation techniques after studying and coming up with solutions from a trial where we increased users from a referral which spamming increased significantly. From that we had to manually moderate. You make a good point. The question is, we have seen people write "Good job" and their friends upvote. So what's your thoughts on preventing this? Currently we are blocking this manually, but this won't be efficient as our user base grows. So our thoughts are to encourage project creators and users to help moderate.

8 months ago

Your system encourages gaming because it is trivial to do it. Going the blockchain way doesn't solve the spam issue either, you can check the quality of most steem posts for reference.

The most glaring issue I encountered and have cultivated an intense dislike for is the way you calculate rewards. For instance if I post something in a thread and then I like some other things in the same thread I effectivelly dilute my reward, which would be suboptimal game theory play for me. My suggestion was to calculate reward from a hidden budget and only display current reward that goes up the more good content there is in a thread and the more likes it has.

For low quality content there is currently no other way than weeding it out manually. You can try increasing the minimum post lenght to something twitterlike, but that will discourage only the laziest spammers.
The effective way of doing it is by using penalties. For instance, there are a few good job posts with a few votes up, a mod flags all involved and detracts their yup, if at 0 banhammer, not just the thread but whole CH, not just an email/name ban but IP.
Less invasive way of disincentivizing spamming is using weights, it can be done with blockchain so it is a venue worth exploring. The steem way where you weight as much as the number of your steem spawns whales that then sell this power. If you weight people by the time they joined, quality and number of their posts you could potentially reduce the new account spam greatly. It doesn't eliminate people selling their votes, but at least their weight would be earned by a time investment, not a financial one.

You can't expect all people to be honest, but most of them will be. The few rotten ones will always have to be rooted out manually. It is however, very important that you have transparent and definitive rules about moderation. Only posts like good job should be penalized. Expressing opinion, even if extremely hostile is OK

8 months ago

Ethan Clime

CEO of Crowdholding.com

Cheers. I see where your getting at and I like your logic. Forwarded your input to our product team, and will be looking into this (especially) the topic of calculating rewards.

8 months ago

0 Votes YUP icon 0
Hide sub-replies
Zexro Indra

former mutual wealth advisor for indonesia

moderation feature is pretty good.
in the meantime, I do not think there is anything to change.
can add anti-bots and spam links

8 months ago

Zexro Indra

former mutual wealth advisor for indonesia

it took a lot of human moderators to realize this dream.
very ineffective.
features turn off notifications and summon that has been so far been quite effective

8 months ago

This comment is not available, it has been removed for violating the code of conduct

8 months ago

Hello, Sarman. According to the Crowdholding Code of Conduct, I am informing you that your comment is considered as a spam.

Please keep in mind, this platform is intended for co-creation between the public and businesses. Comments which produce values to the topic of task deserves vote, hence get token reward. Your comment cannot produce any added value to the business, hence it against our community's ideology. Please provide useful and high quality content.

If you are still not clear with our code of conduct, please read: http://help.crowdholding.com/general-support/code-of-conduct

8 months ago

I think it's been very effective, reduce spam coment, spam links or koment via bot. Many other platforms are lacking in this application, so Crowholding being a clean platform from bot comment, spam comment or link, it's very good

8 months ago

This comment is not available, it has been removed for violating the code of conduct

8 months ago

This is a great idea, community should give a high quality responses to earn rewards. But I think you don't need to blocking a usee right away. Sometimes you need trolls to send or introduce you to a bigger community. Start up admin can just give 3 warnings then block. Will you be posting the guidelines before implementing it?

8 months ago

Ethan Clime

CEO of Crowdholding.com

Yes will definitely make improvements from the feedback and share this in an update, and if still needing some more input, a task.

8 months ago

Thank you for your answers

8 months ago

0 Votes YUP icon 0
Hide sub-replies
John Dexter Villapana

Im very curios on how things work

I think should not be block right away. At first there should be a warning.

8 months ago

Randy Tice

Just take a look at my past contributions to projects; you will see value there.

My thoughts on both:

Startup Admin: Definitely go with the warning system. Giving startup admin the power to block right away isn't necessary at this time because CH doesn't have that level of out of hand spam yet.
For warnings, I think 2 warnings and blocked from project is better instead of three. With the userbase still small, you rarely have people commenting more than once. Blocked from 3 projects then blocked from CH is good (three strikes your out thingy).

Super Admin: Of course, parameters need to be set for who earns Startup Admin title. To stay true to the CH philosophy, Super Admin should crowdhold the idea with the user base. This way, the "hierarchy" can be set up with as much consensus and input as possible. Then, going forward, the Super Admin can directly work with those who meet the criteria to become a Startup. Giving Super Admin the same strength in reports as Startup Admin is good, there doesn't need to be any unnecessary "Power Wielding" in the CH platform.

8 months ago

Sounds about right. My only question is what are the parameters of the moderator selection process (Super admin first bullet point)? Are there clear guidelines for the selection? The composition of the moderation panel will be critical to the integrity of the project.

8 months ago

Ethan Clime

CEO of Crowdholding.com

Not yet officially clear. Super admin will be from our internal CH team and for the time being selecting top users to allow them to moderate. The concept of giving more power to future legendary users, but identifying this into rules is definitely needed.

8 months ago

I assume that this is the right thing to keep the quality of CH content intact, so that the partners who introduce the Platfom still use CH as a means to introduce their platform

8 months ago