Building a Community of Authentic Content Creators
Fraud Prevention and Quality Control
As detailed in our whitepaper, the Vybe mining app tokenizes likes, or, more accurately, content quality. Once a day, it aggregates all the content quality metrics (like hearts, upvotes, thumbs-ups, comments, retweets and hashtags — “likes”) that you get on your content creator sites, converts these into a token called VYBE, and automatically deposits VYBE into your Algorand wallet. This makes it possible to better capture and control the value of your content and convert it to other assets. You can mine your social media, and get paid to post.
The app uses an algorithm to do this. The main variable that the algorithm uses to convert content quality into VYBE is likes. The more likes a user has (in general and up to a point), the more VYBE they get.
Frauds, however, can buy likes on some social media sites, like Facebook, Instagram, Reddit and Twitter. Several sites, for example, sell 100 likes on Instagram for $3. Frauds can buy these quickly and cheaply to make their reach and content appear more valuable than it is. This misrepresents influence and manipulates followers. It can also get frauds banned by the social media sites, but some creators are tempted to put on this disguise and build this Potemkin village because some of their competitors do, and the more friends and influence they appear to have, the more money they can make in branding deals.
Frauds can thus potentially buy likes to misappropriate VYBE. This is a reality and vulnerability that we have to counter. If billion dollar social media companies, however, can’t block bought likes, eradicate bots and ban more frauds on their sites, we can’t solve this problem for them at the source. We can, however, protect, vet and clean our site to minimize cross-contamination. We’re using the following five counter-measures to do this:
1) Metrics. In its most basic form, the algorithm that the app uses to convert content quality into VYBE is:
vd = (ucq/tcq) * 890,411
Where:
vd = VYBE distributed. The amount of VYBE distributed every day.
ucq = User Content Quality. The total number of likes, posts, comments, retweets, subscribers, hashtags and other engagement a user gets on their social media sites that day.
tcq = Total Content Quality. The total number of all likes, posts, comments, retweets, subscribers, hashtags and other engagement that all users get on all sites that day.
We make several modifications to this equation to measure content quality more accurately. We cap gains and weigh variables, for example, and also count variables like hashtags to determine content quality. This limits the impact that buying likes can have. We also continuously review and refine this algorithm to improve accuracy.
2) Cap. No one user can earn more than 1% of any mining pool. Since the app distributes 890,411 VYBE once per day, no single user can mine more than 8,904 VYBE per day. This cap prevents a user, like a celebrity or fraud, from mining or misappropriating the vast majority of VYBE. After hitting the 1% cap, more likes don’t produce more VYBE.
3) Economics. Currently, VYBE cost .0006. At this price, it costs more to buy likes and convert them to VYBE than to buy VYBE directly. Frauds are better off buying VYBE on an exchange than buying likes to mine VYBE.
That said, the more valuable VYBE becomes, the more economic incentive exists to buy likes. This, however, is also countered by the fact that the more valuable VYBE is, the more miners VYBE likely has, and the more miners the app has, the fewer VYBE each miner gets — but the more valuable each of these VYBE are. These conflicting market forces collide to determine incentives and a moving Fake Friend Price (FFP) — it isn’t always profitable to buy likes to try to misappropriate VYBE.
4) Verification. We verify users with a ranking system:
a. Not verified. This includes all new users who connect their creator sites to their Algorand wallet. A new user, however, will have to solve a captcha to register, and verify access to their creator sites by confirming receipt of a DM we send them to that site. So while they’re not verified, they have proven that they’re a human with an Algorand wallet, a verified email address, and access to at least one creator site, which, theoretically, is also routinely vetted for authenticity and ban-able by the host site.
b. Verified Level 1. Users submit photographs of themselves holding a printout with their username and the day’s date. A Vybe Quality Control officer vets the photograph and inspects their sites to ensure they’re posting authentic content with authentic engagement.
c. Verified Level 2 — x. Vybe plans to add additional verification levels. The more verified a user is, the more valuable they’ll be to promoters, and the better bounties and branding deals they can acquire.
5) Vetting. Once the app launches, we will use Quality Control officers to vet users. They will use assessment apps to help identify fake followers and fake likes. Some apps, for example, flag posts that get more likes than the user has followers. The Quality Control officers will focus first on the top creators on our leaderboard. We will ban users who buy engagement and manipulate followers.
Like buying and friend farming is fraud we have to face, like other challenges platforms have, but the more creators we have with quality content and genuine engagement the more useful and valuable our platform is. Given this, we’re committed to combating manipulation with quality control, and have established these measures to do so. Our goal is to measure content quality as accurately as possible, in order to build the best community of authentic creators we can. We want real friends.
