Ecommerce fraud is a huge space, with “over $7 billion to chargebacks in 2016. By 2020, e-commerce chargeback losses are expected to balloon to $31 billion.” False positives, chargebacks, shipping costs, and fraud-prevention are a big problem.
Merchants are losing much more than the replacement cost of goods lost to fraud. They’re also losing revenue to the shipping and insurance costs associated with the fraudulent transaction. Unfortunately, it doesn’t stop there. Merchants surrender even more revenue to the investment and operational costs of fraud-prevention solution, manual reviews, false positives and chargebacks. FInally, settlement can become a long process, which could be solved via a blockchain solution.
The goal of this project was to answer the question: Can we crowdsource the task of settling disputes to a properly-incentivized community to ensure a smoother transaction marketplace?
Influences: Looking at Witnet (proof of witness/oracles), zksnarks, and other solutions, is it possible to prevent someone from Sybil attacking and simply voting through all of their disputes?
Considerations: does someone have an incentive to create multiple accounts to vote? Buiding a KYC/AML
Can you anonymize or edit pieces of the story without changing the decision?
Chargeback fraud: a customer claiming they didn’t receive the item
Reasons for dispute: damaged item misdescribing it Sent wrong item
Considerations: Startup costs: is it reasonable for someone who is new to the network to get a dispute.
PROBLEMS TO SOLVE:
KYC/AML is really important Resolution Time
- Someone lying would also know when the dispute happens. Someone who isn’t lying would treat additional information like the exact amount of the decision,
- The trick: Make multiple duplicates of the same causes Sample case: if there are 10 versions of it, and it’s a blatant case of lying, the cheater would have to vote incorrectly on both.
Voting weight grows over time: