3 ms·
In my opinion, by linking a person's gittip account with their public persona (twitter or, worse, github) and then calling a person out as "suspicious", you are
by nathanb 14y ago
In my opinion, by linking a person's gittip account with their public persona (twitter or, worse, github) and then calling a person out as "suspicious", you are doing your user base a disservice. If I was trying to do the right thing by tipping someone I appreciated and then not only did you tell me I was suspicious but you also told the world, I would not be passionate about how great your service was. At the very least I would have a bad taste in my mouth, and at the worst I would likely flee entirely.
At a scale of a million people, 0.3% is 30k. Can you imagine a human trying to parse that page at that scale? I can't. You wave the magic phrase of crowdsourcing at the objections, but what person has the sort of energy, patience, and free time to sift through all that? Paypal can't hire enough folks to do that sort of thing; what makes you think you can convince them to work at it for free, month after month, year after year? Oh, you might gittip me fifty cents if I do that for you? Well, OK then.
I understand that the last thing you want is for gittip to become a money laundering site. If news got out that your pet project is a great way to get funds out of stolen credit cards, that could easily destroy you completely. Perhaps avoiding that scenario makes the measures you've put into place worth it. But it feels like you've limited both the rate at which you can grow and the scale to which you can grow.
Every now and again you read about someone's Amazon account being closed because it gets caught up in their automated scam dragnets and the customer service drone assigned to their case is monumentally unhelpful. Do you think that a system similar to yours (aggressively flagging accounts, exposing presumed fraudsters, relying on community volunteers and public transparency) would make their system better, at their scale? Do you think they could have reached their current scale if they used such a system?
- whit537 14y agoI'm open to alternatives. What would you suggest?
- whit537 14y agoThe primary goal is to keep stolen money out of Gittip. The secondary goal is to recover stolen money once it's inside of Gittip. There are other kinds of fraud to watch for but let's focus on this for now. If Gittip does nothing, then stolen money gets into Gittip, and some portion of that is reported to the banks, and Gittip gets hit with chargebacks. Chargebacks are financially onerous, and if Gittip gets too many of them then we'll lose our privileges with the banks altogether. Moreover, chargebacks take months to hit. It's neither economical nor efficient to use chargebacks as the primary signal for bad accounts. Therefore, I don't see a choice but to try to predict which accounts are going to inject stolen money into Gittip before it happens. Once stolen money is in Gittip, some of that goes to good accounts and some to bad. Either Gittip makes an effort to recover that money somehow, or we simply factor the entire stolen amount into our cost of doing business.