And why isn't Amazon doing anything to further the project. Stuff like, oh, I don't know... Setting-up International bank debit-card or wire transfers. Improving their interface. Adding a way for Workers to CHALLENGE REJECTIONS, fer Christ's sake! I think that even in it's current broken state MTurk is potentially quite lucrative for Amazon, so why aren't they devoting any resources to the project?
Challenging rejections would require an appeals process, of which the requester would probably have to reply to. This would make mturk ineficient for requesters and probably a lot of requesters would leave.
Fair-enough, then how about a requester ranking system based on Worker feedback. Complete a hit, and 48 hours later you can submit your 5/5 or 10/10 ranking of the requester, a ranking which you can then change at any time. Would that be so hard to implement? And what about those other issues I've stated? Honestly, MTurk is such a HUGE potential money-maker, so why is Amazon just sitting on it? Is it because of those new anti-outsourcing laws I've been hearing about? Is it because Amazon fears that if MTurk becomes TOO popular they might risk instigating the wrath of IRS? All leading questions, I know. But I would still like some answers ...or, at least some well-informed rumors
Amazon doesnt have a need to change a thing.Improving the site would only mean more hastless for them,now they just have to sit back and relax..and money money money...Wetr they change the ways or not...each day more and more turkers are coming to the field...(me too...he he).. Hope they will show some mercy,once in awhile.....
There's No Money In It I think the reason that Amazon's not aggressively moving forward with MTurk is that it doesn't make very much money for them. As a quick experiment, I pull the top numbers of HITs and tallied them up. Of 276,489 HITs available, the top 166,221 were divided into just 10 groups of HITs, with a total value of $10,602. Assuming Amazon adds a 10% charge on each HIT, that means that Amazon stands to make about $1060 from HITs that comprise 60% of the work available on MTurk. I next pulled the top numbers of hits that paid at least $1.00. There are 2555 HITs available in the top 10 with a total value of $4263.35, or a value of $426 to Amazon. I'm going to make a rough guess and say the total value of all the work on MTurk is probably around $30K. If they turn 50% of that over every week, they're probably making about $6K a month with MTurk, which would just about cover the salary of one full-time entry to mid-level developer working on the project. Even if my estimates are low and there's say, $60K worth of work and 80% of it turns over every week, that still works out to about $19K per month to Amazon, which might justify 2 full-time developers on the project, plus a part-time project manager. The economics to put a lot of resources behind MTurk just don't seem to be there.