Is MTurk keeping up with the competition?

I've long felt that Amazon's Mechanical Turk is one of the coolest projects undertaken by any of the giants. However, Amazon has been slow to update their online marketplace for work. Two years ago, I wrote a post on this blog suggesting two specific changes: one, I thought tasks should be categorized for the convenience of the workers, and two, I thought workers should be able to blacklist employers. Neither change has been implemented. The process of cycling through the available tasks remains slow and painful. Worse, scammers have increasingly started using MTurk to make money on the backs of workers by using many accounts and not paying even for good work. Amazon's limited enforcement of MTurk has been increasingly exposed and exploited.

Amazon's conundrum is that Mechanical Turk is a small component of its overall business. It makes some money, no doubt, but arguably not enough to justify putting more workers on the project. I suspect that's also why changes such as the ones I've suggested have been slow in coming. This leaves the door open for smaller, more nimble competitors who would actually be thrilled with those profits that seem small to Amazon. "Task" sites like MTurk seem to be cropping up at an accelerating pace; even worse news for Amazon is that they actually seem to be getting better. MTurk is still by far the most popular of the bunch, but if Amazon continues to be complacent that lead may slowly erode.

Right now, the new task sites offer one primary advantage over MTurk: they can pay people through PayPal, which means many international users are able to make money with them. MTurk allows users from around the world to join, but only Americans and Indians can withdraw cash...the rest of the international brigade have to be content with redeeming their earnings in the form of Amazon shopping credit. Microworkers, the most popular alternative to emerge so far, offers another powerful advantage: they actually have customer service that looks after the workers as well as employers. Case in point: I was actually able to get a task that was wrongly rejected reversed on Microworkers. That would be utterly impossible to do on MTurk! Microworkers also has a more sensible attitude towards rejections than MTurk: workers can be banned from completing tasks for a certain amount of time if they have too many rejections but they are given the opportunity to straighten up their ways. On MTurk, if too many requesters ban you in a certain time period, you're gone for good. Bear in mind that some of these banning requesters are themselves fraudsters. While Microworkers is the best MTurk competitor at the moment, myLot is another site to keep an eye on. They've long been a site which pays users for posting discussions, uploading images, and commenting on the news and blogs, but last year they opened up a tasks section to complement their other offerings. Because myLot offers multiple ways to earn, it's an attractive site for online workers who can divide their time between discussions and tasks.

I'm still waiting for the first tasks site to come along with a good categorization system, similar to what is seen on freelancing sites. You don't need to be a genius to figure out many online tasks can be grouped by category: transcription, article writing, SEO (social bookmarks, linkbuilding, and such) , data entry, signups, etc. Wading through huge lists of uncategorized jobs is a waste of time for workers and discourages them from specializing in particular tasks. I think the site that adopts such a system will have a big advantage. I haven't completely given up on MTurk implementing something like this, but I have a feeling it's more likely that one of the hungry young dogs, like Microworkers, will do it.