I just made a similar post on the Introductions forum. I think this question is suitable for here as well.... I'm new to mTurk and am looking for the most efficient strategy for making the most money in the shortest amount of time. (Isn't Everyone?) I am aware that it's not a cash cow, but I figure a goal of about $4-$5 a day will help me pay some bills. I think this post could become very useful if everyone made a detailed list of their daily turking strategy, coupled with the successes and failures they are experiencing. Obviously, scripts that would streamline the process would increase productivity tenfold, and yet there doesn't seem to be any publicly available. Do we have any programmers on the board? For now, let's focus on raw Turk strategy, sans scripts. What are you doing on a daily basis to make money, the fastest way possible?
But of course! If Amazon Mturk moves to some sort or easier way for the hit types/requesters you like to be easily available, then my productivity would go up. Also if something was done directly though Amazon, similar to Turkopticon, that workers could approve and reject requesters, the requesters might get more trustworthy. We can all dream can't we? Amazon Mturk is artificial artificial intelligence. And they don't want workers using anything that automates the process of doing HITs. Its against their rules, and they have countermeasures against such. However, scripts that allow you to do HITs easier can be employed using Greasemonkey for Firefox. If Amazon thinks you are doing something wrong, the first 'warning' is the 5 minute lockout. They automatically return all HITs, if any, assigned to you but not completed. And for that 5 minutes, do not allow you to accept any new ones. They will also warn you when you are overly active(have a pretty high page view rate). Personally, I just hope for good HITs. I don't want to do hundreds of penny HITs, except if they take less than 6 seconds and are fun. Some days I just end up lucky, checking the site at the right time. Those would be the days above $10. I just keep looking for the requesters I like, and sometimes I go though the HITs sorted by money. My best day recently was when I combined two tasks in a trip to the grocery store, both were for $5, and to me the value for the total time/resources needed on each, was on the low side. The number one thing I think that drops the total $/hour, is waiting around for new HITs. You could spend multiple hours and only make a single dollar. If you could type really fast, and can think even faster, then there's always plenty of writing assignments. Just make sure to check out a requester using the Turkopticon plug-in, as to not waste your time on bad requesters. And finally, this site is a great place to check for the good requesters. I would have missed quite a few good HITs, if it wasn't for the others willing to share them here.
Thanks, that was a great response. I'm going to check out some 'helpful' scripts when I get home from work to see if they'll support my cause. Anyone else have any insight on their strategy?
My strategy in this case is: - Tagline for Confessions of a Dangerous Mind that I feel fits well here, at least (if not only) for me Hope my strategy proves to be helpful.
Ouch, I feel like I just didn't get invited to the cool party. No, but really. I think a brainstorm would benefit everyone.
The problem is there's not enough good HITs for the amount of workers. Amazon can easily get hundreds(if not much more) new workers daily. The problem is there's not enough good requesters coming in. Sure there's more of the 'click my ad to test it', 'test my sign-up', and 'Insert a valid email on this site, and send me the email response'. But those are not even allowed to be on Mturk, yet they exist. I don't know how many times a requester can get reported before something is done, but a lot of times I feel the reports end up in /dev/null. If a good requester puts out 40,000 posts, it's a mad house trying to do those as fast a possible before they run out. There are just too many workers for the amount of work. This also suppresses the need for requester to offer a fair payment. IMO, the number one thing that would benefit people, is if Amazon made improvements. The second thing is get a lot more half-decent requesters. There are plenty of things that could be done on Mturk more economically then however they are currently being done. And there's so much old outdated data out there, that could be refreshed for such a minimal cost. So with the current conditions, and with so many people turning to any extra way they can make some money, some people just want to protect their good thing. I don't tell anybody about Mturk that would probably end up as a worker. Though I do tell people who could use the benefits of Mturk as a requester.
The best days have been when there was a huge dump of hits. Like the AAAS transcribe student comments I did like close to over 300 of their hits and made 28.38. I average about 3-5 dollars a day. When a big dump from a requester comes I try to jump on it such as venue quality. I like the na=amazon books they do huge dumps and pay .01 and it is usually yes or no question and I can make like close to 13 dollars with them. But that is mixed in with a lot of other hits I do in between.. I do a lot of the .05 hits and Andy Dickerson who does .04 sentence rewrites. I do about 3 game reviews a day for .20. I do a lot of the web service .05 hits. I check daily to see if their are any new qualifications out and try to grab those also. the most hits I have done in a day are 735 hits. I usually can do about 200 or so a day. I have slowed down quite a bit though lately.
it depends if there is a big dump usually 2 hours but I do easy ones first. but if I am searching for them it takes forever.
So, last night marked my 7th day Turking. I have averaged $1.28 a day. I would have done much better, but I partied all weekend in Montauk! So far, my strategy has been to search for easy, repeatable HITs for between $.5 and $.10. I was considering doing some of the higher paying HITs, but my math proved that I could make more per hour cranking away at the cheap HITs rather than doing only a few of the bigger ones. I'm going to try to get my average up to $3.00 a day by the end of my 2nd week. Oh, and for posterity, I've been turking an average of 2hrs a day.
As you get more into the swing of things, and develop a sense for what kinds of HITs appeal to and work best for you, I think you should probably be able to get up to minimum wage, or pretty close if you put the time into it. Some workers manage to exceed that, but in general I doubt there are many people really making a living. As supplemental income though, MTurk is great to have around. I tend to wait for the rarer times when I get something I can make more than $10+/hr doing. For a little while I kept getting big HITs and was making over $100/wk for about 8 hours of work. That's going to be the exception rather than the rule, though.
Ok, so I suppose this thread will become my weekly strategy update.... After my second week turking, I am now averaging $2.16 a day while continuing to turk about 1-2 hours per day. I've brought my average up $.88 per day in one week. I believe this is mainly due to working out the newbie jitters and increasing natural efficiency. I have done some things differently this past week that may have resulted in the raise: - Added scripts (greasemonkey, turkberserk, etc) - Actively monitored which Requesters paid best/quickest, and started my workday off with their HITs. - Filtered my searches by "newest first". My goal for the next 7 days is to reach an average of $3 per day. Suggestions?
Quite simple. Just remove - from "turk about 1-2 hours per day". If you do so, you may be able to achieve far more than your current goal, that is, an average of $3 per day. And yeah, I accept your thanks in advance; so you need not mention thanks to me.
Try increasing your time a little bit. Also if you see a decent .20 hit take that. You can also combine a few of the .06 and .07 hits that come out just about daily. Try to get some of the ones that offer a bonus also that helps a lot also.
Accept any HIT that matches your qualifications & pays fair enough taking into consideration the time & effort you need to put in. It (HIT Reward Amount) does not need to be specifically (exactly).20 HIT. Accept the HIT that matches the specifications I've mentioned in my first sentence here; no matter whether Reward amount is greater than, less than or equal to $0.20. Additionally, don't just accept any of such HITs, but work it out as per the instructions given therein & then SUBMIT the HIT, too. Just accepting the HIT won't do any good, executing the task assigned therein & submitting the HIT are of vital importance [& mandatory if you wanna get your HIT to be verified (approved or rejected, whatever it is) by the Requester] too!!! Be specific!!!
I wrote a four page guide on MTurk strategy, actually. I hope it helps you. http://www.associatedcontent.com/article/1959176/making_money_with_amazons_mechanical.html?cat=9