To continue growing Mechanical Turk, we are analyzing possible changes to our commission structure for Requesters. These changes would be intended to allow us to increase our investment in the marketplace and bring future innovation to Mechanical Turk that will benefit both Requesters and Workers. Mechanical Turk has not changed its commission structure since launch in 2005. As part of this, we are considering changing our base commission to somewhere between 20% and 35%. Today, our base commission is 10%. We will share the outcome of this analysis in June.
Who wins? Amazon. Who loses? Researchers who won’t be able to afford to use MTurk, or who would take the money they would pay Turkers to cover the increased costs. Turkers also lose if there are fewer HITs to complete and if Amazon continues to ignore wage issues.
Here is some mathy stuff:
I am about to start a study, I want 125 people and would pay .75 for a five minute study. My budget is $100. That’s about $94 for the data and another $9.40 to pay MTurk: the total is a bit over $103.
If the fees go up 20%, that means I could pay 120 people ($90) and then $18 to MTurk for a total of $108. A bit over budget but not bad. But if they go up 35%? I would be able to pay 100 people ($75) and an additional $26.25 to MTurk. That’s a pretty major decrease in participants.
More importantly, what will Requesters (and workers) get for these increased fees? The post I linked above cited some improvements that MTurk has made (Smartphone HITs! You can screen out people by state!) but I’m not sure that’s worth the increase.
I’m wondering if this is because Requesters rarely use Master’s workers (anybody know)?