First, if they are using Turk then they need to build in a feedback mechanism, so instead of you manually removing the tags from Flickr, which is worthless from a feedback loop perspective, they should instead have you thumbs up/thumbs down the added tags via their site. This would allow them to gauge whatever system they were using as well as not charge you for that photo or pay that Turk cost, which is the other end of how Amazon’s service works. The feedback needs to trickle all the way back down. In time the people that couldn’t tag well wouldn’t be allowed to signup for TagCow’s Turk actions.

Second, you said, “But tagging is still a very subjective thing and getting someone who sees my photos exactly the same way that I do may be a tough job to assign.” On one hand you’re trying to offload some of your own work, which is a great idea, but on the other another reason to have someone else tag your photos, even ones you’ve already tagged, is to get other people’s perspectives on your work thus opening up the search field. Since art *is* so subjective someone may use a tag that you’d never think of and that is worth quite a bit if it generates some work or a purchase.

It seems the business model could actually work very well if they built in a mechanism to improve the quality of their service. If the tag quality rate went up from say 50% to 80% then I can see a lot of people that are just jumping into world of tags wanting a little head start.

Another thing that is subjective is the whole tag in quotes vs. single word tags. The only time I use multiword tags is for proper nouns. It would be nice if they could tighten up their tagging guidelines for the Turks. A tag like “girl in blue dress” is pretty worthless in a tagging system, even if they had quoted it properly.

The Mechanical Turk users make 4 cents per PHOTO, but you get charged 4 cents per TAG.

That said, reviewing the TechCrunch article they seem to imply that the MTurk users actually get 4 cents to properly tag a group of five photos.

In the first instance, so long as more than one tag was added to a photo tag cow would make money. In the second example, they’d be raking it in.

That’s one of those little things that calls the expertise of anything/anyone into question.

Having your Flickr photos tagged is one thing, but having ALL photos in your LIBRARY tagged is what would really be valuable.

Of course, that’s what staff is for.

And at the end of the day there are no landmark recognition tags, which would be important for me. Although they did do a good job on text recognition for me, so I had a different experience to you on that one.

As for the pricing though….jeez, they’d want to charge me more per year to tag my photos than Flickr charge to host them. I could tag them, better, for free. I couldn’t host them better than Flickr do.

I also doubt this business plan will work. But I’ve thought about something like this earlier and I came to this solution:

On a photosharing site where you can also sell your pictures (shouldn’t this also be possible on Zooomr?) you can tell the people that if the tag others photos they will earn a bit if the photo got sold. Optionally, you could say “if it got sold because it was found by your tag”.

I think this would motivate the users to add some tags to the photos they look at and additionally wouldn’t cost the service that much.