MicrostockGroup
Agency Based Discussion => Shutterstock.com => Topic started by: pancaketom on April 22, 2011, 12:34
-
I just noticed that the SS darkroom now shows the keywords used to find the images.
It shows the top 5 keywords used and the percentage for each of them. No real surprises in my short look at it, but many many of the searches seem to be just one word as I suspected. Not only because the chances that a multiple keyword search is used enough to show up in the top 5 is lower and because I think buyers tend to try one word, then if they don't see an image they like in the first few pages then they try more words.
In any case, this is more potentially interesting and useful data from SS. Thanks.
-
Thanks for the heads up on this. Interesting, but like you said not real surprising what keywords are used.
-
I like this.
-
very cool. Thanks for the heads up. A few surprised for me - it is very fun to see the whole list of used keywords and their %. i think it will be quite useful for keywording images in the future.
-
That's great! So helpful. Some surprises in there for me too. I'm looking forward to seeing what other treats they have in store for 'The Darkroom'. I see their speeding up the message boards too (currently down for maintenance). I love Shutterstock :-*
-
I love it! Quite a few surprises, including the number of scientific searches.
-
Neat. I see though a few of mine hit mainly on two words. Majority was one word though.
-
Thanks for notifying us. Very interesting and quite a few surprises in there.
More evidence to me that Istock's CV simply does not allow the buyers the flexibility of the full vocabulary. Several of the main keywords used to find my images would either not have been permitted or would have been side-tracked to a different meaning.
-
Very interesting! I hadn't tried out the Darkroom feature before. Thanks for posting :)
More evidence to me that Istock's CV simply does not allow the buyers the flexibility of the full vocabulary. Several of the main keywords used to find my images would either not have been permitted or would have been side-tracked to a different meaning.
Same experience here. It's frustrating to see how many potential sales I could be missing at Istock because the top searched words on a given image may not be part of the CV.
-
It's frustrating to see how many potential sales I could be missing at Istock because the top searched words on a given image may not be part of the CV.
That is so true. But once you've committed to something like CV there's really no going back. The thing that convinces me it's a loser is that no other microstock sites - even new ones - have gone down that road.
Like many things about IS today, the CV probably works fine for the big sellers of mainstream subjects - and all the niche stuff just goes over the side.
-
I looked at it a bit more and there are a few things that are confusing to me, for instance these keywords were listed for a file with only one download...
central (33%); central park (33%); park (33%)
does that mean that the person was searching on "central, central park, park" or was the search for " "central park" central park " ?
(as an aside, if I was searching for an image of central park in NYC, I'd have considered this image a spam, since I think they keywords I actually had were "central oregon" and "smith rock state park" - I guess the buyer wanted it though.)
In any case, there were a few surprises, but I'm not sure quite how to take advantage of this info. Unfortunately we can't see the searches that didn't show the image but should have (missing keywords). Although maybe looking through the top 100 keywords might help for that.
I do like the features they are putting in the darkroom section. I only wish they would include the files with 0 sales, especially while sorting by upload date.
As far as the CV on IS goes, I always thought that "none of the above" should be a viable CV for any given keyword. So that way I could have Mount Helen - that isn't Mt St Helen.
-
I think it is a bit like Dreamstime, where I recently sold a picture of some wooden flowers with the keyword "madeira". Not at all relevant, and I certainly didn't have that in my list of keywords. But it was the latest keywords the buyer had used, before he found my picture by visual search.
-
I looked at it a bit more and there are a few things that are confusing to me, for instance these keywords were listed for a file with only one download...
central (33%); central park (33%); park (33%)
does that mean that the person was searching on "central, central park, park" or was the search for " "central park" central park " ?
(as an aside, if I was searching for an image of central park in NYC, I'd have considered this image a spam, since I think they keywords I actually had were "central oregon" and "smith rock state park" - I guess the buyer wanted it though.)
In any case, there were a few surprises, but I'm not sure quite how to take advantage of this info. Unfortunately we can't see the searches that didn't show the image but should have (missing keywords). Although maybe looking through the top 100 keywords might help for that.
I do like the features they are putting in the darkroom section. I only wish they would include the files with 0 sales, especially while sorting by upload date.
As far as the CV on IS goes, I always thought that "none of the above" should be a viable CV for any given keyword. So that way I could have Mount Helen - that isn't Mt St Helen.
I think you are right on with the "central park" analysis, Tom, and I am convinced that a lot of the "spamming" paranoia is just that .... misunderstanding and paranoia.
Also, I did learn a lot from just a short browse. Much of what I thought would be used for keywords on my images was WAY off. And, I need to shoot a lot more stuff on which I can use the keyword "Native American."
I'm starting to see what Sean (SJLocke) is talking about. Did I just share too much? LOL
-
Very interesting! I hadn't tried out the Darkroom feature before. Thanks for posting :)
More evidence to me that Istock's CV simply does not allow the buyers the flexibility of the full vocabulary. Several of the main keywords used to find my images would either not have been permitted or would have been side-tracked to a different meaning.
Same experience here. It's frustrating to see how many potential sales I could be missing at Istock because the top searched words on a given image may not be part of the CV.
to the OP, thankyou
Pleased to see other people have picked up on istock as well. One of my best sellers has 33% of sales based on a keyword istock considers not relevant.
Fascinating what is used, some are how you would think but for some images keywords that I would have thought fundamental barely being used, and for others keywords on early images that I dont add anymore because they are only marginally relevant pulling 5-10%
-
I think you are right on with the "central park" analysis, Tom, and I am convinced that a lot of the "spamming" paranoia is just that .... misunderstanding and paranoia.
Also, I did learn a lot from just a short browse. Much of what I thought would be used for keywords on my images was WAY off. And, I need to shoot a lot more stuff on which I can use the keyword "Native American."
I'm starting to see what Sean (SJLocke) is talking about. Did I just share too much? LOL
Yes, no telling secrets like MY best keywords. ;)
It was educational, I found one of my ELs with no words (visual search I guess?) Found two that I always wondered why people were buying them and the information explained it. 90% were using one word that was in with a whole bunch that it could have been.
But it's nice to know that people found my images without using a thesaurus and just used simple English words! I'm not wasting time looking for more words than exactly what's prominent in the image and happy to see it's that way. Of course I'll keep using the biological classifications and proper names, but I won't be using five different ways to label a red rose as four other subtle variations of red!
I think I have some ideas for new images now that I see what people were looking for. Cool!
-
I like this option very much!
-
I looked at it a bit more and there are a few things that are confusing to me, for instance these keywords were listed for a file with only one download...
central (33%); central park (33%); park (33%)
does that mean that the person was searching on "central, central park, park" or was the search for " "central park" central park " ?
(as an aside, if I was searching for an image of central park in NYC, I'd have considered this image a spam, since I think they keywords I actually had were "central oregon" and "smith rock state park" - I guess the buyer wanted it though.)
----------------------------------------------
I think the example above is showing how SS is dividing up the best match points given to an image that has recorded a sale.
Somebody searched for "central" and "park" and purchased your image. SS then awarded .33 of a point to "central park", .33 to "central" and .33 to "park" boosting your image moderately in all three searches. If only a single term is searched rather than two, 1 point is awarded to that term and it gets a more significant boost in best match.
I have a bunch of images that have sold once and now give .33 to a bunch of nonsense terms, things that almost nobody (other than the person who purchased my image) would search on. That fact that they are being awarded at all means that they must be the search that was used to find the image.
-
I like it. If i can see this: red (31%); background (22%); red background (8%) I presume that
- "red" was used when buyer searches in "background & textures" category only
- "red background" was used when buyer doesn't use categories when searching
- "background" was used when buyer doesn't know what he wants to download at all and just tries to use wholly his subscription
Interesting for me is that if I take a look at few similar images with almost or absolutely identical keywords, their "Downloads for Keyword" breakdown is completely different.
-
Also, I did learn a lot from just a short browse. Much of what I thought would be used for keywords on my images was WAY off. And, I need to shoot a lot more stuff on which I can use the keyword "Native American."
I'm starting to see what Sean (SJLocke) is talking about. Did I just share too much? LOL
Since you gave away this valuable info, I can share something back: my Scandinavian tower cakes have been bought several times using the keyword "kransekake". Time to go entering Norwegian keywords for all the other images too. ;)
Oh, and a while ago Dreamstime refused to let me use locality names in keywords for a series with children on top of a mountain. About 20% of the sales on SS are based on keywords such as "Norway", "Andalsnes" and "Rauma". I also see from the map that many of the sales of these pictures are local.
-
Definitely some surprises for me. :) I do love the direction SS is taking.
-
I looked at it a bit more and there are a few things that are confusing to me, for instance these keywords were listed for a file with only one download...
central (33%); central park (33%); park (33%)
does that mean that the person was searching on "central, central park, park" or was the search for " "central park" central park " ?
(as an aside, if I was searching for an image of central park in NYC, I'd have considered this image a spam, since I think they keywords I actually had were "central oregon" and "smith rock state park" - I guess the buyer wanted it though.)
----------------------------------------------
I think the example above is showing how Shutterstock is dividing up the best match points given to an image that has recorded a sale.
Somebody searched for "central" and "park" and purchased your image. Shutterstock then awarded .33 of a point to "central park", .33 to "central" and .33 to "park" boosting your image moderately in all three searches. If only a single term is searched rather than two, 1 point is awarded to that term and it gets a more significant boost in best match.
I have a bunch of images that have sold once and now give .33 to a bunch of nonsense terms, things that almost nobody (other than the person who purchased my image) would search on. That fact that they are being awarded at all means that they must be the search that was used to find the image.
I think you are misinterpreting the information.
I believe it is just information, SS is not awarding points to keywords. They are telling you what percentage of sales came from buyers using that particular keyword. I don't think this affects the best match in any way.
SS is just providing you with a tool to understand how buyers are searching.
steve
-
More on the "spam" paranoia:
One of my best sellers, a bull elk bugling, is found most often with keyword "Moose." My file does NOT contain the keyword 'Moose." I must assume that the buyer was searching for "Bull Moose" and received my image "Bull ELK." To the paranoid, my image would be an example of spam.
I call it "luck of the draw." Keyword. :o
Maybe I should post this in the "Something Positive" thread? ;D
-
I looked at it a bit more and there are a few things that are confusing to me, for instance these keywords were listed for a file with only one download...
central (33%); central park (33%); park (33%)
does that mean that the person was searching on "central, central park, park" or was the search for " "central park" central park " ?
(as an aside, if I was searching for an image of central park in NYC, I'd have considered this image a spam, since I think they keywords I actually had were "central oregon" and "smith rock state park" - I guess the buyer wanted it though.)
----------------------------------------------
I think the example above is showing how Shutterstock is dividing up the best match points given to an image that has recorded a sale.
Somebody searched for "central" and "park" and purchased your image. Shutterstock then awarded .33 of a point to "central park", .33 to "central" and .33 to "park" boosting your image moderately in all three searches. If only a single term is searched rather than two, 1 point is awarded to that term and it gets a more significant boost in best match.
I have a bunch of images that have sold once and now give .33 to a bunch of nonsense terms, things that almost nobody (other than the person who purchased my image) would search on. That fact that they are being awarded at all means that they must be the search that was used to find the image.
I think you are misinterpreting the information.
I believe it is just information, Shutterstock is not awarding points to keywords. They are telling you what percentage of sales came from buyers using that particular keyword. I don't think this affects the best match in any way.
Shutterstock is just providing you with a tool to understand how buyers are searching.
steve
How do you account for a image with one sale and results attributed to 3 different words?
-
SadStock
They used 3 words to search for the image.
-
More on the "spam" paranoia:
One of my best sellers, a bull elk bugling, is found most often with keyword "Moose." My file does NOT contain the keyword 'Moose." I must assume that the buyer was searching for "Bull Moose" and received my image "Bull ELK." To the paranoid, my image would be an example of spam.
I call it "luck of the draw." Keyword. :o
Maybe I should post this in the "Something Positive" thread? ;D
Oh thanks now you gave away my Moose secret, sold when people searched for Moose... :D And no I'm not making this up. I wondered what word was selling that image and now I know it was always just plain Moose.
I suppose Moose uploads will spike now and everyone else knows it's a popular search? LOL
-
More on the "spam" paranoia:
One of my best sellers, a bull elk bugling, is found most often with keyword "Moose." My file does NOT contain the keyword 'Moose." I must assume that the buyer was searching for "Bull Moose" and received my image "Bull ELK." To the paranoid, my image would be an example of spam.
I call it "luck of the draw." Keyword. :o
Maybe I should post this in the "Something Positive" thread? ;D
Oh thanks now you gave away my Moose secret, sold when people searched for Moose... :D And no I'm not making this up. I wondered what word was selling that image and now I know it was always just plain Moose.
I suppose Moose uploads will spike now and everyone else knows it's a popular search? LOL
Sorry. ;D
PS: I like that ... "Levity is the opposite of Gravity."
-
Don't want to sound stupid here, but how do you get to the darkroom feature. I have my bookmark set on the stats page and don't see a darkroom feature..... ::)
-
http://submit.shutterstock.com/darkroom/ (http://submit.shutterstock.com/darkroom/)
-
[url]http://submit.shutterstock.com/darkroom/[/url] ([url]http://submit.shutterstock.com/darkroom/[/url])
Thanks...looks interesting...now gotta go check it out more.
-
SadStock
They used 3 words to search for the image.
--------------------------------------------
Which three words?
"central", "park" and "central park"?
So the search that somebody actually entered was "central, park, 'central park'"? And they had to put "central park" in quotes. I'm thinking its far more likely that somebody searched on "central park".
-
More on the "spam" paranoia:
One of my best sellers, a bull elk bugling, is found most often with keyword "Moose." My file does NOT contain the keyword 'Moose." I must assume that the buyer was searching for "Bull Moose" and received my image "Bull ELK." To the paranoid, my image would be an example of spam.
I call it "luck of the draw." Keyword. :o
Maybe I should post this in the "Something Positive" thread? ;D
Oh thanks now you gave away my Moose secret, sold when people searched for Moose... :D And no I'm not making this up. I wondered what word was selling that image and now I know it was always just plain Moose.
I suppose Moose uploads will spike now and everyone else knows it's a popular search? LOL
Sorry. ;D
PS: I like that ... "Levity is the opposite of Gravity."
Pure Bullwinkle all the way. ;)
Sorry I have to take the day off tomorrow and search for more Moose photos, now that I know they are the secret to getting rich with microstock.
Actually I'm going on an Easter Egg hunt at a friends house. Adult party,like 40 year olds will be the "kids", jazz, and adult beverages, hopefully they remember to cook all the eggs this year. (no kidding!)
Hey, I'll take pictures, NCV of course.
(http://img14.imageshack.us/img14/7010/blueeggfoundweb.jpg)
Since one example would be an image with 8 downloads and the keywords are 40% - 40% & 20% I'd guess, this data is just the start and no one should jump to any big conclusions from limited data! It's firstname, lastname and event. Same as Robin Hood or Central Park being split and distributed, so are others. Some images have been downloaded more times and have no keywords for searches.
Give it some time. The data does not represent Percentage Of Sales, as some have suggested or assumed.
-
It's very interesting and in some ways worrying. On a couple of my oldest files the third highest ranked keyword is one that I would now regard as spam and would not use. For example, with pictures of flowers "apology" accounts for several percent of sales, so does "lover" though there is no lover in the picture.
Rather than fighting spam I can see this encouraging it.
-
SadStock
They used 3 words to search for the image.
--------------------------------------------
Which three words?
"central", "park" and "central park"?
So the search that somebody actually entered was "central, park, 'central park'"? And they had to put "central park" in quotes. I'm thinking its far more likely that somebody searched on "central park".
I think this is the correct explanation. I have similar results on Robin Hood sales - nobody, but nobody, would use the single word robin or hood.
-
the more curious is that "designers" look for colors, I place colors on keywords but never had that idea, this looks handy :)
-
I looked at it a bit more and there are a few things that are confusing to me, for instance these keywords were listed for a file with only one download...
central (33%); central park (33%); park (33%)
does that mean that the person was searching on "central, central park, park" or was the search for " "central park" central park " ?
(as an aside, if I was searching for an image of central park in NYC, I'd have considered this image a spam, since I think they keywords I actually had were "central oregon" and "smith rock state park" - I guess the buyer wanted it though.)
----------------------------------------------
I think the example above is showing how Shutterstock is dividing up the best match points given to an image that has recorded a sale.
Somebody searched for "central" and "park" and purchased your image. Shutterstock then awarded .33 of a point to "central park", .33 to "central" and .33 to "park" boosting your image moderately in all three searches. If only a single term is searched rather than two, 1 point is awarded to that term and it gets a more significant boost in best match.
I have a bunch of images that have sold once and now give .33 to a bunch of nonsense terms, things that almost nobody (other than the person who purchased my image) would search on. That fact that they are being awarded at all means that they must be the search that was used to find the image.
I think you are misinterpreting the information.
I believe it is just information, Shutterstock is not awarding points to keywords. They are telling you what percentage of sales came from buyers using that particular keyword. I don't think this affects the best match in any way.
Shutterstock is just providing you with a tool to understand how buyers are searching.
steve
No, he is not misinterpreting the information.
Only in the very beginning (the first few days), images are showing up for all keywords given. But after that, images will only show up in searches with keywords that were previously used to purchase the image.
For instance... I have a picture of a feather in the sky. It sold 7 times in the first 2 days. However, a few days later it still did not show up in the most popular search for 'feather', nor in any of the other keyword searches. With the exception of the keyword 'background'. The darkroom revealed that this was the keyword used for all purchases.
Then I have a background image. It was the most popular summer background image within 48 hours, but it did not show up when searching for 'spring background', 'nature background', etc. There again... the image was purchased 16 times on 5 different keywords, but the keywords 'spring' or 'nature' wasn't one of them. So, it just does not show up in the spring background search (I checked 25 pages with the filter on photographs)
Anyway, to make a long story short... I checked at least 15 files (incl. very good selling ones), but none of them were to be found in searches, other then the ones that had keywords the files were previously downloaded on.
Then yesterday my feather image showed up in the 'feather' search and when I checked the darkroom, there was indeed someone that purchased it, using the keywords 'feather'.
However... it's obvious that only new images have a chance to get downloaded on their key keywords, since they fall back more and more in the newest search and don't show up in the most popular search, unless and until the files were downloaded on these key keywords. The longer they are online, the slimmer chances are that they will get downloaded, meaning that we are missing out on sales big time.
I mean... I have recent images downloaded on the keywords 'green' several times, but they don't show up when searching for leaves, etc. So, these images have not sold again after a very good start, the first few days. And that does not surprice me, since there are millions of images with the keyword green and it basically says nothing about my image.
I do understand that the keywords used to purchase the files should be included in the algorithm, but I also think that it should not have this much weight. And also, that the files should at least show up on the keywords used in the title.