MicrostockGroup Sponsors


Author Topic: what is keyword spam?  (Read 7330 times)

0 Members and 1 Guest are viewing this topic.

cascoly

  • Photography, travel & online games at cascoly.com

« on: July 30, 2013, 15:38 »
0
as I've looked at the various keyword stats at http://symbiostock-search.com, I've discovered several instances of what i'd consider keyword spamming

there are over 11000 different keywords being used right now.  folks will have different opinions as to whether  images of mule, elk or monkey should also include the tags horse, deer or ape, respectively.   but to me, when someone has an image of paris, and also includes new York and Singapore as keywords, that's a bit much (I've made this up completely but here are analogous tags out there)

my question is how should sym sites and the global searches handle this?  simplest is to stay with a completely hands off policy, but this risks user annoyance if poor keywording is too common

any ideas?


Ron

« Reply #1 on: July 30, 2013, 15:52 »
0
I try to stay away from spam, but if I have a flag of a french colony and its called whatever, and it has traces of the French flag, I will tag it with the Island name and with France and French Colony. Thats not spam to me. It might be to others. I have never spammed, but I have stretched on occasions. And if you find spam, it could be a stray keyword. If I run into them, I delete them myself.

What can you do about it, nothing I think. If I will be 'punished' for it in the future, I will be more than happy to disconnect from the network.

Just my two cents.

PS: My keywords are a mess because I split up all my keyword phrases as workaround for the local search, but it created a heap of spam. I am eagerly waiting for the premium version so I can clean it up.

« Reply #2 on: July 30, 2013, 16:19 »
+1
Hi

I realised i messed up by using dumbell for kettle weights, as far as i know i changed them. However, if anyone notices any bad keywords on mine, please tell me!

Thanks
Mark

stocked

« Reply #3 on: July 30, 2013, 16:47 »
0
I'm rekeywording most of my files before I used the keyword tools and the 50 keyword maximum that ends always in SPAM

« Reply #4 on: July 30, 2013, 16:54 »
0
Same here. I don't keyword spam, but in the early days of stock, we were allowed to do some weird things. I don't think it's an issue for me because I am going through each and every photo and cleaning up. Just like I had irrelevant ones in there, I have also found images where I missed a bunch I should have used. Please tell me if anyone finds wrong ones and I will fix them.

When people keyword spam on their own sites, it's their business. If it annoys buyers, they just won't go back. But now that we are all connected through the global search, keyword spamming that will annoy buyers punishes the whole network. And I'm not talking about mis-keywords here and there, as Ron describes. I am talking about what cascoly talked about: images of mule, elk or monkey that also include the tags horse, deer or ape, respectively. A horse is not the same as a mule, and if I were looking for a horse, I want a picture of a horse. If I wanted a mule, I would search horse AND mule, or mule.

Do a search for corn on some of the agency sites...I see all kinds of images that have nothing to do with corn, at least here in the US...images of wheat, etc.

Since you, cascoly, put all of your time and effort into the global search, I think that if you see a site that does this keyword spamming excessively, maybe you could contact them and ask them to clean it up. If, after a reasonable period of time, it has not cleaned up, then maybe they should not be included in the search? Not sure if technically you can do that, but I think you have every right since you are the one doing the work on the search.

If people don't like it, they can be independent. I would like to think that most people here would be amenable to fixing the keywords, because they are likely leftovers from early microstock and are an oversight. The ones that put up a big fuss, in my opinion, are the ones who probably have no intention of fixing them and have done it intentionally. Everyone knows there are always people who want to cheat the system.  ::)

ShazamImages

  • ShazamImages.com
« Reply #5 on: July 30, 2013, 17:38 »
+1
Without the ability of having a search engine that can search on complex phrases, it is almost impossible to avoid what some people would call spam.  For example, a lot of artists have images with copy space on them (so that customers can add text to the image).  But in order to include that in the keywords, they currently have to include both words "copy" and "space".  So anyone searching for "space" (as in the stuff between the stars and planets), will have images that show up with white copy space and not one star or planet in them.  You could also include the keyword "copyspace" but you never know how someone will spell the word(s).  The same is true for many other keywords that are part of complex phrases.  For example, if someone wants to find images using the word "pop" (as in burst), they will find images with "popcorn", "soda pop", "pop artists", etc.  So I think that the search engine needs to be upgraded to include the ability to search for complex phrases.

In addition, everyone has a different definition of what spam is, so it is hard to come up with hard and fast rules.  This is especially true with conceptual images.

Another idea might be to have the ability to "flag" a keyword on an image.  This way someone could flag a keyword and the artist would get an email.  If the artist starts getting a bunch of emails about a certain keyword in an image, then that would let them know that the keyword should probably be removed from the image.

cascoly

  • Photography, travel & online games at cascoly.com

« Reply #6 on: July 30, 2013, 18:33 »
+1


..... I am talking about what cascoly talked about: images of mule, elk or monkey that also include the tags horse, deer or ape, respectively. A horse is not the same as a mule, and if I were looking for a horse, I want a picture of a horse. If I wanted a mule, I would search horse AND mule, or mule.....


here's the problem - I don't consider those examples spam - someone looking for an ape, may be looking for the more general idea of 'monkey' etc.   since we can't be there to see that buyers use perfect keywords, I don't have a big problem when people use terms more loosely than if it were a peer reviewed journal article.


Since you, cascoly, put all of your time and effort into the global search, I think that if you see a site that does this keyword spamming excessively, maybe you could contact them and ask them to clean it up. If, after a reasonable period of time, it has not cleaned up, then maybe they should not be included in the search? Not sure if technically you can do that, but I think you have every right since you are the one doing the work on the search.



that's why I started this thread - because I don't WANT to be the one who decides these things

cascoly

  • Photography, travel & online games at cascoly.com

« Reply #7 on: July 30, 2013, 18:36 »
0
Without the ability of having a search engine that can search on complex phrases, it is almost impossible to avoid what some people would call spam.  For example, a lot of artists have images with copy space on them (so that customers can add text to the image).  But in order to include that in the keywords, they currently have to include both words "copy" and "space".  So anyone searching for "space" (as in the stuff between the stars and planets), will have images that show up with white copy space and not one star or planet in them.  You could also include the keyword "copyspace" but you never know how someone will spell the word(s).  The same is true for many other keywords that are part of complex phrases.  For example, if someone wants to find images using the word "pop" (as in burst), they will find images with "popcorn", "soda pop", "pop artists", etc.  So I think that the search engine needs to be upgraded to include the ability to search for complex phrases.

In addition, everyone has a different definition of what spam is, so it is hard to come up with hard and fast rules.  This is especially true with conceptual images.

Another idea might be to have the ability to "flag" a keyword on an image.  This way someone could flag a keyword and the artist would get an email.  If the artist starts getting a bunch of emails about a certain keyword in an image, then that would let them know that the keyword should probably be removed from the image.

yes, I think those are good examples

the email info in the symbiocard files is encrypted, so emailing isn't a possibility,but maybe the ability to flag potential spam might be doable

-----
the bulk editor is going to be a great way to clean up tags and descriptions.  it makes it easy to see and edit images that have incorrect tags, etc

« Reply #8 on: July 30, 2013, 18:37 »
0
then disregard everything I said.  ;)

« Reply #9 on: July 30, 2013, 22:07 »
0
Since it was used as an example of possible spam, here's a brief natural history lesson. 'Deer' is a family of animals which includes the following North American species:
1. White-tailed deer aka Virginia deer
2. Key deer (a dwarf variant of the white-tailed deer found in southern Florida)
3. Coues deer (a dwarf variant of the white-tailed deer found in the Arizona desert)
4. Mule deer
5. Black-tailed deer (a smaller variant of the mule deer found in coastal areas of southwestern BC and Washington)
6. Elk aka Wapiti
7. Moose
8. Woodland caribou
9. Barren-ground caribou
10. Reindeer
... there may be more - these were just the first ones that popped up in my abbreviated pocket field guide.

I would keyword any of the above species with their correct name, their family name ('deer') and their Latin names. I take the same approach with wildflowers and trees as well. I don't consider that spam.

Ron

« Reply #10 on: July 31, 2013, 02:36 »
0
Without the ability of having a search engine that can search on complex phrases, it is almost impossible to avoid what some people would call spam.  For example, a lot of artists have images with copy space on them (so that customers can add text to the image).  But in order to include that in the keywords, they currently have to include both words "copy" and "space".  So anyone searching for "space" (as in the stuff between the stars and planets), will have images that show up with white copy space and not one star or planet in them.  You could also include the keyword "copyspace" but you never know how someone will spell the word(s).  The same is true for many other keywords that are part of complex phrases.  For example, if someone wants to find images using the word "pop" (as in burst), they will find images with "popcorn", "soda pop", "pop artists", etc.  So I think that the search engine needs to be upgraded to include the ability to search for complex phrases.

In addition, everyone has a different definition of what spam is, so it is hard to come up with hard and fast rules.  This is especially true with conceptual images.

Another idea might be to have the ability to "flag" a keyword on an image.  This way someone could flag a keyword and the artist would get an email.  If the artist starts getting a bunch of emails about a certain keyword in an image, then that would let them know that the keyword should probably be removed from the image.
That! Man I couldnt have said it any better.

shudderstok

« Reply #11 on: July 31, 2013, 04:41 »
+2


..... I am talking about what cascoly talked about: images of mule, elk or monkey that also include the tags horse, deer or ape, respectively. A horse is not the same as a mule, and if I were looking for a horse, I want a picture of a horse. If I wanted a mule, I would search horse AND mule, or mule.....


here's the problem - I don't consider those examples spam - someone looking for an ape, may be looking for the more general idea of 'monkey' etc.   since we can't be there to see that buyers use perfect keywords, I don't have a big problem when people use terms more loosely than if it were a peer reviewed journal article.


Since you, cascoly, put all of your time and effort into the global search, I think that if you see a site that does this keyword spamming excessively, maybe you could contact them and ask them to clean it up. If, after a reasonable period of time, it has not cleaned up, then maybe they should not be included in the search? Not sure if technically you can do that, but I think you have every right since you are the one doing the work on the search.



that's why I started this thread - because I don't WANT to be the one who decides these things

The general rule of keywording is this... no more, no less
Who:
What:
Why:
Where:
When:

So for example a monkey is a monkey not an ape, and an ape is an ape not a monkey. so keyword "what" it is and nothing more it can't be both. now an ape and a monkey are both primates so using "primate" is acceptable. monkeys have tails and apes don't.

Using this as an example, it can be either an ape or a monkey - but not both, but both can be a primate.

i hate nothing more than bad keywording as it affects us all in the long run and I report it when I spot it, we should all do the same.

i recently saw a photo of a tea plantation in Malaysia and the keywords were also Sri Lanka, India, Thailand, and China... how can a known tea plantation in Malaysia be located in five different countries? I have also seen a temple in Kathmandu with keywords that contained as keywords three other temples in Kathmandu and also used Tibet, which is not Nepal at all.

Personally, if i was an art director/designer and have never left small town anywhere, but was assigned to make a brochure of sorts or a magazine spread featuring agriculture in Sri Lanka, and I ended up using a photograph on the cover that was Malaysia, I'd be pissed off.

That is what keyword spam is - falsifying information.






« Last Edit: July 31, 2013, 04:44 by shudderstok »

shudderstok

« Reply #12 on: July 31, 2013, 04:49 »
+2
Since it was used as an example of possible spam, here's a brief natural history lesson. 'Deer' is a family of animals which includes the following North American species:
1. White-tailed deer aka Virginia deer
2. Key deer (a dwarf variant of the white-tailed deer found in southern Florida)
3. Coues deer (a dwarf variant of the white-tailed deer found in the Arizona desert)
4. Mule deer
5. Black-tailed deer (a smaller variant of the mule deer found in coastal areas of southwestern BC and Washington)
6. Elk aka Wapiti
7. Moose
8. Woodland caribou
9. Barren-ground caribou
10. Reindeer
... there may be more - these were just the first ones that popped up in my abbreviated pocket field guide.

I would keyword any of the above species with their correct name, their family name ('deer') and their Latin names. I take the same approach with wildflowers and trees as well. I don't consider that spam.

moose, deer family, latin name... but not deer.

cascoly

  • Photography, travel & online games at cascoly.com

« Reply #13 on: July 31, 2013, 17:13 »
-1


The general rule of keywording is this... no more, no less
Who:
What:
Why:
Where:
When:

So for example a monkey is a monkey not an ape, and an ape is an ape not a monkey. so keyword "what" it is and nothing more it can't be both. now an ape and a monkey are both primates so using "primate" is acceptable. monkeys have tails and apes don't.

Using this as an example, it can be either an ape or a monkey - but not both, but both can be a primate.

i hate nothing more than bad keywording as it affects us all in the long run and I report it when I spot it, we should all do the same.

i recently saw a photo of a tea plantation in Malaysia and the keywords were also Sri Lanka, India, Thailand, and China... how can a known tea plantation in Malaysia be located in five different countries? I have also seen a temple in Kathmandu with keywords that contained as keywords three other temples in Kathmandu and also used Tibet, which is not Nepal at all.

Personally, if i was an art director/designer and have never left small town anywhere, but was assigned to make a brochure of sorts or a magazine spread featuring agriculture in Sri Lanka, and I ended up using a photograph on the cover that was Malaysia, I'd be pissed off.

That is what keyword spam is - falsifying information.

I agree on the bad place names, that was the main reason I started the thread  -- altho the caption should clear that up

-- there are, however, 2 ways to keyword -- first, to accurately & precisely describe the image  - the Ws

but second,  a broader array of tags to help people find the image even when they don't know exactly what they want, or when they're willing to accept something beyond the narrow keywords they entered  -- as the various examples make clear, that second type is interpreted differently.   the more professional your buyers, the narrower you can make the keywords

Ron

« Reply #14 on: July 31, 2013, 17:16 »
0
Very easy on simple images, not on conceptual images where a lot of keywords can be used, to describe the concept or a concept.

I have images of a man in a reptile skin, the keywords used on that image are mostly conceptual and far fetched. I added liar to the keywords, and it was bought using the keyword liar.

shudderstok

« Reply #15 on: July 31, 2013, 17:18 »
-2
Very easy on simple images, not on conceptual images where a lot of keywords can be used, to describe the concept or a concept.

I have images of a man in a reptile skin, the keywords used on that image are mostly conceptual and far fetched. I added liar to the keywords, and it was bought using the keyword liar.

and how do you know it was purchased by using a certain keyword? this is a concept i would call exaggeration.

cascoly

  • Photography, travel & online games at cascoly.com

« Reply #16 on: July 31, 2013, 17:19 »
+1
.....

i hate nothing more than bad keywording as it affects us all in the long run and I report it when I spot it, we should all do the same.

.....


our problem/opportunity with sym is that there is no place to report right now - other than going to the individual sites & leaving comments or emails

one possibility would be to create a list of flagged images & keywords that was only accessible to symbio owners.  each owner could check reports about their images & remove them from the list

other ideas?

shudderstok

« Reply #17 on: July 31, 2013, 17:21 »
-2


The general rule of keywording is this... no more, no less
Who:
What:
Why:
Where:
When:

So for example a monkey is a monkey not an ape, and an ape is an ape not a monkey. so keyword "what" it is and nothing more it can't be both. now an ape and a monkey are both primates so using "primate" is acceptable. monkeys have tails and apes don't.

Using this as an example, it can be either an ape or a monkey - but not both, but both can be a primate.

i hate nothing more than bad keywording as it affects us all in the long run and I report it when I spot it, we should all do the same.

i recently saw a photo of a tea plantation in Malaysia and the keywords were also Sri Lanka, India, Thailand, and China... how can a known tea plantation in Malaysia be located in five different countries? I have also seen a temple in Kathmandu with keywords that contained as keywords three other temples in Kathmandu and also used Tibet, which is not Nepal at all.

Personally, if i was an art director/designer and have never left small town anywhere, but was assigned to make a brochure of sorts or a magazine spread featuring agriculture in Sri Lanka, and I ended up using a photograph on the cover that was Malaysia, I'd be pissed off.

That is what keyword spam is - falsifying information.

I agree on the bad place names, that was the main reason I started the thread  -- altho the caption should clear that up

-- there are, however, 2 ways to keyword -- first, to accurately & precisely describe the image  - the Ws

but second,  a broader array of tags to help people find the image even when they don't know exactly what they want, or when they're willing to accept something beyond the narrow keywords they entered  -- as the various examples make clear, that second type is interpreted differently.   the more professional your buyers, the narrower you can make the keywords

accuracy does not have a broader array.

you are either pregnant or you are not.

cascoly

  • Photography, travel & online games at cascoly.com

« Reply #18 on: July 31, 2013, 17:21 »
+2
....
and how do you know it was purchased by using a certain keyword? this is a concept i would call exaggeration.

some agencies like DT show the keywords that were used

eventually we'll have similar data about searches in sym reports and an owner could match the searches to sales

Ron

« Reply #19 on: July 31, 2013, 17:23 »
-1
Very easy on simple images, not on conceptual images where a lot of keywords can be used, to describe the concept or a concept.

I have images of a man in a reptile skin, the keywords used on that image are mostly conceptual and far fetched. I added liar to the keywords, and it was bought using the keyword liar.

and how do you know it was purchased by using a certain keyword? this is a concept i would call exaggeration.
Shutterstock shows you the keywords used to purchase the images. Probably technology unfamiliar to dinosaurs.

ShazamImages

  • ShazamImages.com
« Reply #20 on: July 31, 2013, 17:24 »
+1
Very easy on simple images, not on conceptual images where a lot of keywords can be used, to describe the concept or a concept.

I have images of a man in a reptile skin, the keywords used on that image are mostly conceptual and far fetched. I added liar to the keywords, and it was bought using the keyword liar.

and how do you know it was purchased by using a certain keyword? this is a concept i would call exaggeration.

Dreamstime sometimes provides the keywords for a purchased image.

cascoly

  • Photography, travel & online games at cascoly.com

« Reply #21 on: July 31, 2013, 17:26 »
0
.......
accuracy does not have a broader array.

you are either pregnant or you are not.

that's precisely the point - the reason for keywords is to help people find the image they will buy, not to provide the most accurate description.  if you choose to ignore buyers who don't know the difference between a monkey & an ape, that's your decision

what about an image of a woman on an exam table talking to her doctor - pregnant might be a possible keyword?

shudderstok

« Reply #22 on: July 31, 2013, 17:28 »
+1
 interesting, did not know that about some agencies showing you which keywords were used. i stand corrected. i am not with any of those agencies so had no clue.
« Last Edit: July 31, 2013, 17:43 by shudderstok »

shudderstok

« Reply #23 on: July 31, 2013, 17:39 »
+1
.......
accuracy does not have a broader array.

you are either pregnant or you are not.

that's precisely the point - the reason for keywords is to help people find the image they will buy, not to provide the most accurate description.  if you choose to ignore buyers who don't know the difference between a monkey & an ape, that's your decision

what about an image of a woman on an exam table talking to her doctor - pregnant might be a possible keyword?

you can caption however you want. i have been in this business for way too many years working with many magazines worldwide, newspapers world wide, and publishers worldwide to know that you don't caption a monkey an ape or vice versa in hopes your buyer/client/editor/art director/designer does not know the difference and do so in a desperate attempt to sell an image. that shows your lack of professionalism.

now in your case of the woman on exam table, it would be a stretch to use pregnant - unless she was visibly pregnant, but you could also use cancer and syphilis.
so what are you going to do keyword it for what it is or use the entire medical dictionary of every known ailment to mankind as a keyword in desperate hopes of a sale?




« Last Edit: July 31, 2013, 17:44 by shudderstok »

cascoly

  • Photography, travel & online games at cascoly.com

« Reply #24 on: July 31, 2013, 21:11 »
0

you can caption however you want. i have been in this business for way too many years working with many magazines worldwide, newspapers world wide, and publishers worldwide to know that you don't caption a monkey an ape or vice versa in hopes your buyer/client/editor/art director/designer does not know the difference and do so in a desperate attempt to sell an image. that shows your lack of professionalism.

now in your case of the woman on exam table, it would be a stretch to use pregnant - unless she was visibly pregnant, but you could also use cancer and syphilis.
so what are you going to do keyword it for what it is or use the entire medical dictionary of every known ailment to mankind as a keyword in desperate hopes of a sale?
you're just proving my point - it's NOT ABOUT professionalism, and it's not about the caption. keywords are VWERY different from  captions.   you're in the first class of buyers I described.  many of us are aiming for the  2nd group-- the much larger population of people who are NOT professionals but who sometimes need images.  the pros are likely to stay with their subscriptions where they can get XXL images for pennies


 

Related Topics

  Subject / Started by Replies Last post
45 Replies
8905 Views
Last post October 16, 2008, 07:02
by basti
14 Replies
3263 Views
Last post February 20, 2009, 07:41
by Peter
31 Replies
6353 Views
Last post July 29, 2009, 16:08
by puravida
11 Replies
3183 Views
Last post January 30, 2011, 11:18
by jbarber873
11 Replies
1408 Views
Last post July 30, 2013, 05:28
by plrang

Sponsors

Microstock Poll Results

Sponsors