I just put as many keywords I find relevant. For some images it's only 10, for some 25 and for some 47.

I generally apply between 20-30 keywords, sometimes less.  Since I shoot primarily macro I don't see the need to apply more than that as my subject matter is pretty simple.  In looking at keyword relevance (based on buyer behavior) at iS and Image Gallery stats at SS it I get the impression that minutia isn't necessary when keywording....it's usually the obvious that grabs the buyer.

There is no 'perfect number of keywords'. Just stop keywording before you start adding words which you think 'might be applicable'. If you don't *know* those keywords apply to the image, you've added too many

I've tracked keywords bing used on the 100+ symbiostock sites (something no agency helps with).  I only track keywords used at least 5 times on a given site:

here are the 4000+ most used keywords:
http://cascoly.com/symbio/list.asp?list=401

the keywords used on the most sites:
http://cascoly.com/symbio/list.asp?list=401

alphabetical list of keywords used:
http://cascoly.com/symbio/list.asp?list=403

But there is places that can only use photos made in their area, such as the Ontario government agencies only allows photos in their brochures of people from Ontario and shot in Ontario.  A photo of a dandelion and someone blowing on it must be shot in Ontario and the model be from Ontario.  Government of Canada has some agencies that do the same, the models, photographer, and people in the photos must be from Canada and in Canada, any scenes (prints on a wall in offices) for decorating, etc all must be shot in Ontario or Canada respectively.  I'm unsure if other provincial governments also do this but it is practiced.  There are businesses that will for what ever reason want only local shots even if it's not obvious so putting in where the shot was taken could be relevant even when not obvious.

Quote from: gillian on September 28, 2013, 21:57
i agree with cuppacoffee, concepts like "beautiful" "cute" and "gorgeous" are subjective, and we've all seen cases of it and rolled our eyes. it must drive designers nuts. How about the images that contain "london" or "new york" but don't have any connection to that city, apart from being shot there? ugh! we could go on with examples. excessive spammy keywording hurts us all.

Quote from: cascoly on October 07, 2013, 23:34
I've tracked keywords bing used on the 100+ symbiostock sites (something no agency helps with).  I only track keywords used at least 5 times on a given site:

here are the 4000+ most used keywords:
http://cascoly.com/symbio/list.asp?list=401

the keywords used on the most sites:
http://cascoly.com/symbio/list.asp?list=401

alphabetical list of keywords used:
http://cascoly.com/symbio/list.asp?list=403

To confirm, these are keywords the photographers had applied to their images, NOT what customers have used to search with?

Quote from: Me on October 08, 2013, 05:38
To confirm, these are keywords the photographers had applied to their images, NOT what customers have used to search with?

Yep.
Like that list, but for searches I have here: http://picworkflow.com/blog/research/top-2000-image-buyers-searches-vs-keyword-popularity/

Probably relevant is the best not to annoy the client.

Whatever it takes to cover the image and everything that's in it. I'm still against adding words that no one will ever search for, just because they are more words.

If someone wants something red, they will search for red. If they want a tomato they will search for that, not Solanum lycopersicum crimson (cardinal, cinnabar or erythraean) Why do people waste so much time and effort trying to get trick words into keywords?

Let me put it another way... KEY words.

Whatever it takes to describe the image accurately. It's not a number it's a condition.

Quote from: Uncle Pete on October 08, 2013, 18:45
If someone wants something red, they will search for red. If they want a tomato they will search for that, not Solanum lycopersicum crimson (cardinal, cinnabar or erythraean) Why do people waste so much time and effort trying to get trick words into keywords?

The scientific name is not 'trick words'. It enables people whose first language is not English to find what they want. It makes sure that where they are looking for something specific, they get what they want, e.g. there are many species of 'Robin'.

Granted, I label all my birds with their specific recommended English vernacular, common English vernacular and Latin names, yet the top keyword on my gull pics is inevitably 'seagull', a word so beneath contempt that I didn't even use it as a keyword for a couple of years after I started.

Still, there might be someone who's grateful for the scientific name; and it is uber accurate, unlike the spam on most uploaded files these days at iS, but judging by results, at the others, though SS usually makes a good whack at filtering out spam.

Real simple as many as possible and sometimes that is well over 50 and close to 100.

What is found in the image?
Where are the items?
What is the condition or state of the items?
What actions or moods are taking place or implied?
What colors are present?
What textures are found?
What shapes are found?
What concept, theme or idea is found?"

"...and it is uber accurate..."

No, not uber. Some look up what they think is a certain species, bird, butterfly, etc. on Wikipedia and use that info whether or not it is actually correct just to add keywords without knowing the true identity of the beast. I've seen it. There are many variations in species and not all uploaders take the time to find the correct scientific name. I've seen grasshoppers keyworded as grasshoppers where they should/could be katydids or locusts. If I were a true scientist I would not trust the scientific names on many microstock images. That being said, if they really knew what they were looking for I bet that they would first search for grasshopper and then maybe Orthoptera Caelifera. Grasshopper and the scientific name would be in the same search.

Perhaps this is a more accurate place for those kinds of photos? http://www.sciencephoto.com/aboutSPL.html

Quote from: cuppacoffee on October 10, 2013, 13:01
"...and it is uber accurate..."

No, not uber. Some look up what they think is a certain species, bird, butterfly, etc. on Wikipedia and use that info whether or not it is actually correct just to add keywords without knowing the true identity of the beast. I've seen it. There are many variations in species and not all uploaders take the time to find the correct scientific name. I've seen grasshoppers keyworded as grasshoppers where they should/could be katydids or locusts. If I were a true scientist I would not trust the scientific names on many microstock images. That being said, if they really knew what they were looking for I bet that they would first search for grasshopper and then maybe Orthoptera Caelifera. Grasshopper and the scientific name would be in the same search.

Perhaps this is a more accurate place for those kinds of photos? http://www.sciencephoto.com/aboutSPL.html I don't know anything about them.

OK, accepted, if the person misidentifies the species, adding the latin of the species doesn't help.
I just wikied (to what purpose) a few days ago a bird which was a genuine misidentification (i.e. no other bird was spammed). The tog had gone to the bother of copying a description of the wrong species from somewhere, by which I was left wondering - 'how on earth did they arrive at that identification from that description?' as so many key features from the description were not visible on the bird. Although it was a species not known to me, it didn't take me long to have a positive ID on it. But I suppose it helps if you know where to start.

It's usually not difficult to get an ID. Only last week I had a photo of a Hoverfly I needed IDd. I put it onto Flickr, got a well-meaning suggestion which turned out to be a blind alley, then within an hour of upload, the correct identification from a real top expert in the field (I Googled him!).

I also agree that most people needing accuracy will go to specialist agencies. A clear example of how the clueless and the evil (spammers) have spoiled it for the rest of us.

Just from your persona here I'm sure that all of your images are keyworded extremely well. Take a look at the sciencephoto.com site and see if it is worth your while to become a contributor there. I think that would be a perfect fit.

http://www.sciencephoto.com/static/media/contact/SPL_DigitalGuidelines.pdf

Proper keywording takes a lot of time especially when you trying to find right description of some animal or plant. I usually start with Shutterstock keywords suggestions, then Google image search and wikipedia. Sometimes it is almost impossible like once when I got photo of some dragonfly. Likely it was very specific to the location when I took a shot.

Quote from: cuppacoffee on October 10, 2013, 14:59
Just from your persona here I'm sure that all of your images are keyworded extremely well. Take a look at the sciencephoto.com site and see if it is worth your while to become a contributor there. I think that would be a perfect fit.

http://www.sciencephoto.com/static/media/contact/SPL_DigitalGuidelines.pdf

Thanks for the link.

The "Buyer searched for" column in earnings on Dreamstime is revealing

A FEW EXAMPLES OF RECENT SALES ARE DIRECTLY FROM BASIC KEYWORDS

Title: Burlap texture
Searched for: texture

Title: Bowlerl hat
Searched for: hat

Title: Maple leaf
Searched for: fall leaf

Title: Empty head
Searched for: memory

Title: Sunglasses
Searched for: sunglasses

BUT THERE ARE A FEW ODD ONES

Title: Rowing team
Searched for: care (care was not a keyword)

Title: Beach umbrella
Searched for: house for sale (no keyword here)

Title: Clef and star
Searched for: broadway sign (broadway was a keyword)

I've begun to think you don't need 50 keywords - 30 should be plenty.

From the DT FAQs

Q A BUYER FOUND MY IMAGE USING KEYWORDS THAT ARE NOT PRESENT IN MY IMAGE INFO. WHY IS THAT?

A Buyers may navigate several pages before buying an image so you may sometimes see that the "buyer searched for" section lists keywords which are not included in your info, may be unusual or completely unrelated to your image. You could use these keywords to identify the concept as in some cases unrelated keywords can reveal more about the buyers' project. Although they start from a search based on keywords, the buyers may then navigate the "similar images" or the "more images with this model" sections. A search for "smiling woman" can continue in the "images with the same model" section from where the buyer can move to a different expression and end up downloading an "upset man". If the keywords are never changed while navigating these pages, then the initial "smiling woman" set is saved and shown as used by the buyer. The "buyer searched for" feature is designed to help you better understand how searches are made and which of your keywords are most relevant sale-wise speaking. You can use it as guideline and edit the image info to include relevant or obvious keywords you may have missed. We do ask you to keep in mind the visual search explanation above and NOT include unrelated or irrelevant terms. Spam is not allowed so make sure all keywords apply to your image. If a buyer downloads your "dog looking up" but initially searched for "business man looking up," the only keywords you can add to your image are "looking" and "up" in case they are missing.

Quote from: ShadySue on October 10, 2013, 11:17

The scientific name is not 'trick words'. It enables people whose first language is not English to find what they want. It makes sure that where they are looking for something specific, they get what they want, e.g. there are many species of 'Robin'.

Granted, I label all my birds with their specific recommended English vernacular, common English vernacular and Latin names, yet the top keyword on my gull pics is inevitably 'seagull', a word so beneath contempt that I didn't even use it as a keyword for a couple of years after I started.

Still, there might be someone who's grateful for the scientific name; and it is uber accurate, unlike the spam on most uploaded files these days at iS, but judging by results, at the others, though SS usually makes a good whack at filtering out spam.

just had several images rejected by SS because captions 'must be in english' -- only non-English words were the scientific names!

I think sci names are one of the exceptions to not being too specific

Quote from: Me on October 08, 2013, 05:38

Quote from: cascoly on October 07, 2013, 23:34
I've tracked keywords bing used on the 100+ symbiostock sites (something no agency helps with).  I only track keywords used at least 5 times on a given site:

here are the 4000+ most used keywords:
http://cascoly.com/symbio/list.asp?list=401

the keywords used on the most sites:
http://cascoly.com/symbio/list.asp?list=401

alphabetical list of keywords used:
http://cascoly.com/symbio/list.asp?list=403

To confirm, these are keywords the photographers had applied to their images, NOT what customers have used to search with?

correct - these keywords come from the image files symbio creates.  soon(?) there will be a file that records searches and we'll be able to do comparisons

steve

Quote from: cascoly on October 10, 2013, 22:34
just had several images rejected by SS because captions 'must be in english' -- only non-English words were the scientific names!

Is that general on SS or an eejit inspector?

Hmmm, not sure what the 'caption is there.
I just found one:
Stock Photo: Lesser redpoll, Carduelis cabaret, single bird on branch, Gloucestershire

Is that the caption or the title?

Image ID:

Then again, on iStock (stupidly), although many scientific names map to that species, in far too many others, the scientific maps to a generic. So for example "Vespula vulgaris" (common wasp, the clue's in the name) maps only to 'wasp', so a person knowing that they want to look for Vespula vulgaris has to wade through all wasps. I forgot about this; I've seen lots of examples of the same thing. The worst about this is because it's in the CV, you can't do "Vespula vulgaris" and get only that, as you could if it was a non-CV term.

.

Quote from: ShadySue on October 10, 2013, 22:44

Quote from: cascoly on October 10, 2013, 22:34
just had several images rejected by SS because captions 'must be in english' -- only non-English words were the scientific names!

Is that general on SS or an eejit inspector?

Hmmm, not sure what the 'caption is there.
I just found one:
Stock Photo: Lesser redpoll, Carduelis cabaret, single bird on branch, Gloucestershire

Is that the caption or the title?

Image ID:

I usually get the latin names approved
SS pulls their caption from the iptc caption metadata

this was the same reviewer who rejected thunderstorm and backlit mountains with rainbow for 'lighting' problems

Quote from: tickstock on October 10, 2013, 23:56

Quote from: ShadySue on October 10, 2013, 23:16
Then again, on iStock (stupidly), although many scientific names map to that species, in far too many others, the scientific maps to a generic. So for example "Vespula vulgaris" (common wasp, the clue's in the name) maps only to 'wasp', so a person knowing that they want to look for Vespula vulgaris has to wade through all wasps. I forgot about this; I've seen lots of examples of the same thing. The worst about this is because it's in the CV, you can't do "Vespula vulgaris" and get only that, as you could if it was a non-CV term.

Contributors aren't experts on these things and neither are the inspectors.  The issue (of imperfect mapping) can be fixed by sitemailing the keywording admin or posting in the keywording forum

Yeah, but I've been banned for almost three years, so that can't happen.

Quote
but nothing will fix bad keywording. 

BTW compare the results you get at Shutterstock to Istock, which would you rather see if you were a buyer?  It at least shows that not very many people will put in the scientific name on their own.  On Shutterstock Vespula Vulgaris 69 results, common wasp 103 results, wasp 8,000+ results.  The best search for that subject looks like the istock one and the shutterstock one for 'wasp'.

Searching SS for Vespula vulgaris, I get 69 results and most of them are actually common wasps. On iS I get 4000 random wasps and hornets that I have to really look at and various other insects . (the 'various other insects' are spamming, of course.)
So if I were a buyer and Vv was what I wanted, SS would be preferable.