The Methodology Behind the Analysis of Image Keywording for Stock Houses

Stock House Thursdays: The Methodology

Welcome:

This is all the work that went on behind the scenes for me to give an accurate, in-depth analyses of the stock houses I have reviewed in my series of posts call Stock House Thursdays.  For readers who haven’t yet read any of these posts, you would do well to first read my introductory post to the series: 10 Ways that Stock Houses can be different in their Keywording Standards.

For those of you who have arrived here having already read one of my stock house analyses, congratulations for being curious enough about what I’m doing to read the gruesome details of how I’m doing it.  This will greatly benefit you.  By knowing how I did my analysis, you can use my methods at any stock house, or even your own personal website.  Understanding why I did each of my searches, and why I chose particular keywords to test will inevitably help you to truly understand the concepts behind keywords that will give you an advantage over most other photographers.

Introduction:

My basic plan was to start with the bigger houses as they have a lot more contributors, and after all, I would like this information to be useful to somebody.  With this information, I hope to inform contributing photographers how to best keyword their images to maximize their exposure.

The short of it is that having the most appropriate keywords for each stock house will get your images found in more searches by buyers, and will lead to more sales. If a buyer at your stock house searches for dogs, will your image with the keyword dog be returned in the results? If not, then you need both dog and dogs in your keywords if you want your image to be seen in both search results.

How did I figure out if you need both dog and dogs?  Simple, by doing what a buyer would do–search the site for each dog and dogs and see how many images were returned.  If a different number of images is returned for dog than for dogs, then their search engine cannot tell that these searchers are looking for the same thing.  This tells me that you need to keyword both dog and dogs.

All of my results and prognoses are based upon searches I did at each stock house’s website. The proof is in the searching, not in what you read in forums, or even on the stock house’s website.

So, please keep in mind that these analyses aren’t classic reviews with thumbs up or 4 out of 5 stars at the end.  This is not intended to be a recommendation or trashing of any of these stock houses.  Honestly, most of the elements I’m looking at don’t necessarily have an objective right or wrong answer.

The only thing that really matters is that  photographers use this information to be learn the correct way to keyword for each individual stock house they contribute to.

The Variables:

  1. Plurals
  2. Synonyms
  3. Variant Spellings
  4. Verb Forms
  5. Other Forms of the Word
  6. Keyword Phrases
  7. Words in a Keyword Phrase
  8. Prepositions
  9. Periods
  10. Apostrophes
  11. Hyphens
  12. Capitalization
  13. Abbreviations
  14. Individual Letters
  15. Numerals
  16. Keyword Separators
  17. Alphabetization of Keywords
  18. Keyword Weight
  19. Number of Keywords Allowed
  20. Disambiguation of Homographs
  21. Faceted Searches

Plurals:  I started with a variable that should pretty clear to everybody.  Do you need both the singular and plural form(s) of your keywords?

  • I did three comparative searches on words with regular plurals, two common and one relatively uncommon:
    • dog & dogs
    • boy & boys
    • tureen & tureens
  • Then I did ten comparative searches of words with irregular plurals:
    • child & children
    • man & men
    • policeman & policemen
    • knife & knives
    • ibex & ibexes
    • couch & couches
    • baby & babies
    • goose, geese & gooses
    • mouse, mice & mouses
    • fungus & fungi

I have found stock house’s abilities that run the spectrum of returning the same results for the singular and plural forms of every word I could throw at it, to one that seemed to only recognize plurals of common people types (babies, men, women, etc).  I’ve found some that work for only regular plurals.  I’ve even seen some that can handle all regular and irregular plurals like children, men, couches, babies, and ibexes, but not geese, mice or fungi.

Synonyms:  Do you need to include all synonyms of a keyword to make sure your image is found?

  • I did six comparative searches on common words with synonyms, six common and one relatively uncommon:
    • sofa & couch
    • child & kid
    • frustration & aggravation
    • boy & male child
    • happy & joy
    • mad & angry
    • daffodil & narcissus
  • I also did three comparative searches of common American English words with their British English synonyms:
    • boy & lad
    • elevator & lift
    • apartment & flat

To be blunt, most stock houses simply don’t deal with synonyms, but I check each one just in case.

Variant Spellings:  Many words have variant spellings, and they are usually overlooked by photographers when keywording.

  • I did two comparative searches on common words widely accepted variant spellings:
    • gray & grey
    • catsup & ketchup
  • Along with those, I did two more comparative searches of somewhat less than common English words of foreign origin along with their variant spellings:
    • tabouli, tabbouleh & tabouleh
    • Hanukah, Hanukkah & Chanukah
  • Then I did two comparative searches on common words with American English and British English variations:
    • airplane & aeroplane
    • industrialization & industrialisation
  • Lastly, I did three comparative searches on common words with one and two word variations:
    • closeup & close up
    • makeup & make up
    • snowman & snow man

Verb Forms: This along with the next section are what stock houses are referring to when they say “stemming” which is the technical word for recognizing all forms of a word as being the same thing.

  • I first did two comparative searches on common regular verbs:
    • frustrating, frustrated, frustrate & frustrates
    • baking, baked, bake & bakes
    • hiking, hiked, hike & hikes
  • And three comparative searches on common irregular verbs:
    • running, run, runs & ran
    • standing, stand, stands & stood
    • sleeping, sleep, sleeps & slept

I’ve found stock houses run the range from doing an excellent job of stemming, recognizing all forms of a keyword to not doing any stemming at all.  Many stock houses fall somewhere in between, recognizing all forms of regular verbs, but not non-verb forms, or recognizing only only common words.

Other Forms of the Word:

  • I first did five comparative searches on forms of common words:
    • frustrating, frustratingly & frustration
    • amazing & amazement
    • happy & happiness
    • angry & anger
    • joy, joyful, joyfulness & joyous

You’d be surprised how many stock houses that claim to do stemming, but only do it for verbs, but not other forms of the keyword.

Phrases:  There are plenty of “keywords” that are made of more than one word: bald eagle, New York, salt and pepper, mandarin orange, etc.  In fact, I suspect that they are more useful keyword phrases than there are single keywords.

The first thing I check is whether the stock house keeps keyword phrases together, or just breaks them up into individual keywords.  Then, regardless of whether they divide them up or not, I do a series of searches to see how the search engine finds searches for keyword phrases, or multiple keywords.

  • Here I did three comparative searches on 8 different forms of common keyword phrases:
    • “wooden chair”, wooden chair, wooden + chair, [wooden and chair], “chair wooden”, chair wooden, chair + wooden, [chair and wooden]
    • “Scottish terrier”, Scottish terrier, Scottish + terrier, [Scottish and terrier], “terrier Scottish”, terrier Scottish, terrier + Scottish, [terrier and Scottish]
    • “New York”, New York, New + York, [New and York], “York New”, York New, York + New, [York and New]

By seeing which combinations returned the same results, it is pretty easy to see how the stock house deals with multi-word keyword phrases.  Some return different result for “New York” and New + York, which means they value keyword phrases and its to your benefit to make sure you have all relevant common keyword phrases saved together.  If all the searches return the same results, it means having New York as a single keyword is no more valuable than New and York being separate keywords.

Words in a Keyword Phrase: Here what I’m looking for is whether a search for an individual word will return an image that does not have the word as an individual keyword, but only as part of a keyword phrase. For instance, let’s say you have an image with the keyword phrase blue whale but not the individual keyword whale. Will this image be found if the buyer searches for whale?

  • For this I did two comparative searches:
    • York & New York
    • North & North America

Yes I did this on keyword phrases that shouldn’t be broken up into individual keywords, Yet they are useful for my purposes here.  If there are more results returned for New York than York, we know that the search engine is not looking inside keyword phrases for individual words.  This means you will need to keyword blue whale and whale, or bird nest and nest, or apartment building, apartment and building.

Prepositions: What does the stock house do with the and in bride and groom? Or the in in lying in bed?

  • For this I did three comparative searches:
    • “salt and pepper”, salt and pepper, “salt pepper” & salt pepper
    • “waiting for”, waiting for & waiting
    • “waiting in line”, waiting in line, “waiting line” & waiting line

If the the stock house returns the same images for these sets of images, then its clear that they completely ignore prepositions, so you don’t need to use them in your keywords.

Periods:  In general, stock houses either treat periods as valid characters to search for, or they just ignore the period.  When they ignore the period, they will either replace it with a space or delete it without leaving a space in its place.

  • I did three comparative searches on common keywords and keyword phrases with and without periods:
    • Mt Everest & Mt. Everest
    • Washington DC & Washington D.C.
    • NY & N.Y.

Apostrophes:  These are either treated as a valid search character, or deleted.

  • Then I did two comparative searches on common keyword phrases with and without apostrophes:
    • chef hat, chefs hat & chef’s hat
    • baker dozen, bakers dozen & baker’s dozen

Hyphens:  We earlier investigated in closeup, makeup and snowman in their one and two word variations.  Now we will compare them with their hyphenated versions.   Hyphens, like periods, are either left as valid search characters, deleted or replaced with a space.

  • If there are different results between the one word, two word and hyphenated variations, then we need to do two more searches for each:
    • close-up + closeup & close-up + close up
    • make-up + makeup & make-up + make up
    • snow-man + snowman & snow-man + snow man

If the results for close-up equal the results for close-up + closeup then they treat the hyphen as if it wasn’t there and push the two parts of the word together with no space.  However if the results for close-up equal the results for close-up + close up then they simply treat the hyphen as a space.

Commas and Semicolons: Since these are the common keyword separators, never include them within keywords or keyword phrases as they will be treated as separators.

Capitalization:  Proper Capitalization is almost never necessary, but my personal opinion is that it looks more professional to use proper capitalization for keywords. Several stock houses also make clear in their recommendations to contributors.

  • Just for completeness I did two comparative searches:
    • Scottish terrier & scottish terrier
    • Washington DC & washington dc

Abbreviations:  Not much explanation needed here.  Does the abbreviation return the same results as the full word?

  • Here I did three comparative searches of keywords and their common abbreviations:
    • apartment & apt
    • New York & NY
    • Mount Everest & Mt Everest

Individual Letters:  What does the stock house do with the keyword “A“?

  • Here I did five searches two in quotes and three without quotes:
    • “A”
    • A
    • “A B”
    • A B
    • A + B

Some search engines ignore single letters, others treat it as a complete keyword.  Good to know if you have an image of a Model T.

Numerals:  Can they handle numerals  (yes, they can, but can they handle a single numeral?)

  • Like with the individual letters I did searches with and without quotes:
    • “9″
    • 9
    • “99″
    • 99
    • 999

This is important to know if don’t want your beautiful picture of a number 9 returned with every image of 9 year old children.

Keyword Separators:  This has nothing to do with search results, simply something you need to know when submitting keywords with images. The most common separators are:

  • comma space (far and away the most common separator)
  • semicolon space
  • space
  • comma (no space)
  • semicolon (no space)

Alphabetization of Keywords: Really the only reason to check this is to see if your keywords are weighted based on order. If they are alphabetized, then there is no chance they are weighted by order.

Keyword Weight: I need to preface this with the fact that very few stock houses do any weighting of keywords. But for the few that do, it means that the search engine will use an algorithm to rank the order of the images shown in results.  Images will place higher in the results if they have keywords that are weighted to match the search.  The weighting can be based upon:

  • Keyword order, when when the first keywords you list are considered more important.
  • Number of keywords, where fewer keywords means each is more important to the search algorithm.

Number of Keywords Allowed:  This is usually done by either number of keywords (where keyword phrases count as one keyword) or by the total number of characters.

Disambiguation of Homographs:  These fancy words mean, what does the stock house do to differentiate an elephant’s trunk from a car trunk, or a cardinal the bird from a cardinal the religious figure?

  • Simply search for trunk

Do all kinds of trunks appear in the results, with no prompts helping you to find the right trunk?  Well then, no disambiguation.  This means you better make sure you have whatever other words the searcher might use to narrow their results.

Faceted Searches:  These are the additional choices that most stock houses offer buyers when they are searching for images.  Common options are:

  • Orientation (horizontal, vertical, etc)
  • Category (Nature, urban, street, landscape, etc)
  • Age (child, adult, etc)
  • Ethnicity

What you, as the contributor need to know, is if a buyer chooses not to use the faceted search, but instead keyword searches for “nature + horizontal”, will they find your image that doesn’t have those keywords.  The answer usually is no, but you should do searches for the same word, once from the faceted search and once from the keywords to see if the same results are returned.  If yes, then you don’t need these keywords.

——–

I do hope you found this understandable and useful.  If you are reading this paragraph, I congratulate you for wanting to truly understand how to keyword your images.  Learn these situations, understand the concepts involved, and you will surely see improved sales of your stock images.

Share and Enjoy:
  • Print
  • Digg
  • StumbleUpon
  • del.icio.us
  • Facebook
  • Yahoo! Buzz
  • Twitter
  • Google Bookmarks
This entry was posted in Illustration keywording, Image Keywording, Image keywording software, Stock House Thursdays, Stock Photography, Uncategorized and tagged , , , , , , . Bookmark the permalink.
  • http://blog.keywordsmart.com/2011/12/08/image-keywording-and-image-search-at-dreamstime-com/ Image keywording and Image Search at Dreamstime.com

    [...] Blog Home ← The Methodology Behind the Analysis of Image Keywording for Stock Houses [...]

  • http://blog.keywordsmart.com/2011/12/15/the-search-engine-at-fotolia-com-and-how-to-keyword-for-it/ The Search Engine at Fotolia.com and How to Keyword for It

    [...] For those of you with some gumption, you should read the complete introduction and methods behind these analyses: Image Keywording for Stock Houses: The Methodology. [...]

  • http://blog.keywordsmart.com/2011/12/22/nationgeographicstock-keywords-and-search-for-stock-photographers/ NationGeographicStock: Keywords and Search for Stock Photographers

    [...] For those of you with some gumption, you should read the complete introduction and methods behind these analyses: Image Keywording for Stock Houses: The Methodology. [...]

  • http://blog.keywordsmart.com/2012/02/02/masterfile-how-they-search-for-keywords/ Masterfile: How They Search for Keywords | KeywordSmart Blog

    [...] For those of you with some gumption, you should read the complete introduction and methods behind these analyses: Image Keywording for Stock Houses: The Methodology. [...]

  • http://blog.keywordsmart.com/2012/02/23/how-to-keyword-your-images-and-video-for-pond5/ How to Keyword Your Images and Video for Pond5 | KeywordSmart Blog

    [...] For those of you with some gumption, you should read the complete introduction and methods behind these analyses: Image Keywording for Stock Houses: The Methodology. [...]

  • http://blog.keywordsmart.com/2012/03/01/effective-image-keywording-for-canstockphoto/ Effective Image Keywording for CanStockPhoto | KeywordSmart Blog

    [...] For those of you with some gumption, you should read the complete introduction and methods behind these analyses: Image Keywording for Stock Houses: The Methodology. [...]

  • http://blog.keywordsmart.com/2012/03/15/search-and-keywording-at-bigstockphoto/ Search and Keywording at BigStockPhoto | KeywordSmart Blog

    [...] For those of you with some gumption, you should read the complete introduction and methods behind these analyses: Image Keywording for Stock Houses: The Methodology. [...]