Top 2000 Image-Buyer’s Searches vs Keyword Popularity

I’ve been playing around with some of my data the last few days (having a bit of an overhaul to get picWorkflow running faster) and before I archive off some of my data-tables I thought this one might be of interest to you :)

The list shows the top 2000(ish) short-form searches (no more than 2 words) made by image-buyers on my various search tools, and the rank (i.e. relative popularity) of that search in my dataset of a somewhere near a million searches. The far-right column shows it’s comparison to the popularity of that word’s usage on stock images (collated from various agency’s over the years, across about a quarter million images).

This is essentially an epic update of my previous post from 2 years ago: TOP 1000 Stock Photo Buyer’s Searches – THE List, that post contains a lot more details about the data quality, relevance, demographics and more.

How to read the data?

A green entry suggests the search is more common than the keyword, a red entry suggests the keyword is more common than the search. It’s not bulletproof as I haven’t vetted the raw data too heavily, but it IS a good quick-reference to many of the most popular stock photo ideas and whether they’re under or over supplied.

What do you think?
Any gems in there for you? :)

Top 2000(ish) Image-Buyer's Searches versus Applied Keywords Popularity

Please feel free to share with others; published under Creative Commons (CC-BY) so you must state it’s by me if you republish it (or derivatives). Would be great if you can drop me a link-back to here or the homepage too :)

  • Twitter
  • Facebook
  • Google Bookmarks
  • Digg
  • del.icio.us
  • StumbleUpon
  • LinkedIn
  • Reddit
  • Live
  • Technorati
This entry was posted in Research and tagged , , , , , , , , , , , , , , , , , . Bookmark the permalink. Post a comment or leave a trackback: Trackback URL.

7 Comments

  1. andre
    Posted February 27, 2013 at 4:49 pm | Permalink

    thanks for sharing this, this looks super interesting. one question though, I am not sure if I understand “usage rank”. I guess this is measuring how often a keyword is used in images. however “beautiful woman” is throwing me of. Its usage rating right beside “Brazil” but gets 2 million more search results.

    again thanks for putting this together for the community.

    • bobbigmac
      Posted February 27, 2013 at 4:57 pm | Permalink

      aye that’s right. Best guess is search engine algorithms (depending on which site you’re searching for images on) normalising “beautiful woman” to related phrases like “beautiful girl”, “hot girl”, etc etc or just to “beautiful” + “woman”. Both beautiful (often incorrectly) and woman are individually used significantly more often than Brazil. With algorithms normalising terms like this it will give far more search results. In the context of the list above (presence of it as a single keyword “beautiful woman” on images) the number of search results in the various agency search engines won’t be an exact match to the number of times a keyword is actually used. This is generally true for any multi-word phrase, though including one as a single keyword gives the engine a ‘hint’ that the image is about a woman, who is beautiful, rather than an ugly woman looking at a beautiful landscape :)

  2. Nagesh
    Posted March 31, 2013 at 4:42 am | Permalink

    Hi,
    Is it possible for me to get an excel sheet version of this? I am looking at using this data to fine-tune my stock photography.

    Thanks,
    Nagesh.

  3. Julie
    Posted September 2, 2013 at 6:52 pm | Permalink

    Hi Bob,
    I think this is such a great tool, thanks for sharing :)
    I was just wondering why there is some words missing in the list.. ex: #20, 26, 33, 35… just curious.

    Thanks again!

    j

    • bobbigmac
      Posted September 2, 2013 at 9:09 pm | Permalink

      The missing numbers apply to words which are in one list, but not the other; or were spammy terms with artificially raised search profiles (like “boobs” etc). This list covers the range of the top 2000, but actually contains a lot less than 2000 because the entire range was not all in both databases.

Post a Comment

Your email is never published nor shared. Required fields are marked *

*
*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>


− 5 = four