Ch-Ch-Ch-Changes – Stock Photography over Time

I’ve added a very cool new feature to the picNiche rating system today. I’ve been waiting to add this one for months but didn’t have quite enough raw data to make it worthwhile until now.

It’s a little more complex than usual so I figured it would be useful to highlight the key points, and some very interesting bits of note here.

When running a keyword phrase through the picNiche search engine, it will now also check for any historical data for this same search, and if previous searches exist, it will draw a graph.

Here’s an example:

Food – Historical Data

As marked on the legend, the 3 plots are Downloads, Number of Images, and the calculated picNiche Rating. All three are represented as the change against their average for this word or phrase, so 100% will always be somewhere near the middle. This is part of the mean/mode/median thing many of us remember being confused about in school 🙂

What does it mean?

When I was first playing with the idea of charting historical data, I was outputting the raw figures, and at that point I must confess I wasn’t really seeing useful information. Though as soon as I changed to this percentage-change metric, it all became clear…

There are a few important things to be seen here:

  • The steepness of each plot (line) indicates not only it’s direction, but it’s speed of growth or decay.
  • The difference between downloads (demand) and the number of images (supply) can clearly be seen by their relative steepness.
  • The rating is obvious as a result of this supply/demand relationship.
  • Clearest of all is to see when during the year demand or supply are at their highest or lowest.

Note that it doesn’t matter ‘when’ or even ‘if’ the lines cross, the rating is unaffected, it’s all about the gradient (the steepness of slope).

How does this help us?

This is where it got really useful, I tested a few of the most regularly run phrases and was able to see at a glance not only how well they’ve been doing over the last year or so, but also to have some idea what they’re doing now.

Since the ‘stock photography’ market is a lot less volatile than the financial stock markets, this can be used as a guide to what they’re likely to do over the next few months.

Here’s some more detailed examples:

I decided first to take a look at very busy stock subjects, starting with business related images:

Business Man – Historical Data

Business Woman – Historical Data

Both the above show a fairly steady growth, with a library clean-up at the end of 2008 causing a slightly better sales growth (probably as a result of buyers finding the more suitable images faster).

This is nice to know, these two subjects are clearly still bloated but good portfolio management clearly results in a good return.

Feeling SAD? (Seasonal Affected Downloads)

The next obvious step was to see when the peaks and troughs in seasons occur, to optimise shoot planning and uploading. (Limited datapoints to draw firm conclusions, but they show definite indications).

Spring – Historical Data

Spring sales pick up a little early, September/October but downloads didn’t even start to rise till around January.

Summer – Historical Data

Summer seems to have less lead-time than spring. Though sales persist longer once the season arrives.

Autumn – Historical Data (or Fall)

Autumn and Fall seem to show very steady growth all-year through without any significant change, this could be a reflection of the limited and generic range of image concepts usually associated with this season (Falling leaves and raindrops for example).

Winter – Historical Data

Winter starts to rise just before spring sales (Sept/Oct), but rises very sharply till the new year then ceases almost immediately.

If I were to shoot stock images to focus on only one season, I’d shoot for spring, sadly I live in the UK, and shooting in September/October would get me far too many rainy days, great for April Showers, but of pretty limited value for globally available spring shots 🙂

Big Ticket Events

Onto the last few big events (in the western world at least) and an interesting contrast I wasn’t expecting to see here:

Easter – Historical Data

I’d have liked a few more datapoints here before Jan 09, but even here it’s clear to see that sales-growth starts significantly earlier than the available images do.

Christmas – Historical Data

Both Septembers show a small rise in Christmas images before the image count rises over a month later (very suddenly last year, I expect it will show the same this year). There’s clearly about a month here where this year’s buyers are finding last year’s images.

New Year – Historical Data

New Year on this dataset is a tough call, it looks like oversupply to me, though perhaps that will be resolved this year as both sales and image count seems to have shown matching growth from September.


I’m obviously a stat-junkie, and I’m thinking I perhaps got carried away digging through this data (and I’m still trying to figure out a way to do time-based non-cumulative comparisons over associated phrases without breaking my server), but I know when it comes to planning for, and preparing to produce my stock photography, it’s given me a few solid tips:

  • Watch for growth and make images for those buyers
  • Oversupplied subjects will hurt my RPI
  • Shoot earlier!
  • Process earlier!
  • Upload earlier!

The other bit of good news from all these charts (and many others I’ve run) is that sales growth across many topics is consistently slightly higher than portfolio growth, showing a good forecast for microstock’s future.

You can now see these charts when you run any search on picNiche (if previous data exists). There’s a lot of power in relating the charts for similar topics to each-other. Just don’t get too carried away, pick up your camera and go shoot something, staring down the internet-hole doesn’t produce photos 🙂

This entry was posted in Microstock, Tips and tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , . Bookmark the permalink. Post a comment or leave a trackback: Trackback URL.


  1. Posted October 29, 2009 at 6:52 pm | Permalink

    Awesome stuff Bob! Very cool use of historical data and great to see downloads and image counts overlaid. Kudos!

  2. Posted October 29, 2009 at 6:57 pm | Permalink

    Very nice, great work and thank you.

    Yes, maybe you got a little carried away, but nothing wrong with too much data. I would always prefer to have more info than I need, because once I get a little, I always want more.

  3. Posted October 29, 2009 at 7:18 pm | Permalink

    Ch-Ch-Ch-Changes from shrek OST?? 😛

    • bobbigmac
      Posted October 29, 2009 at 7:23 pm | Permalink

      lol, no idea if it was in Shrek (though that film did rock :), I was thinking more of David Bowie 🙂


  4. Posted October 29, 2009 at 7:42 pm | Permalink

    yeah that’s it! wonderful song!

  5. Posted October 30, 2009 at 1:57 am | Permalink

    Wow! (I think???)

    Am I interpreting this correctly? in all but the easter graphs demand for images outgrows the number of new images – surely that’s too good to be true?, and only a few times a year (pre Easter upload) does the increase in images available match growing sales.

    Or is that just a feature of the graphs being normalised? otherwise it seems despite new image rates growing towards millions each few months growth in demand is still growing faster? and the picniche rating is, in general, on the way up for lots of subjects. (go on make me a graph of average rating against time across all your key phrases 🙂

    marvelous if true, and it will be a sad day when the rank starts to drop year to year… too much microstock not enough buyers.

    – ziggy stardust (rating 156) was better I think, it all went a bit downhill in the 80’s

    • bobbigmac
      Posted October 30, 2009 at 12:19 pm | Permalink

      You’re right… In generic topics the demand ‘does’ outstrip the supply by quite a significant amount, though I don’t think exactly the same applies to all topics, there are many where the increase is much smaller, and likewise many with an oversupply (though it’s good to see this has been tempered somewhat in severely overcrowded topics by selective editing/acceptances.

      There is a definite positive lean towards sales growth, but I think they strength demonstrated here is probably exaggerated a little by my own choice of relatively carefully managed topics 🙂

      I’ve been trying to build an overview graph showing the rating across all queries, and my early results show a substantial rise in overrall rating over the last 12 months, but I’ve so far been unable to reliably factor in how much of this is a result of people who use picNiche getting better at choosing what they search for.

      I’m thinking I will have to do a ‘representative sample’ graph similar to how economists estimate the CPI (consumer price index). But at the moment I’m not certain my dataset is robust enough to be sure of the result. Some day though 🙂

  6. DAN
    Posted October 30, 2009 at 12:42 pm | Permalink

    This can only make the site better. What’s in store next?

    • bobbigmac
      Posted October 30, 2009 at 1:49 pm | Permalink

      Who knows? I’m always open to suggestions 😀

  7. Posted November 20, 2009 at 10:20 am | Permalink

    Hey Bob…
    I’m a bit in late but I’d like to make my congraz, stunning work!
    Good steepness to all of you 😉

3 Trackbacks

  1. By November News Digest | Microstock News on November 21, 2009 at 4:01 am

    […] added graphs of historical data to their stock photo demand […]

  2. […] picNiche Blog Blog for and it’s toolbars « Ch-Ch-Ch-Changes – Stock Photography over Time […]

  3. […] parole chiave ricercate. I dettagli nel blog di PicNiche dell’encomiabile autore Rob Davies, Ch-Ch-Ch-Changes – Stock Photography over Time. Lookstat permette ora di creare collezioni arbitrarie delle proprie foto per poterne seguire […]

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=""> <s> <strike> <strong>