Controversy with the Image Count at Fotolia
Posted by Lee Torrens
The number of images that Fotolia reports on the front page of their website has been the subject of much debate and accusation. Since the Fotolia V2 upgrade the figure has dropped significantly, further fueling the debate. I’ve put together in this post all parts of the puzzle. The details of the debate, the indisputable facts, and the official Fotolia response.
The Controversy
Prior to V2, the Fotolia front page showed more than 3.3 million images in their online library. Many speculated that this figure was artificially high. They couldn’t understand how Fotolia could have more images than the market giants iStockphoto and ShutterStock. Simple research turned up evidence that supported the suspicions. The top selling images at Fotolia had lower sales than the those at other microstock websites. Also, common search terms returned far fewer results at Fotolia in comparison.
The Facts
Fotolia started accepting contributions in November 2005. They initially incentivised contributions by paying $0.20 for each accepted photo. They overtook iStockphoto and ShutterStock in total image figures in July 2006, being the first to pass the 1 million milestone.
Searches for popular keywords at Fotolia (V1) did indeed return significantly lower numbers of photos than their competitors. However, search terms that were not so popular resulted in less distinctive results.
Prior to the V2 upgrade the website reported over 3.3 million images online. Fotolia announced that the V2 upgrade would involve removing images over 18 months old that had not yet been sold.
After the V2 website was launched, Fotolia’s front page reported 1.9 million images.
Controversy Explained
The statistic on the Fotolia website represents the number of images available to purchase. It does not include deleted, free, or rejected images.
Prior to the V2 upgrade, Fotolia programmed their search results to include only the most relevant results from their database. This was an intentional solution to limit the load on their servers, something the V2 upgrade has addressed. Subsequently, the same searches performed now with V2 provide more results.
Search engines are vital competitive parts of microstock websites, and they’re all complex and diverse. The results from identical searches with different search engines will always provide different results. Fotolia’s V1 search engine provided different result to what it’s V2 engine does now.
So?
The conclusion is that there’s a logical explanation for the figures at Fotolia, but previous and current. It was high before V2, but during the upgrade they removed a lot of old non-performing images as part of a quality over quantity strategy. Fotolia no longer have the largest portfolio of all the top microstock agencies.
The controversy was based on solid evidence, though digging a little deeper it turns out things are more complicated and the evidence has a logical explanation. Everything with the Fotolia numbers is above board. And that’s good to know.
Further Analysis: Andy Goetze at StockPhotoTalk
Much of the research for this article was derived from the contributions of Stephen Finn (fintastique) at the Microstock Group forums.
Posted July 4th, 2007 by Lee Torrens
Hmm, what was the research? I was a developer in my early career, and worked at the largest enterprise search company for a long time, and have always been puzzled by Fotolia’s interesting results… I’ll call bullshit.
Do a search for ‘brown dog’. 471 files returned. Many searches return far more results, so they aren’t running up against a max number. That’s all they have. Period.
Shutterstock: 2667 results
iStockphoto: 2265 results. iStock tells you the total number of files found it 2265. It only returns the first 1000 to ease pain on its search engine. The rest can be pulled if the user requests.
Next try just ‘dog’.
Fotolia: 8762
Shutterstock: 15745
iStockphoto: 16671
That’s more than 50% less files in pretty much any test case I could come up with. A search engine isn’t rocket science. It’s just a database with keywords in it. If the keyword is there, the image is pulled. If it’s not, it doesn’t. Especially with a single word like ‘dog’. What makes an image more or less relevant when its tagged with dog? Maybe the position of the keyword within the list? So if it’s not in the first 50 keywords it’s not returned? Doubt it.
Biased or not–being the iStock head of PR, I think (?)–Kelly has a point. I can’t see how you can adequately explain that popular keywords did and continue to result in fewer images returned on a website with supposedly more images in total. I get that Fotolia has removed over a million images with the recent upgrade, but I still don’t get the discrepancies. Can the Fotolia collection be so substantially different as to have a broader subject matter range?
All I can say is that I’m giving FT a break until they get their act together. Maybe I’ll check back in a month or two. They haven’t made it worth my while as of late.
Notice how Fotolia has now locked their forums in to only members. I have the feeling they don’t want the general public to know how bad the situation is.