Big UGC (User Generated Content) websites know a lot about users’ behaviour. But one of the things they’re most acutely aware of is that users are not going to post content if they feel no one’s interested. Each site has its own way of trying to reassure users that what they post is being viewed by other interested parties, thus maintaining a desire in contributors to keep feeding the mighty content machine. But over time, some sites’ methods have become increasingly dubious. In particular, there’s been a tendency on the part of various UGC facilities to exaggerate the amount of interest in users’ work in order to keep them motivated.
How they accomplish this varies. Some sites incentivise users to essentially consume their own work. WordPress.com, for example, gives automatic self-promotion to anyone who ‘Likes’ the posts of others or ‘Follows’ someone else’s blog. The result is that users are ‘Liking’ and ‘Following’ blogs they’ve never read or even visited – just so they can promote themselves. WordPress aren’t bothered. The system motivates millions of gullible users. What does it matter if the interest in them is fake as long as they think it’s real and keep the content coming?
Until the early part of this decade, Flickr’s system was pretty straight up, and it could afford to be. Unlike someone’s rambling poetry or self-absorbed journal on WordPress, wide sections of the public do actually care enough about photographs to seek them out, and Flickr had a huge presence on Google Images, as indeed it still does.
Back in 2012, Flickr was still able to use code to convert many image clicks on Google into actual page visits on the user’s Flickr account. But in January 2013, a huge, sweeping and devastating change to Google Images meant that suddenly, the source sites for all photos appearing on Google’s search display were being bypassed. Surfers were accordingly much more likely to download the images directly from Google. Flickr users’ pics were still being viewed on (and downloaded via) the world’s number 1 search site, but now, those Flickr users were not getting anywhere near as many actual hits, because Google was more typically bypassing the original page – stealing Flickr’s hits, essentially. That would inevitably impact negatively on Flickr users’ stats.
I guess there followed a pretty grim time for Flickr in which motivation among users began to tail off as their stats headed downward. But in July 2013, Flickr users started to notice a huge increase in their stats, without any apparent reason. After a lot of speculation, Flickr confirmed it had changed the way stats were measured, and that a ‘view’ no longer meant someone had loaded the full image on the user’s Flickr page. They might instead have seen the image in a remote location, or alongside other users’ photos on Flickr’s search pages. Certainly a more liberal interpretation of the word ‘view’, but technically, provided the image was actually loaded somewhere within eyeshot of the viewer, still arguably a fair statistic.
Of late, however, some experiments I’ve been doing with Twitter promotion have very strongly suggested that actually, at least some of the ‘views’ Flickr counts, are not real, and the image has not even been loaded.
One of the things I hate about Twitter is the way it’ll display by default images you upload directly, but default to hiding images uploaded to other sites, such as Flickr. If you share your Flickr upload on Twitter, your followers and people using Twitter’s search functions can view your photo without clicking through to Flickr, but they have to at least click the Tweet to view.
For some time I assumed that the boost in Flickr views that came when I shared a pic on Twitter did not primarily comprise click-throughs, but was instead mainly made up of views on Twitter. I did however, imagine that people had at least clicked the Tweet and seen the image.
But then I came up with the idea of sharing my Flickr photos on Twitter, but additionally uploading a lower res version of the same image directly to the Tweet. So the Tweet would no longer have a hotlink to my original upload on Flickr – it would instead contain the actual photo, which would of course (because it was uploaded to Twitter and not a third party site) display by default. Great! So it’s the best of both worlds. I’m now sharing my Flickr pics on Twitter, but I’m tricking the system with a direct upload, so my followers will automatically see the image in their timelines.
But how does that then work as regards Flickr views? Because once you upload a pic directly to a Tweet, it disables the ‘Twitter Card’, and the original hotlink from Flickr no longer loads. What Twitter followers see is a completely different image. It might be the same picture rescaled, but it doesn’t necessarily have to be. I could, for example, upload a photo of a guitar to Twitter, and link to a photo of a cat on Flickr, and in a case like that I’m definitely NOT getting views for the cat photo unless people do genuinely click through from Twitter to my Flickr page.
To cut a long experiment short, I tested click-through rates by linking from Twitter to two different image sources. One source was my Flickr account, and the other was a blog whose stats are monitored by Google Analytics. I used old and activity-dead uploads to Flickr which are not publicly searchable, in order to minimise the prospect of the experiments being corrupted by visits from elsewhere.
The result? Well, the hits registered by Flickr invariably dwarfed those registered by Google Analytics on my blog. Typically, Flickr in fact registered at least ten times the number of hits registered on the blog. My conclusion? Flickr is counting the loading of a Tweet containing a link to a Flickr image as a ‘view’ – even if there’s a totally different image in the Tweet and there’s no way the Flickr image could be visible to a Twitter user without them clicking through to Flickr. It’s a completely fudged view, and you cannot in any way trust ‘views’ registered on your Flickr account after Twitter shares as an accurate measure of reality.
So take Flickr’s ‘view’ stats with a pinch of salt, because with at least some of them, the ‘viewer’ has categorically not seen the image at all. It’s merely Flickr’s way of trying to persuade you that people are taking an interest in your work.