We all know that online readership is a nigh impossible nut to crack.  Why? You ask.  Well, where do I start?  There’s the fact there’s no one authoritative source for data.  Nielsen NetRatings is a well respected name and self proclaimed, “global standard for Internet audience measurement.”  They use country-specific panel based research to statistically model monthly user figures for websites.  All well and good but this involves primary research, which makes the data cost prohibitive.  It's fine when you're only dealing with the UK market but for an agency like ours that operates on a global basis, serving more than 48 countries, it just gets ridiculous.  The other big kid on the block is ABCe who is, “independent, not-for-profit and a subsidiary of the Audit Bureau of Circulations (ABC)” which by its very nature makes them very proper, like.  They basically log unique monthly users for each site.  It’s very accurate but doesn’t take into account the fact that one person can log in from many computers and essentially counts people twice (a bit like stacked opportunities to see compared to our cross-readership reach research).  In fact, remember the dingdong earlier this year between the major UK papers about their online readerships? If not, have a read here.

So they are the two gorillas in the room.  Other options include: Alexa (biased sample because it relies on the Alexa toolbar but can give pretty accurate international data); comScore (self-selected sample using software download); and Compete (browser toolbar dependent, primarily US only).  And the list goes on.

So, nothing really comes close.  Then there’s the next challenge… mainstream media measures audiences per issue, so a daily paper uses daily readers, a weekly its weekly audience and so on.  Online media is different.  Once it’s up there it pretty much stays there.  More sophisticated business gurus might call it the ‘long tail’ effect.  So do you take daily, weekly, monthly or infinity and beyond readers?

How about the fact that, like mainstream media uses a figure for the whole paper, you get a readership of the whole site.  Hands up who has read a paper cover to cover *I’ve got my hand up*.  Now hands up who has read an internet site in its entirety?  Didn’t think so.

The answer is that ABCe is as close as we can get to accurately measuring online readership today. It uses the Joint Industry Committee for Web Standards (JICWEBS) industry agreed metrics.  However because it looks at it globally, on a monthly basis and looks at the whole site, the figures are huge.  Literally massive.  So we end up with figures that we, and the rest of the industry, just don’t trust.  For example, a piece buried deep inside BBC Online ends up with a figure based on anyone in the whole world who reads any one page on the BBC site within a whole month.  Ludicrous.   What’s important though is what we, as PR people, do with this data.

Following a consultation period, we are in the process of changing the way we output online readership figures.  Based primarily on the ABCe data, but using other sources where they are accurate, we are going to model the long tail of an article into the readership.  The piece has most impact once it’s put up on a site and that fades away, so monthly is too much.  We are also factoring in where people in the worlds are viewing the article.  Finally, we are taking into account that one person may log on to the site multiple times, from different computers.  What it means for our clients is that the once very impactful online coverage that boosted their reach figures is going to drop down to something far more realistic.  But realistic is good if it’s more credible.  

So, we’re not there yet, I’m afraid nobody is.  But as long as we can supply our clients with better data to help them communicate their successes we’re winning in my eyes.