I tend to be unmoved by the declining state of traditional media. Don’t get me wrong, I love reading the paper, but bankrupt publications and out-of-work journalists are evidence of nothing more than atrophied business models. People still want news; people still get news. Who cares if it comes on newsprint?
Nevertheless the current disruption has a real effect on our business; so much so that I’m frequently asked about the impact of layoffs, magazine closures, newspaper bankruptcies, etc. When a client asks these questions I tend to answer them by pointing to a recent development such as the closure of the Rocky Mountain News or the Washington Post’s decision to close its business section.
Intuitively most people understand that these developments translate into fewer opportunities for traditional media coverage but is that really true and, if so, by how many?
With these questions in mind I decided to count the number of articles appearing in technology trade media since 2004. To make this manageable I decided to search these publications (using Factiva and its ‘computers and electronics’ source list) for articles featuring the most frequently used word in the English language: ‘the’.
This is what I found:
- From 2006 – 2008 the number of articles in technology trade media with the word ‘the’ in it declined from a total of 135757 to a total of 92021
(Total number of articles in technology trade media featuring the word ‘the’ from 2004-2008)
- This represents a 32% decline in the number of articles with the word ‘the’ in it since 2006.
(Percent change in articles featuring ‘the’ since 2004)
Given the fact that it is virtually impossible to write an article without the word ‘the’ in it, this would seem to indicate that there are approximately 32% fewer opportunities for coverage in technology trade media today then there were two years ago.
To confirm this theory I decided to run the same analysis on the next two most frequently used words in the English language: ‘of’ and ‘and’. This is what I found:
- Articles featuring ‘of’ decreased from a total 132451 in 2006 to a total of 89298 in 2008
- Articles featuring ‘and’ decreased from a total of 134127 in 2006 to a total of 90544 in 2008
- All three words had an identical negative growth rate of 32% from 2006 – 2008.
This seems to confirm it; there are 32% fewer articles – and fewer opportunities – for coverage in traditional technology trade media today than there were just two years ago.
Later today or tomorrow I’ll post numbers and graphs for mainstream media, and then play around with some projections for 2009.
Great measurement idea – these results match what I have experienced anecdotally.
I’m looking forward to seeing your data for the mainstream media.
Thanks for the comment. I think we’ve all felt this change, but I’ve had an awful time proving it to people. Not sure this entirely does the trick . . .
Hi Nick – thanks for this interesting piece. One possible caveat I wondered about. Is there an external factor limiting the size of the 2007 and 2008 Factiva ‘computers and electronic’ corpus, and could it be distorting the effect?
For example, is there a delay by Factiva in making some materials available (as there is with Jstor, for example)? Or have some trade magazine publishers switched their electronic archiving deals since 2006? This is also fairly common, and can mean that back-issues of magazines can be stored electronically on different databases according to date. (Before moving into marketing communications I used to be a lexicographer, and ran into corpus problems like this a lot when looking for data.)
Alex,
Thanks for the comment. You raise an excellent point – one of several potential limitations and caveats.
I can’t give a completely accurate answer but I do know that when I first did this analysis 2-3 weeks ago the raw 2008 numbers were slightly lower than when I tried to replicate it yesterday. The increased volume didn’t change the percent decline, but it’s worth keeping in mind.
Incidentally I rechecked prior years and those numbers remained unchanged.
So, bottom line, yes there’s some distortion due to a delay in articles being uploaded but not enough – I believe – to change the conclusion.
In a day or two I’ll write about this and other limitations I see.