Sunday, May 23, 2010...11:26 am

Modern media is rubbish #5: how to misrepresent the uSwitch broadband survey 2010

Jump to Comments

From early this month (because it’s been knocking around the house and I’ve only just got around to glancing at it before I put it in the recycling bin) – here’s the Guardian Money report on uSwitch’s annual broadband satisfaction survey.

What’s wrong with this piece of simple reportage?

It couldn’t be simpler, really. In fact, this is the kind of job you should give to the intern or editorial assistant, because all it involves is looking at the uSwitch survey results (right) and writing them up for your reader. Because readers can’t you know, make sense of the results by looking at the table itself.

Well, certainly Miles Brignall of the Guardian couldn’t make sense of the results by looking at the table itself.

Here’s what he said:

The survey, based on responses from more than 7,500 broadband customers, found that 92% of O2‘s customers are satisfied overall – compared with just 44% of Orange users who said they were happy with their firm’s customer service

So, that’s clear – O2 has a whopping 92% satisfaction result, while Orange is terrible, with just 44% satisfaction.

But this isn’t right. Later on, the story says:

Orange has continued its dismal run. Despite a brief respite last year, it is back at the bottom with a 70% score for overall satisfaction.

I had to go back and read everything again, because I could have sworn Orange had a lousy 44% satisfaction rating. Look, I’m right – there in the little graphic table in the story (right) is a list of stats that shows Orange with 44% “Service satisfaction”.

But hang on – it also shows O2 with a 79% satisfaction rating. I thought it said 92% earlier. What’s going on?

What’s going on is that the Guardian story is mixing two entirely different metrics – overall satisfaction and satisfaction with customer service. It does actually say so in the text (though you could be forgiven for missing the detail). And I think the author knows the difference – but it’s confusing for the reader. Especially as the helpful graphic that the Guardian uses to show what it’s talking about has made the same mistake.

It’s as if the sub-editor or designer who was asked to create the graphic had a quick look at the story, saw the reference to the 44% Orange rating and thought “ah, that’s the bit of the uSwitch survey they’re talking about – I’ll pull that out of the results”.

Instead, whoever put this page together should have used the overall satisfaction rating – because that’s the ultimate result. And the writer should have compared like with like, instead of muddying the water and confusing both the Guardian‘s production staff and its readers.

And it would have been helpful to have a proper comparison chart of how the companies did compared to 2009. Instead, the story just gave us the change for AOL and TalkTalk (up 6% and down 3%, respectively). Checking on the uSwitch web site showed that O2’s 92% was the same as last year – but the numbers are not available for Orange, Virgin and BT.

Come on – if you’re going to look at a survey, try to be organised and consistent about it. Otherwise just print the survey results untouched and let the readers interpret it. It’ll save time and effort, and probably be more accurate.


  • There are two very different types of broadband offerings from O2.
    Their Standard, Premium and Pro packages, available only at exchanges that O2 has ‘LLUed’, have a good reputation. Access, the only scheme available at other exchanges, is best avoided.
    Obvious questions: how was the sample of O2’s customers selected? What was the ratio of LLU to Access?
    Just thought I ought to try to muddy the waters a bit more!

  • Hi Chunter – thanks for the input, and some interesting and relevant points…

Leave a Reply