As I've said in previous posts, Twitter is a wonderful source of data for highly cost-effective, real-time market research; and yesterday's speech by David Cameron provides an ideal opportunity to demonstrate this.

The charts (below) show response to the speech given by David Cameron at the Tory party's spring conference on 28 Feb, 2010, in terms of popularity (a measure of positive tone) on an hourly basis.  The top chart reveals a ding-dong battle between Twitterers on both sides of the political ocean.  HMS Cameron's supporters fired the first salvos, with HMS Brown's responding shortly after, at 16:00, before unleashing a broadside at 20:00.  HMS Cameron might have been holed but actually fought back with real aggression to quieten HMS Brown's guns, which fell silent at midnight.  HMS Clegg's gunners finally woke-up at 23:00, by which time the battle between the dreadnoughts was virtually over.

But if the battle between HMS Brown and Cameron was won by the latter, what about the fight between their supporting vessels?  The lower chart reveals that the Tory navy outgunned its Labour counterpart right up to 23:00 when the advantage swung back to the reds.  

So what about the morning after the battle, now that the smoke has cleared?  The latest position, taken just 20 minutes ago as I write, puts Cameron ahead of Brown by 3.5% to 1.1% though both trail to Nick Clegg, whose popularity currently stands at 5.8%.  The story is different for the parties, however, with Labour on 2.3% compared to just 1.2% for the Tories while the Lib-Dems are on 2.6%.

There is, of course, a long way to go in this electoral 'war' but these latest results from Twitter reflect what's happening in the polls, with Labour coming back hard.  They also show just how Twitter can be used for real-time assessment of consumers' views and opinions.

Charts show popularity -- a measure of positive tone as a percentage of overall coverage -- for the political leaders and their parties on 28 February, 2010

Following on from yesterday's post (immediately below), we can reveal that Gordon Brown's popularity has indeed dipped following the 'sympathy' vote gained in the immediate aftermath of the bullygate story.  At the same time, David Cameron's popularity has bounced back over the past 24 hours.  Significantly, these developments cannot have happened as a result of the bullying story dissipating as the percentage of tweets on the topic rose yet again to 28%.  This is twice the volume of the previous day and more than five times the volume on the day the story broke.

Interestingly, despite its leader's misfortunes, the Labour party has increased its lead over the Conservatives over the past few days, as the lower chart shows.  This appears to be an ongoing trend that reflects the opinion polls.

Yet again, this example demonstrates the value of Twitter as an alternative and valuable source of data that can be tapped into on a real-time basis.  While we're showing its use on a daily basis here, exactly the same analysis could be conducted more regularly, even hourly, making it a highly cost-effective real-time market research tool.

One of the most useful applications for Twitter is as a source of data for market research, where it offers two major benefits over traditional methods.  First, the data is available without the need for time-consuming and relatively expensive interviews.  The second major benefit follows largely from the first, which is that Twitter can be studied as frequently as needed.  Because Twitter updates so rapidly it is perfectly feasible to take findings daily -- or even hourly -- without the need to resort to panels. 

Last but not least, as an online version of word-of-mouth -- widely recognised as one of the most powerful forms of influence available -- Twitter is a powerful medium in its own right; and understanding what is being tweeted is just as important as understanding what the national newspapers are saying.

The ongoing spat over Gordon Brown and 'Bullygate' offers an excellent example of how Twitter can be used to test the water of public opinion on an ongoing basis.  The following charts show how the popularity of Gordon Brown, David Cameron, the Labour Party and the Conservatives has changed over the past few days since the (alleged) bullying story broke.

What's immediately obvious from the top chart is that while David Cameron's popularity on Twitter has dropped significantly over the past three days, Gordon Brown's has risen.  This is not to say that Brown hasn't had a lot of criticism on Twitter over the bullying claims, which he has, but overall his popularity has improved while Cameron's has fallen.

The position of the main parties is less clear, as the lower chart reveals, though even here the Labour party has improved its position over the past few days.  Indeed, as our Spin-Tweet dashboard reveals (see blog post immediately below), we are not surprised that the Labour party has been catching the Tories in the polls because we've seen this trend developing on Twitter.

So what might be happening to explain the above?  In the short-term Gordon Brown is benefiting from the sympathy vote from those who see him as a real person with flaws similar to many of us.  The question is whether such sympathy will persist as this latest episode gets more wind its sails.  Indeed, the volume of tweets associating Gordon Brown with bullying has increased three-fold, from 4.7% to 14.4%, over the past 48 hours. 

So, despite the current sympathy vote the Labour spin machine must be hoping this whole thing soon dies down because the longer it drags on the larger the chance something will come out that might really be damaging to the PM; and what better way to monitor that than with Twitter.  We'll keep you posted.

On Monday we reported that "Gordon Brown's popularity on twitter increased over the [last] week but the big loser was the Labour party, whose popularity on twitter plummeted by 4.11%."  Yesterday, 24 hours after our twitter findings were published, a Populus opinion poll widely reported in 'The Times' and 'The Scotsman' put Labour down two percentage points on last month while also reporting "some consolation for Mr Brown, with a sharp improvement in his personal rating – 41 per cent said they believed he was the best leader Labour could have at present".

Now, we're not saying that twitter can predict the future but the findings from our Tweet-tracker compare very favourably with those of this opinion poll and there's no doubt that twitter is a very powerful source of information for market research. 

Perhaps the most telling words above are "last month", referring to how often traditional opinion polls are typically carried out and reported.  A really huge benefit of using twitter for market research is its immediacy and cost-effectiveness.  While it's only practical and cost-effective to run opinion polls such as those produced by Populus on a monthly basis, twitter can be monitored, analysed and report using twitter every day.  In fact, we can do this hourly as demonstrated in our summary of what occurred on 6 January as the Labour coup unfolded on twitter in real-time (click here for more).

The real point, of course, is that while this example refers to politics it's just as applicable to any and every form of market research that might be required.  This might be instant reaction to a new advertising campaign, a product launch or a crisis; and a major benefit of twitter is that it's capturing the spontaneous views and opinions of thousands of people in real-time.

All of this has major implications for the use of twitter as a source of market research data.  Of course, some will undoubtedly point out that twitter is a biased sample and they're right.  But, and this is very important, it's also the case that those who twitter are typically advocates whose views and opinions are highly influential and of great value to marketers.  Indeed, twitter is the online equivalent of word-of-mouth, which is widely regarded as the most powerful form of marketing available, so understanding the sentiment behind what's being said is very powerful. 

See the latest results for the party leaders' popularity on twitter up to last midnight, on the Spin-tweet dashboard, by clicking here.

Over the past few days I've exchanged some interesting emails with David Reed, the editor of 'Data Strategy', around the research Spectrum conducted that revealed a lack of consensus among UK academics  for man-made global warming (see 'Lab' for details).  The exchange has been most enlightening, revealing Mr Reed's complete lack of understanding of basic statistics (as well as an inclination to rubbish anybody who dares to keep an open mind on a subject that he clearly supports with almost religious fanaticism).  In itself this wouldn't be an issue -- after all there are plenty of journalists who write well about subjects in which they're not experts -- but it is a problem when the journalist thinks he knows what he's talking about but clearly doesn't.  In such instances a journalist can easily mislead readers and given that Mr Reed edits 'Data Strategy' I find this very worrying.

In response to a point I made about sampling errors that had been conveniently ignored in his own publication's State of the Nation survey, Mr Reed emailed me with the following: "Surprising that you should want to take issue with data generated by an online surveying engine. No doubt it was designed by a qualified engineer [a dig at my engineering qualifications]. An answer that can only be given as yes or no, with no third option or ability to proceed without giving one or the other looks pretty binary to me. It's called a filter question. You clearly aren't bothered about including those in your own surveys."  In other words, Mr Reed thought that a "yes/no" answer in a survey allowed no room for error, regardless of the size of the sample (in the case to which I was referring it was 124). 

I pointed out to Mr Reed that sampling errors have nothing to do with a question being of a binary nature any more than they are to do with who designed the 'survey engine' on which a study has been conducted.  I also pointed out that far from reducing errors (supposedly by determining people with a specific viewpoint), filter questions are almost bound to increase them by splitting a sample into even smaller parts.  Having responded to every one of my previous emails, Mr Reed failed to do so this time, presumably because he finally realised his expertise in statistics leaves rather a lot to be desired.

Being involved in research means we don't always come up with the results people expect nor, frankly, want to see.  But as our job is to report independently and without bias we have to accept this goes with the territory.  It was, however, with a good deal of astonishment that I read an email from David Reed, the editor of 'Data Strategy' and an editorial contributor to 'Marketing Week' who responded to a press release we issued on the findings of our study into the 'consensus' behind man-made global warming.

Having accused me of being part of a right-wing plot to undermine the Copenhagen summit, Reed signed-off with "I hope you drown when the oceans rise" and "here's hoping your retinas get burned out when the ozone layer disappears".  I've always been a fan of 'Marketing Week' but I might have to revise my views if this is the quality of editorial contributors they choose to employ. 

I'm all for debating the findings of research but it would appear Mr Reed is not so prepared to have an open mind.  This is worrying for readers of 'Data Strategy' and 'Marketing Week' who rightly expect balanced editorial. 

In communication terms it is difficult to conclude anything other than that the public are not being given the full picture on man-made global warming.  From our study, reported on the 'Lab' page of this website, it is clear that the consensus agreement for man-made global warming among academics claimed by the Governmenht and international bodies is itself a lot of hot air, at least in the UK. 

Whether this is by fault or design we cannot say, but we suspect it will be the former.  In other words, while the Government and international bodies are unlikely to be deliberately misleading the public, they have clearly not done their homework and are making claims that simply don't stack-up in reality.

This tells us, yet again, that the public are not the mugs too many politicians take them to be.  People can see through the spin and, if anything react negatively against it, which is probably why recent polls have suggested that the majority of the UK population does not believe in man-made global warming.

Remember, too, that we have no axe to grind here.  We're neither for nor against the man-made global warming lobby.  We are, however, very much in favour of good communication, which means determining the facts and communicating them honestly and responsibly.  In today's highly cynical world, this is the only thing that people will believe; and rightly so.

Something else this tells us is how useful social media can be as a source of information on opinion, which can change very rapidly.  After all, just a few weeks ago the Climate Research Unit at the University of East Anglia was one of the world's leading centres in its field, with millions of pounds in funding.  Now, rightly or wrongly, its reputation has been severley damaged and will most likely take a very long time to recover.  Nothing makes the point more strongly about the need for organisations of any type, shape or size to ensure that they monitor, analyse and manage their reputations on a continuous basis.

Regardless of your views on man-made global warming, something that cannot be doubted is the BBC's clear bias towards its supporters.  This might be about to change.  Two days ago the 10 o'clock news had a remarkably balanced piece, which included an interview with a man-made global warming sceptic.

There are two possible reasons for this.  First, the hacking of the UAE's computer and shady emails from its Climate Research unit have cast serious doubts over the validity of the data used by man-made global warming advocates.  But second, this also coincided with publication of research into public opinion, which revealed that the majority of the UK population remain to be convinced about the link between global warming and man's contribution to it.  Has the penny finally dropped with the BBC that audiences are far more inclined to believe balanced reporting than obvious spin?

Let's hope so.

Metrica, the media evaluation company I founded, was acquired by Durrants at the end of last week.  It was an interesting few weeks leading up to the sale and wasn't the best time to be posting blogs, which is why things have been very quiet on here.

Anyway, everything's now done and dusted and I have to say that I'm delighted for both parties.  For Durrants it's a very good strategic move as it positions them as a world-leading evaluation company while for Metrica it brings the critical mass required to compete on a global basis.  It's also great for clients because it integrates the back-end delivery of monitoring with the front-end delivery of evaluation.

I am, of course, biased but if I were a competitor in the media evaluation world I'd be pretty worried by this development.  Metrica's brand is one of, if not the most respected in the evaluation industry and Durrants is probably the best known name in monitoring with a very significant customer base.  Combined, they create a powerhouse that should have the media evaluation industry looking over its shoulder.

So what does this deal mean for the industry?  I think it signals what's likely to become a significant trend towards consolidation as traditional monitoring companies seek to add value to their increasingly commoditised products.  As has happened in the PR agency business there are likely to be just a few major players within the next couple of years or so, one of which I am sure will be Durrants-Metrica.  Others are likely to be MBPrecis and TNS but what about Cision?  They seem to be moving more towards a Vocus-type service and are in danger of missing the boat, having made what look like fire-sale deals (not least to Durrants) and with a brand that seems to have lost direction. 

Beneath these major players we're likely to see an increasingly small number of independents squeezed from below by numerous press cutting companies positioning themselves as evaluation suppliers. 

These are interesting times for the evaluation industry and now that I'm no longer involved directly through Metrica it will be fascinating to see how everything develops!

Last week, Internet advertising spend overtook TV advertising spend for the first time in history (Telegraph 30/9).  So those who manage advertising budgets have shifted focus, recognising their best chance of turning ad viewings into sales is to get people clicking through online.  That's why Internet ads amounted to £1.75bn in the first six months of 2009 - 23.5% of all ad spending.

But Google Adwords and online adverts alone are still not enough to maximise the potential of customer attraction online.  Marketing managers need to boost their results not only in paid search, Adwords and sponsored links, but also in organic search.

So although the emphasis for ad budget ihas rightly shifted from more fragmented media, it would be foolish to become focused narrowly on paying for Internet exposure.  This can be generated even more effectively by other elements of the marketing communications mix, for example PR campaigns.  Look at the results achieved by Jaguar recently during the launch of the XJ - this was a PR-driven launch yet the amount of online buzz it generated was incredible.  In fact, Jaguar captured close to 10% of the total online buzz around a group of 37 car brands during the 3-week period around the launch.  The coverage was driven by a mix of a cool lauch event, celebrity endorsement, management commentary and online reviews.  A terrific online result - without the ad spend.