CorpComms Magazine

In the final fever-pitched days leading up to the US Presidential election, media coverage fixated on two diverging lines - one a triumphant red, the other a sorry shade of blue. Red, for Barack Obama and the Democrats, looked like an airliner soaring skywards, reflecting his rampant and ever-increasing lead in the opinion polls.

The blue line, for John McCain, was heading the other way, reflecting voters' growing disenchantment with the Republican cause. The widening chasm between the two presidential contenders suggested that the chance of Obama not winning the election was negligible. As one armchair pundit quipped, there was a greater chance of being struck on the head by a meteorite than for McCain to win the keys to the Oval Office.

Yet, in an agony of introspection, or perhaps an extreme case of hedging one's bets based on bitter past experience, the pollsters responsible for those red and blue lines left open the possibility that they might have got it horribly wrong. Respondents in market research surveys have a nasty habit of saying one thing when speaking to a researcher then doing the complete opposite when they enter the polling booth.

No amount of weighting and adjustments can totally compensate for the risk of voters lying. Yet with as much as 15 per cent of the $3 billion (£2 billion) spent on campaigning for the US presidential elections going on pollsters and market research analysts, the firms can only hope that their forecasts will stack up and that any flaws in their methodology will not be exposed to the cruel light of scrutiny. Like journalists and stories, market research firms are only as good as their last report.

Fortunately for Gallup, Newsweek and the other big names, the American voters did what they said they would and backed Obama. There was no recurrence of the dreaded ‘Bradley effect' seen in the California governor elections in 1982 when exit polls pointed to victory for African-American Tom Bradley. Embarrassingly, he lost. Pollsters surmised that respondents had lied about saying they would vote for Bradley because they did not want to admit to being racist.

NOT ALWAYS RIGHT

The pollsters got it wrong again in the 2008 New Hampshire primary election, when, a day before voting, the polls gave Obama a ten point lead over Hillary Clinton. Despite that, Clinton went on to win the state. And it is not just an American voting phenomenon. John Major's Conservative government was elected into power in 1992 even though all the polls pointed to a Labour victory.

In the London mayoral election in April, Ipsos Mori had Ken Livingstone down to win while YouGov had Boris Johnson on a ten point advantage. Clearly one of them had got it badly wrong.

Calling it right generates loads of free press coverage for the research company in the frame, which for an aggressive, growing firm like YouGov is worth its weight in gold. Prospective clients want to sign up with the winning team. Calling it wrong can be embarrassing if not financially ruinous, with companies from British Airways to Marks & Spencer spending millions of pounds on market research to help tailor advertising spend and brand positioning.

However it pans out, research firms are not above trying to turn a dire situation to their advantage, even if calling it wrong in the recent presidential elections would perhaps have been irredeemable.

In November 2007, ICM generated eye-catching newspaper headlines with claims that Britons were set to spend £21 billion in the run-up to Christmas. Giving retailers the information they needed to finesse their product line-up, ICM cited clothing, CDs and DVDs, books and health and beauty products as most likely to be on Christmas shopping lists. ICM canvassed about 1,500 consumers for their views.

With a week to go to Christmas, ICM came up with a rather different headline. Having benchmarked the figure of £21 billion, consumer shopping was now heading for a £1.4 billion shortfall. Regular monitoring suggested that the earlier respondents had been a little over-optimistic. Yet, in the finest tradition of market research ‘spin', ICM simply changed its stance, blaming the change on ‘dwindling consumer confidence fueled by general doom-mongering in the media'. So if it is not going your way, just shoot the messenger.

Market research is so much a part of business life that clients, it seems, will tolerate such blatant shifts in position. If it is true that any publicity is good publicity, market research firms certainly know how to cash in. From pedigree cats that turn up their noses at inferior brands of food to babies that will only tolerate the finest product range of toilet paper, there is no shortage of consumers to be canvassed, and of companies willing to retain their services.

None of it comes cheap, of course. Although firms are notoriously coy at disclosing what they charge for their services, partly for competitive reasons but also out of fear of scaring away prospective clients, published profits and revenue figures for the big players dispel any doubts about the level of fees generated.

The global market research industry is worth an estimated $25 billion (£17 billion) annually, with Nielsen generating revenues last year of $4.7 billion and Taylor Nelson Sofres generating $2.1 billion. Even niche firms such as YouGov, known for its BrandIndex which tracks public perception of brands such as British Airways based on ‘buzz' and other measures, made a pre-tax profit of £5.6m last year. Interim figures for YouGov showed pre-tax profits up 30 per cent at £3m on turnover 208 per cent higher (due to acquisitions) at £18.8m And YouGov is at the cheaper end of the fees scale. Its use of online consumer panels means fees are as much as 75 per cent lower than traditional telephone/clipboard-style polling.

OMNIBUS APPROACH

The industry is resilient, too; even in an economic slump, big companies need to feed their appetite for data and extrapolations. And along with traditional telephone canvassing, other, cheaper methods of conducting research are making market research more accessible to clients, deepening its catchment. Increasingly popular is so-called omnibus research in which pollsters load up each survey with several different topics. Clients buy as few or as many questions on the list as they like, a process likened to buying seats on a bus.

Any one omnibus survey can cover a wealth of topics from favourite holiday destinations to political views to awareness of particular advertisements. In a typical month this year, ICM placed nearly 900 questions on behalf of 125 clients among different omnibus surveys. Here, too, online polling is growing in popularity, with a presumption that respondents answer more truthfully from the anonymity of a computer than in telephone interviews or face-to-face.

But does market research yield anything more useful than calculated guesses? Cynics might claim that it is a self-sustaining industry that feeds on advertising men as well as company managers seeking statistics to endorse their existence. Communications professionals, too, are a soft touch when it comes to commissioning expensive surveys. Who wouldn't jump at figures endorsing their client or brand? And if the research highlights some unpleasant attitudes, it provides a base from which to build improvements.

While market research often looks like a case of firms dreaming up a theme that they can sell to clients - and then going out to canvass for views - the results, either way, can be extremely influential. A YouGov survey in October, at the tail-end of the banking crisis, saw voters rate National Savings & Investments (NS&I) as the safest place for their savings, with a rating of 7.6 out of 10. Conversely, and disastrously for the PR men concerned, Halifax/Bank of Scotland came out second last (ahead of Bradford & Bingley) on 4.8 out of 10. The ‘free' advertising for NS&I was eye-watering while the impact in forcing HBOS into damage-limitation mode was equally powerful. Here was a case of market research landing its punches.

As with Obama and his widening lead over McCain, the polls become self-fulfilling as everyone clamours to jump on the winning team.

Yet it is all-too easy to forget that surveys which give the appearance of speaking for the nation generally rely on as few as 1,000 respondents. YouGov's financial services survey canvassed 1,416 savers over just three days; on the strength of that pitifully small window, NS&I was handed millions of pounds in free advertising. Although firms insist that sophisticated balancing and weighting means that the small scale accurately reflects the opinions of millions of people, they get it wrong with enough regularity to make you wonder.

CORRECT RESULTS

Market research firms claim that a poll of 1,000 people will be accurate 95 per cent of the time within a margin of plus or minus three per cent. Yet in an election race, with the Democrats on 32 per cent and the Republicans on 38 per cent, this margin for error means that both parties could be on 35 per cent. Flawed method¬ology is an enduring problem. You have to suspect that the pollsters hope to bluff their way through, basking in the glory when they get it right.

Political pollsters face additional problems such as the phenomenon of the ‘Shy Tories' in which survey respondents are too embarrassed to admit that they are going to vote Conservative in the General Election.

Technology and changing demographics are further pressures facing market research firms. Canvassers use random digit dialling of landlines to build their samples, cutting out younger, often less-affluent, mobile phone users. So their findings may not be as representative as they would claim.

In such a fragmented industry, with dozens of firms touting for business, reputation is everything. Client companies want to be able to state with authority that their product has been endorsed by NOP or Ipsos Mori. If a particular firm consistently gets it wrong, clients will defect to firms with a better track record. Such are the brutal realities of market research.

Pollsters live in fear of ‘Michael Fish' moments such as during the 2004 US Presidential election when Sir Bob Worcester, the Mori grandee, chose the heat of election night to proclaim, before the results were in, that victory had gone to John Kerry. It was rather embarrassing for him when George W Bush duly won the election.

Market research firms sometimes end up as political footballs. In October, Opinion Leader Research was accused by Greenpeace of ‘fixing' the public consultation on nuclear energy in Britain to give its client, the government, the answers it needed to press for more nuclear power stations.

Acting on a complaint from Greenpeace, the Market Research Standards Board found that the way the questions were phrased along with ‘misleading' information about wind power and other energy sources steered respondents towards a particular answer.

Opinion Leader Research could laugh off such criticism as all part of the game, but when it comes to forecasting election results, firms are on much thinner ice. In 2004, YouGov forecast that Kerry would beat Bush by a three per cent margin. It pinned the blame on a ‘sampling error'. Conversely, favourable YouGov research published in the Daily Telegraph at the start of the 2004 European Parliament campaign gave a huge boost in support for the UK Independence Party, generating more than a fortnight of positive press coverage.

Accurately forecasting an election result generates fabulous free publicity for the firm responsible, less so this November when Obama's win looked certain so far in advance. Away from the polling booths, market research firms' staples of brand tracking and customer satisfaction surveys can be just as powerful. The market research industry is about gathering facts and offering strategic business insights to help clients gain a competitive edge.

Increasingly, clients won't make investments in advertising or online marketing without the statistics and insights to back up the decision. While US Presidential elections might grab all the attention, it is this unceasing work that keeps market research firms in business. With such appetite for figures, the industry can only gain in strength, even with the occasional gaffes to blacken firms' names. In the fast-paced world of business such blots are quickly forgotten, or at least so the firms hope. Clients vote with their wallets, after all. 

share me: del.icio.us | digg | reddit | Tweet