Opinion

Research suggests that research may not work

Six months after the general election, if there's anything to be learnt, it's that research might not work, says Greg Delaney, co- founder of Watchable.

Research suggests that research may not work
Research suggests that research may not work
couple of people lost their jobs, but since it was only Ed Miliband and Ed Balls, they don’t seem that bothered

Apart from a bit of handwringing, it is pretty much business as usual for the research companies that got the result of the General Election so spectacularly wrong in May 2015. OK, a couple of people lost their jobs, but since it was only Ed Miliband and Ed Balls, they don’t seem that bothered.

To be fair, the pollsters did set up an enquiry into what happened (the final report is due in January) no doubt because their failings were very much in the public eye. Professor Sturgis of Southampton University, who is chairing the report, gave some early indications last month as to why they might have got their surveys so wrong: "Not everyone puts themselves forward to take part, and those who do will not be representative". In other words, they only asked people who don't mind doing surveys. I don’t know about you, but I can’t think of many people like that. But maybe I'm not representative either.

Research isn't mobile-first

It also turns out that many of the researchers only called people up on their landlines: "Many households are increasingly reluctant to pick up their landline, and a growing proportion of young people in particular rely exclusively on mobile phones", says the good Professor, who may also believe the Pope is statistically likely to be Catholic.

Many voters claimed they were Labour, but in reality they voted Tory. Again, this shouldn't surprise anyone - unless you're a research company apparently

Another widely quoted finding is that people didn’t actually vote in the way they said they would - they didn’t tell the researchers the truth. Many voters claimed they were Labour, but in reality they voted Tory. Again, this shouldn't surprise anyone - unless you're a research company apparently.

I’m left wondering how often this inaccuracy (or maybe the word is incompetence?) has happened in all the masses of predictive research undertaken in the marketing world? After all, in many cases it’s the same companies who are conducting the research looking at market initiatives, advertising and new product development, to give just a few examples.

£3 billion industry

For the past couple of decades, research companies have made fortunes claiming that they are the key to predicting success or failure. The research market is worth over £3 billion a year in the UK, so clearly plenty of people believe it can help them to make decisions.

Brands and political parties should stop wasting time and money trying to work out what people will like and what their intentions are

But the early conclusions from the General Election suggest that they are not always that good at it. What we now know is that they're often only researching people who like doing surveys; they like to call up on landlines when many people just use mobiles; and that even when they do ask people what they are going to do, those people won’t necessarily tell them the truth.

How many other research failures and consequent job losses have there been over the years which weren't subject to any public scrutiny? For example, it’s common knowledge that most new products are researched extensively and yet most new products fail.

Embrace the digital age

There is one solution to this problem of predicting the future: it is that brands and political parties should stop wasting time and money trying to work out what people will like and what their intentions are. Instead, they should embrace the digital age, put their messages and products out there and simply watch to see how people react. If their audience believes it, likes it, or wants to buy or vote for it, this will soon become apparent.  

Some smart marketing companies are already doing this: producing a stream of content (news, comment, entertainment, information) which they distribute and then observe and, if necessary, change or adapt. If it looks like it’s working, they put more money behind it. If it isn’t working, they abandon it. That is real time, real world research.

In other words - and this may be difficult for people who are used to researching everything, including which colour socks they should put on in the morning -  you need to live with a constant risk of judgement and possible failure in order to be able to succeed. The future is as unpredictable as it has ever been. Just ask Ed Miliband.