Ben Page, chief executive of the UKs leading market research and polling firm, talks to Richard Purcell ahead of the Life Conference about predicting elections, going beyond big data, and observing human behaviour
I catch up with Ben Page, CEO of Ipsos MORI, ahead of his speech at the Institute and Faculty of Actuaries Life Conference in Dublin. Although a graduate of Oxford who counts Boris Johnson and David Cameron among his fellow classmates, and one of the most influential people in the public sector, everything about Page eschews stuffiness and convention, from his colourful shirts to his energetic and straight-talking style.
We begin with Page revealing his unconventional route into market research. "It was when I was running a nightclub in Oxford that I thought about moving to London. I found a job advertised in Time Out magazine that involved interviewing people over the telephone. After a while I thought I had better get a proper job, and started as a graduate at MORI in 1987." Page hasn't looked back since. He was promoted, then in 2000 helped to lead a management buyout, before selling to Ipsos five years later, and becoming chief executive of Ipsos MORI in 2009.
Most people probably associate the pollster with general elections, so we touch on why canvassers got the results of the last UK election so wrong. Page explains: "It was not a disaster, we were the least inaccurate pollster." He adds: "The problem we have is people saying they are going to vote a certain way tomorrow, and then three in 100 either vote a different way or ultimately do not vote at all. That's why the exit polls were so accurate by comparison, because - by standing outside 141 individual polling stations - we know they've at least voted."
On learning lessons for the future, Page believes they need tougher filters to adjust for this "over claiming" of likelihood of voting in polls. But he also says you have to understand the limitations of research. "Most people using market research know 1,000 fast-turnaround interviews won't give you pinpoint accuracy, yet the press put too much reliance on these polls." He stresses that the way you conduct research and how much you spend is also important, adding: "I think that's why the Conservative Party seemed to have a better handle on the way the election was going and where to focus resources, because they put more investment into polling in individual marginal constituencies, data that simply wasn't available to the mainstream media."
The failure of the polls has also led to some calling for a ban on their publication. But Page is quick to point out: "A media blackout will mean banks and other organisations will commission their own polls and use this to their advantage. While this would actually mean more business for us, it is not really in the public interest."
While Ipsos MORI is well known for its work in the public sector, Page explains that the vast majority of their work is for the private sector. "We do market research on anything related to human behaviour and activity, from surveying sex workers at truck stops in Africa, to understanding bullying in abattoirs. It keeps it interesting!"
Power of observation
Page describes how market research is changing. "We know people don't always do what they say they will, because they're sometimes not able to rationalise why they prefer one product over another, for example. They also edit experiences massively in their own mind. So increasingly we are using ethnography or passive observation to see what people actually do. This means we are starting to use techniques like facial coding and psychological response to help us understand behaviour."
When talking to Page about the company's use of technology to observe people, such as cameras on glasses or wearable devices, there seem to be strong parallels with insurers collecting vast amounts of data.
He explains: "We have so much data now, and we are finding some interesting correlations such as a link between divorce rates and cheese eating. But the challenge is to spot patterns in data that we haven't asked the computer to find. This is the gold rush for big data."
He points out that we need to go beyond big data though: "Big data tells us what people are doing or how many are doing it, but doesn't tell us why. Polling can tell us the answer and helps us narrate the data." For Page, this underlines the need to layer these new techniques with traditional research, rather than replacing it entirely. He adds: "Knocking on doors is still effective, and I'm always amazed when we still get a 15% response rate from postal surveys. Much higher than anything online."
Learning from each other
The importance of analysing and narrating data in the market research world draws similarities with the work of actuaries. When I ask Page about his perception of actuaries, he responds "dry and numerical" and then pauses before concluding, "a bit like market researchers". While he associates actuaries with predicting life spans and having the advantage of more certainty in the demographics in the short term, he sees some similarities with the work of market researchers. "There is a spread of outcomes with actuarial work, and getting people to accept there is a range is important in our field too."
On data, another area where there is common ground, he believes privacy is a big unanswered question that we need to think about. "Some people still think you can say anything you like on the internet and that there are no consequences. What's more, people are worried about their data but only 25% have ever changed their browser settings. This "cognitive polyphasia" - holding conflicting ideas - is a challenge." He believes that we all have a fine balancing act to tread in using data to help people, but without intruding in their lives and being perceived as creepy.
Actuaries can also be consumers of market research too, using it to help us inform new product design. "It can be difficult to do research on things that don't exist yet. Even I couldn't see the benefit of iPads when they were first launched. I now have three." Page believes the key is to identify if there is a need for a product by getting under the skin of the problem, rather than asking 'Would you buy this?'. He also points out the benefits: "Research can save you wasting a lot of money in developing new products, and also improve your return on investment on marketing." When I ask about examples of research that has not been used effectively, he points to Pepsi Blue. "People said they liked it during testing, but when they re-branded it failed to take off". He goes on to say that you can't use market research in isolation. He believes "the user experience is key", and that in cases like Pepsi Blue, the experience didn't live up to what they said it would.
Future of market research
Besides the importance of observation alongside traditional polling techniques, Page recognises that the industry is becoming more diverse. While technology is making surveys easier and reducing the barriers to entry, he sees the importance of innovation and multi-disciplinary teams to help solve problems. "We already employ a broad range of people from economists to statisticians to documentary film-makers."
When I volunteer the idea of employing and using the skills of actuaries, he replies: "That's an interesting idea." I'm not sure he is 100% sold on the concept, but I did come away feeling that our skills in analysing and interpreting complex data sets means that actuaries could well be at the centre of predicting elections in the future.
We begin with Page revealing his unconventional route into market research. "It was when I was running a nightclub in Oxford that I thought about moving to London. I found a job advertised in Time Out magazine that involved interviewing people over the telephone. After a while I thought I had better get a proper job, and started as a graduate at MORI in 1987." Page hasn't looked back since. He was promoted, then in 2000 helped to lead a management buyout, before selling to Ipsos five years later, and becoming chief executive of Ipsos MORI in 2009.
Most people probably associate the pollster with general elections, so we touch on why canvassers got the results of the last UK election so wrong. Page explains: "It was not a disaster, we were the least inaccurate pollster." He adds: "The problem we have is people saying they are going to vote a certain way tomorrow, and then three in 100 either vote a different way or ultimately do not vote at all. That's why the exit polls were so accurate by comparison, because - by standing outside 141 individual polling stations - we know they've at least voted."
On learning lessons for the future, Page believes they need tougher filters to adjust for this "over claiming" of likelihood of voting in polls. But he also says you have to understand the limitations of research. "Most people using market research know 1,000 fast-turnaround interviews won't give you pinpoint accuracy, yet the press put too much reliance on these polls." He stresses that the way you conduct research and how much you spend is also important, adding: "I think that's why the Conservative Party seemed to have a better handle on the way the election was going and where to focus resources, because they put more investment into polling in individual marginal constituencies, data that simply wasn't available to the mainstream media."
The failure of the polls has also led to some calling for a ban on their publication. But Page is quick to point out: "A media blackout will mean banks and other organisations will commission their own polls and use this to their advantage. While this would actually mean more business for us, it is not really in the public interest."
While Ipsos MORI is well known for its work in the public sector, Page explains that the vast majority of their work is for the private sector. "We do market research on anything related to human behaviour and activity, from surveying sex workers at truck stops in Africa, to understanding bullying in abattoirs. It keeps it interesting!"
Power of observation
Page describes how market research is changing. "We know people don't always do what they say they will, because they're sometimes not able to rationalise why they prefer one product over another, for example. They also edit experiences massively in their own mind. So increasingly we are using ethnography or passive observation to see what people actually do. This means we are starting to use techniques like facial coding and psychological response to help us understand behaviour."
When talking to Page about the company's use of technology to observe people, such as cameras on glasses or wearable devices, there seem to be strong parallels with insurers collecting vast amounts of data.
He explains: "We have so much data now, and we are finding some interesting correlations such as a link between divorce rates and cheese eating. But the challenge is to spot patterns in data that we haven't asked the computer to find. This is the gold rush for big data."
He points out that we need to go beyond big data though: "Big data tells us what people are doing or how many are doing it, but doesn't tell us why. Polling can tell us the answer and helps us narrate the data." For Page, this underlines the need to layer these new techniques with traditional research, rather than replacing it entirely. He adds: "Knocking on doors is still effective, and I'm always amazed when we still get a 15% response rate from postal surveys. Much higher than anything online."
Learning from each other
The importance of analysing and narrating data in the market research world draws similarities with the work of actuaries. When I ask Page about his perception of actuaries, he responds "dry and numerical" and then pauses before concluding, "a bit like market researchers". While he associates actuaries with predicting life spans and having the advantage of more certainty in the demographics in the short term, he sees some similarities with the work of market researchers. "There is a spread of outcomes with actuarial work, and getting people to accept there is a range is important in our field too."
On data, another area where there is common ground, he believes privacy is a big unanswered question that we need to think about. "Some people still think you can say anything you like on the internet and that there are no consequences. What's more, people are worried about their data but only 25% have ever changed their browser settings. This "cognitive polyphasia" - holding conflicting ideas - is a challenge." He believes that we all have a fine balancing act to tread in using data to help people, but without intruding in their lives and being perceived as creepy.
Actuaries can also be consumers of market research too, using it to help us inform new product design. "It can be difficult to do research on things that don't exist yet. Even I couldn't see the benefit of iPads when they were first launched. I now have three." Page believes the key is to identify if there is a need for a product by getting under the skin of the problem, rather than asking 'Would you buy this?'. He also points out the benefits: "Research can save you wasting a lot of money in developing new products, and also improve your return on investment on marketing." When I ask about examples of research that has not been used effectively, he points to Pepsi Blue. "People said they liked it during testing, but when they re-branded it failed to take off". He goes on to say that you can't use market research in isolation. He believes "the user experience is key", and that in cases like Pepsi Blue, the experience didn't live up to what they said it would.
Future of market research
Besides the importance of observation alongside traditional polling techniques, Page recognises that the industry is becoming more diverse. While technology is making surveys easier and reducing the barriers to entry, he sees the importance of innovation and multi-disciplinary teams to help solve problems. "We already employ a broad range of people from economists to statisticians to documentary film-makers."
When I volunteer the idea of employing and using the skills of actuaries, he replies: "That's an interesting idea." I'm not sure he is 100% sold on the concept, but I did come away feeling that our skills in analysing and interpreting complex data sets means that actuaries could well be at the centre of predicting elections in the future.
Ben Page will be speaking at the Life Conference in Dublin, 18-20 November, 2015
Filed in
Also filed in