ConservativeHome's General Election Review (executive summary here/ contents page here), concluded with ten recommendations. One of them was this recommendation for a full-time party pollster.
A good pollster-politician relationship is not about the pollster telling the politician what to believe but the pollster will perform three main roles. He/she will give the politician an early warning of the issues worrying voters - particularly the voters in swing seats. Second, they will advise on which of the policies already chosen by the government/ political party should be front of shop and which should be back of shop. Thirdly the pollster will advise on how best to communicate policies, especially on language and tone.
It seems clear that the Tory leadership did not use polling properly in opposition. Three very good sources said that the 'Big Society' agenda was never properly tested or focus grouped. That is a political crime. It would be like Alex Ferguson buying a footballer even though he'd never seen him play. Although an impressive governing philosophy the Big Society was too nebulous for an election campaign. Key figures in the leader's circle acted as "creative directors". A typical creative director is so convinced of their idea that they *know* the public will come round to accepting their great idea, if only it is presented to them again and again. Only a powerful, full-time pollster - enjoying the party leader's confidence - can stand up to such a belief and ensure mistakes like the Big Society are avoided in future.
In government the party cannot continue with what has been described as a pick'n'mix attitude to researcxh - choosing findings that reinforce instinctive views and ignoring others. Team Cameron, so careful to study the Blair project, need their own Philip Gould.
