You don’t always need data to hit your target

Having been warned about taking care in how it targets election material, Labour is facing a media storm over leaflets focusing on cancer care.

Some of these were received by individuals currently getting treatment for breast cancer, leading to suspicions that the campaign had been deliberately targeted.

At first sight, this looks like it would be hard to do. After all, the Data Protection Act is clear about the use of sensitive data, such as medical history. It would be impossible to argue that this data should be used to target information about how the opposition might cut treatments, even if it were available.

But you do not have to have the direct variable on health to find yourself targeting a cancer patient. It may be that the Labour supporter database has been built including affinity data, such as charity donations. That would look like an obvious enhancement to carry out, especially if the party is looking to solicit funds as well as votes.

Some health charities do make their own support lists available, although usually as swaps. More likely any such variable could have been derived from lifestyle surveys asking people if they give money and to what type of cause. Charity donors tend to be older and female – seen as the key group in the electorate during this election – so it would not be a surprise if the data had been enhanced in this way.

There does not even have to be direct data on health in order to target sufferers. Last year’s Grand Prix winner at the Data Strategy Awawards was British Lung Foundation. Working with Experian, it combined lifestyle data and hospital admissions rates to identify which areas had the highest probability of people suffering from an undiagnosed pulmonary condition. It was clever use of data and highly effective.

Labour is also working with Experian on its targeting data, so may well have carried out profiling in order to align campaign messages to areas where they would have most salience. That is what data-driven marketing is all about, even when being done for national issues, like who should form the next government.

But it could be that this particular distribution did not use any targeting at all and just happened to have more relevance to its recipients than was expected. The epidemiology of breast cancer indicates that 1 in 12 women are likely to contract it before the age of 95.

If the catchment area for Labour’s leaflets had an older, more female profile, then that probability could be higher – perhaps 1 in 9. So completely unintentionally, the party may have delivered messages right to the people most affected, with no clever data work (or breach of the DPA). Whether it was right to campaign on this issue and raise fears of lower treatment standards, however, is another matter.