When will politicians and pollsters tap marketing data?

Scott-LogieBetween now and July 4, we are going to get a lot of data and stats thrown at us. It has already started with the “£2,000 increase in household tax if Labour win” claim from Rishi Sunak at the recent debate. 

That one claim (lie?) has itself generated a huge amount of data – he made the claim nine times in the debate, but there are some holes in this claim, e.g. the £2,000 is based on some maths around the increase in tax needed to meet spending plans; it’s across four years, not per year, and so on.

From a data geek’s perspective, elections are rich fodder. We can’t get enough of the numbers, whether it is polls and predictions, swings and shifts, or claims and counter claims. And don’t get started on election day itself; once voting closes it is basically one long night of “numbers porn” with seats being counted, results coming in and forecasts shifting on a real-time basis. Phwoar!

In the US, there are star observers and prediction experts like Nate Silver (a bit of a hero of mine) and Allan Litchman. In the early 1980s, Lichtman developed the ‘Keys to the White House’ system – a checklist of 13 statements where if five or fewer are false, the candidate for the incumbent party will win; but if six or more are false, the challenger will.

The keys are based on a range of factors, including the state of the economy, the previous mid-term election results, social unrest, charisma and more. Basic, but it has been successful in every election for the past 30 years (he backs Biden in ’24, incidentally). Nate is much more probability driven: having built models in every state, he has predicted every state correctly for the past few years.

And then in the UK, there was Cambridge Analytica, where The New York Times, working with The Guardian, obtained a cache of documents which proved that the firm, where the former Trump aide Stephen K Bannon was a board member, used data improperly obtained from Facebook to build voter profiles. The news put Cambridge Analytica out of business and thrust Facebook into its biggest crisis ever.

As a data-driven marketer, and not just a data geek, I have even more interest in an election. For me, there are a few areas where data could be better used. For example, and picking up on Nate’s approach, most polls in the UK tend towards “nationally representative” sampling.  For an overview, that is fine, but we don’t vote at a national level. And, as marketing data specialists, we also know that there are volumes of data already available at local levels, including parliamentary constituency level. That includes datasets released by the Government themselves, such as petition data. The level of knowledge that allows forecasters to make more granular predictions, as well as the parties to ensure they talk locally about local issues, is immense.

Previously, we have used that data to predict every constituency and who is going to win. We predicted an SNP landslide in Scotland in 2015 and no one would believe us. The UK is not one big election, it is 650 or so small elections and looking at who is going to win each one is more valuable. And marketing data is available and useful to help with this.

And then there is how parties go about persuading us who to vote for. As data-based marketers, we know all about using data to target better: it is what we do. A lot of time and effort has gone into social media targeting for younger voters, but can any party claim to have a true multi-channel approach to their campaign?  Do they even approach it like a campaign?

Given the wealth of data that is available, and that can be contacted under GDPR, the actual contact that happens ends up being very perfunctory and doesn’t feel very personal. Surely by 2024, we should have worked out how to engage voters using personalisation? And we haven’t even touched on segmentation, or predictive modelling, or retention analysis. The possibilities are endless.

Overall, elections bring out the data geek in those of us who have an interest, but also infuriate and frustrate us because the data that is available is not being used that well.

Scott Logie is chief commercial officer at data solutions experts Sagacity

Print Friendly