Elections have always been a numbers game, but in the last decade we have seen a seismic shift. It is now a data game, and there is no turning back. This revolution started in earnest with Obama’s 2012 re-election campaign and has since barrelled through US and global politics, leaving disrupted strategies and confused pollsters in its wake.
Politicians want to know what messages resonate with their audience. Such topics were traditionally derived in person through focus sessions, but those outcomes can be distorted by small sample sizes and the artificial setting. In 2012, Obama’s team proved that you could use data science techniques to identify granular voter types and spot opportunities to win otherwise hidden votes. The Romney campaign struggled to keep up. They could see that the Democrats were advertising at irregular times and in unusual ways, but they could not reverse-engineer the model to work out why: “We could tell that there was something in the algorithms that was telling them what [adverts] to run”, said Brent McGoldrick1, a then member of Romney’s ‘Strategy, Polling and Media Analytics’ team.
The Obama campaign was not cheap but by rethinking their approach to advertising they would have been able to take advantage of off-peak rates and know that they were still making a sound investment. Jared Kushner, discussing Trump’s 2016 campaign, indicates that they also were looking for good ROI: “We ran the campaign like a business”, said Kushner. “You have ten different ways that you could reach the potential consumer. You ask, ‘How can I get my message to that consumer for the least amount of cost?”2
The question of how to find new clients and approach them successfully determines business development in any corporation. At ECOVIS Wingrave Yeats we help our clients to get to know their client base through segmentation analysis and then employ machine-learning techniques to help focus their advertising and drive sales. For a client in the fashion industry we identified how sale offers resonated with individual customers and developed a machine-learning solution which they calculated would allow them to save 17% of their catalogue mailout costs with minimal risk. Machine learning can determine a wider range of customer profiles than traditional methods could ever hope to deliver. They allow us to optimise campaigns on the fly and predict successful marketing choices ahead of time.
For organisations where referrals or people-networks are an important source of new business, we provide a client network analysis tool which maps a client’s portfolio and highlights business development opportunities in their existing network. When linked to a suitable CRM system this product can be used to help tailor the method of introduction to each potential client individually. It is a flexible tool which can also be layered with external datasets to give our clients insight into their industries and a stronger platform on which to do business.
Professionals can now use data science to identify opportunities for growth before they are evident to the human observer. Many companies are on the cusp of technical innovation but are yet to take the leap to a progressive data-driven strategy. We see that data science has released potential for growth and innovation across sectors and by working with our clients to realise their ambition we help them to offer a better service and get ahead of the competition.
Alison Champernowne, Senior Data Analyst, ECOVIS Wingrave Yeats UK, London, UK
1) Sasha Issenberg. 2012. MIT Technology Review. [Online]. [1 February 2019]. Available from: https://www.technologyreview.com/s/509026/how-obamas-team-used-big-data-to-rally-voters/
2) Steven Bertoni. 2017. Forbes. [Online]. [1 February 2019]. Available from: https://www.forbes.com/sites/stevenbertoni/2017/05/26/jared-kushner-in-his-own-words-on-the-trump-data-operation-the-fbi-is-reportedly-probing/