As marketers understand well, driving the path to purchase on any product is far more complex and nuanced than a simple case of ‘find person, show ad and the rest will follow’.
The motivations that drive consumers are intricately layered. Take my recent acquisition of a copy of Pride and Prejudice. I’d recently watched the film and had been considering rereading the novel. That day, I had read an article that told me the most productive people read a book a week – and then I saw an offer online for a reduced price on a beautiful copy. Something I’d normally never notice suddenly jumped out at me as a ‘must-have’.
This context is a lot for a marketer to account for. But new developments in machine learning and vast datasets have given marketers an advantage that once they could only have dreamed about. The insights they can create to understand consumer behaviour and the actions they can take to tap into those mean that the hard work of data crunching and insight can be automated, leaving time to get creative around how to make make best use of these capabilities.
What does this look like in practice? One great example springs to mind here.
The Royal Mint – the organisation responsible for the UK’s coin production and the retailer of coin collections, bullion and gifts – worked with their partner agency Manning Gottlieb on a search strategy that was nothing short of exceptional.
First of all, some background. The demand for gold is highly volatile, and The Royal Mint wanted to develop an alternative to traditional paid search strategies such as seasonal trend optimisation. Marketing a product with such fluctuations in demand presents an interesting challenge – and one I find interesting as a case that is applicable across product categories.
Manning Gottlieb’s task was this: to capitalise on gold demand trends with no incremental budget while improving sales by 15% and reducing cost per acquisition by 40%. The team had already observed that around the votes for Brexit and the US presidential election, huge decreases in the FTSE 100 price index coincided with surges in searches around investing in gold.
Of course, seismic events such as this are infrequent – and so the idea was to look at how this trend could be tapped into in a more micro, but more regular, fashion. They hypothesised that significant events cause instability in the FTSE 100, and that performance drops on the FTSE 100 coincided with an upswing in searches for gold.
Their strategy was to study a three-month date range and note any events that were significant in steering the mood of the economy, including, for example, the announcement of the UK budget.
When they uncovered clear correlations, they tested boosting bids and budgets around key dates in the UK political calendar – but found that market reactions were too quick to be tapped into through automated budget adjustments alone. They then integrated the search strategy with moment-marketing technology provider TVTY, enabling them to link The Royal Mint’s Google Ads accounts to FTSE 100 performance. Effectively, whenever the FTSE 100 dropped by more than 1%, the account’s ad groups would boost bids by up to 50%, enabling real-time reactions to moments of economic uncertainty.
What they found was that during the moments TVTY triggered, they could see an increase in the click through rate of 8%, and a huge conversion rate increase of 26%.
After just a month using this strategy, they beat the initial sales target by 85%, with a cost per acquisition 37% cheaper than their goal. Being something of a data nerd myself, I’d love to see how this could be taken further by combining this with first party data so that they could identify which customers engaged with them and which didn’t – something that we offer though our Audience tool.
What’s clear for other marketers facing turbulence in product demand is that through smart use of data, those extremes in demand can be used to manage more regular, micro-boosts. The Royal Mint has tapped into third party data in an innovative and exciting way. Automation and machine learning have given us the chance to make our insights sharper and more effective than ever before – freeing up our time to think creatively around how we can apply those insights in new and exciting ways. Only through innovation and experimentation can we execute really smart campaigns based on the genuine understanding of consumer behaviour that machine learning affords.
Picture credit: "New One Pound Coins", by M. Zhu, used under CC BY-NC-ND 2.0 / Modified from original
Interested in hearing leading global brands discuss subjects like this in person?
Find out more about Digital Marketing World Forum (#DMWF) Europe, London, North America, and Singapore.
