Marketing Fast & Slow

Marketing Fast & Slow

Gerry HopkinsonCEO, Selbey Labs
Marketing Fast & Slow

Some of us take pride in “getting shit done” and others celebrate the joys of deep thought and reflection. In the world of Marketing there is room for both, but how do you know when to take a step back and when to step on the accelerator? The perceived wisdom is we plan in the planning cycle and then execute our campaigns and hope for the best. This works, up to a point but it misses a greater opportunity. What if we divided to conquer?

There is a really interesting algorithmic theory that looks at the trade-off between “exploration” and “exploitation”. Exploration is an attempt to acquire new knowledge and “exploitation” is an attempt to optimise decision making based on existing knowledge.

In the fields of probability and machine learning, there is a well-known problem called the “Multi-armed Bandit” problem which trains machines.

In this problem, there are fixed resources which must be allocated between competing choices to maximise gain. The twist is that the choices are not always informed with sufficient information, but future choices may be improved by rolling the dice and learning i.e. chancing your hand. It’s a classic re-enforcement learning problem and is deployed widely.

Sound familiar? For me, that’s a perfect parallel with an average marketing team’s challenge on any given day. There is a finite amount of resources, they have to be allocated in such a way as to deliver growth and they have to swing between decisions that are built on what’s gone before and decisions that require new understanding to succeed.

Crucially, the only way you can really succeed is to occasionally test and learn by trying things you’ve not done before to see what happens.

Of course, if that’s all you did, you’d lose as the odds would be stacked against you. But equally, if you just trod the well-worn path, you’ll like fail too as you will miss out on the chance to innovate.

So the solution seems to be to balance the “exploit” behaviour of doing what you know works and the “explore” behaviour of seeing what else might work.

Rather than doing this like a gambler, and just rolling the dice, we think there’s an easier way.

Coca-Cola, Google and other smart brands use a 70-20-10 rule for getting the balance right between exploring and exploiting.

For example, they would invest 70 percent of their marketing budget on tried and trusted channels, messaging and tactics, 20% on variations on what has worked e.g. they might know that emotionally weighted comms work, but they try a different emotional tone and 10% is reserved for pure innovation, trying things they’ve never done before and testing and learning as they go.

What makes this even easier is if you build a dedicated team of explorers and evaluators who constantly look to tomorrow, seeking new ideas, information and knowledge and keep a steady eye on the present, evaluating the impact of the decisions you are making today.

The fancy name for these teams is Foresight & Innovation, and they are the secret sauce that help brands maximise the gains on tried and trusted strategies and find the next wave of ideas that will keep them one step ahead of the competition.

We set up Selbey Labs to offer clients precisely this service and we know it works.

If you’d like to learn more about how to set up your own team or work with us to explore how we could help you, please get in touch.

We’d love to hear from you.

Gerry Hopkinson
Gerry Hopkinson
CEO, Selbey Labs

Following a successful career in communications, including co-founding award-winning agency Unity (now part of the Selbey Anderson Group), Gerry set up Selbey Labs in April 2022. Through his work at Unity over the past two decades he has developed a perspective on the role of brands in culture, the importance of observation and benefits of testing and learning.

Latest Insights