Don’t hate, iterate: making the most of agile marketing

One of the key principles of Agile software development is the idea of ‘continuous delivery’. It’s a contrast to the old way of struggling for months or years to deliver a complete application that’s out of date by launch. It should come as no surprise to us that these methods significantly crossover with the aims of the marketer.

Against the backdrop of a digital world, marketing has become a digital profession – in fact, digital dynamics now pervade almost every corner of marketing. The proliferation of new technologies and platforms over the past decade has placed marketers in a world as disruptive and unpredictable as the developer’s.

Digital channels and touc hpoints influence people’s buying decisions for all kinds of products and services, at every stage of the customer life cycle. So how do we adapt what the software industry has been doing for years to suit our own purpose?

Developers iteratively test and modify code instead of performing a separate QA process on completion. Marketers can mirror this process; iteratively test campaigns on the fly and incorporate the learning as you go, rather than measuring success at the end and using the results to do something different in the next cycle. This is the state of ‘constant beta’ we aim for.

Achieving ‘constant beta’

A Google search returns countless definitions of Agile marketing, often confusing it with reactive social strategy. For us, Agile marketing is a lean and simple process-driven manner of working that stimulates transparent communication, collaboration and the sharing of intelligence and experience. In practice, it does this by utilising empowered and multidisciplinary teams with a strong sense of individual accountability, regular short meetings, short delivery cycles, continuous measuring and a ‘fail fast and learn’ philosophy. Its ultimate goals are speedy adaptation and proactive innovation in response to a market that is as fast as it is unpredictable.

The mindset must uncompromisingly support a state of flux; it’s now simply complacent to accept that a project is ever done. ‘Tried and tested’ is no longer relevant for more than a few weeks, so trying and testing must go on repeatedly. This can seem a daunting and insurmountable shift in approach to many teams – the trick is to chop up big plans or projects into little chunks. It allows us to experiment in a safe environment. It tells us to stay critical about how we work and to keep reinventing.

Agile marketing teaches us to scale experiments down to a manageable size. If we try out smaller things and they do not work out, we just move on quickly to other small bets with little harm done. Agile teaches us how to fail fast, learn from it and then readjust. It helps us adapt in baby steps and reduces the threat of spectacular failures. It is less about incredibly expensive Big Bang marketing efforts, and more about divided little experiments.

Spotify is a great example of this, launching speedy experiments that are killed quickly if they don’t work. Their ‘squads’ (similar to scrum teams and designed to feel like a mini-start-up) are even encouraged to spend roughly 10% of their time on ‘hack days’. On those days, employees do whatever they want, typically trying out new ideas and sharing them with colleagues. Google, Facebook, and Uber also share this practice. Continuous monitoring and measuring of how our efforts influence consumers allows us to keep learning what is working and what we should change.

Practice makes perfect

Let’s imagine this process in action. Your team has been firing out solutions throughout a focused work sprint. Conversion rates are helping you in your drive to boost sales, but you’re falling just shy of the target. You hypothesise that the changes you made have solved the problem, but can be optimised by testing iterations. This is the perfect opportunity to squeeze all remaining value from the good work you’ve already done.

This is where the principle of A/B (particularly when automated and prioritised for effect or multivariate testing) comes into its own. Real-time testing of two or more different options is an unbeatable way to understand which option has the greatest impact in terms of meeting your KPIs. At TNW 2013, Dan Siroker, a product manager from Google, revealed how A/B testing of varied combinations of homepage image and button copy had effected email signups and donations – two vital KPIs – in Barack Obama’s 2008 presidential campaign.

By uncovering the optimum combination (which, incidentally, turned out to be highly counter-intuitive), his tests enabled the campaign to attract 2.9m extra email subscribers and 288,000 more volunteers, who raised an additional $57m in campaign donations. Stellar results like these are commonplace among organizations that routinely A/B test and are undeniable evidence of the business value of testing.

Richard Tidman is managing director at creative digital production company Space66.