Experimentation in marketing is almost a century old, and there have been some world changing examples along the way. Pepsodent toothpaste got the ball rolling by testing the effectiveness of magazine and newspaper ads in the 1920s using nothing more than coupon copy.
Predictably, it took a science-based company to pioneer these first A/B split tests and Google use exactly the same principles today to optimise AdWords. And how can anyone overlook the impact that the first Pepsi Challenge taste test had in 1975?
So why have so few companies taken advantage of this approach, especially when digital makes it so much more accessible? The culture around web optimisation has been long established, but very few regularly test their creative, messaging or propositions. And still, scores of marketers place more value in emerging platforms and technologies than in the things they publish for people to see and read. That is until now…
How do you build a research framework that allows you to gather information in real time?
This is what the guys at VanMoof bikes famously did. They suffered at the hands of couriers delivering their bikes to their customers damaged. Bad for the customer, bad for the brand and bad for the bikes!
Their solution showed how experimentation can be cost effective by working with their packaging; at first increasing the box size and filler, latterly experimenting with box labelling like ‘fragile’ and ‘this way up’. But their most effective solution was packaging their bikes up like they were a TV. In a box, marked TV! Fragile! They even printed a TV on the side of the box. All to make the delivery guys treat the packages with care. It was an iterative process and it worked. Bikes were delivered intact.
This fake TV packaging for VanMoof bikes has become a fan favourite, and they now sell the box as a stand-alone product (with the bike inside of course).
So how can you apply this to your marketing?
The steps the team at VanMoof went through is the same process that's needed to facilitate any radical concept through innovation. I'm just thinking back to the halcyon days in the school science lab.
Minimum Viable Experiment (MVE)
Start with an MVE – the least you can possibly do to give you the information that you need.
Promote your product or service with different campaign creative, messaging and promotions. Take a control forward (that's the best performing). And do this before you commit to one option.
Offer choice to highlight preference
If people choose a new product off the back of your campaign that you can't provide, simply stipulate that it's out of stock and offer them the original. If enough people show an interest, put the alternative into production. How is offering people a product that doesn't exist a good user experience? Well it's not ideal, but the learning value outweighs the customer disappointment and helps improve product strategy in the future.
Hypothesise how to affect user behaviour
Highlight the data that supports your experiment. And don't settle for ‘why aren't people converting?’ It's too broad. Break each big question down into subsets. Always define which experiments will help you find the answer to each subset before you start, and build them out from there.
This framework can be applied to anything from your advertising strategy to your customer journey. We’ve found that experimentation can now form the basis of most marketing strategies, and will quickly lead to a performance advantage in media, on-site engagement and product innovation. We recommend you start using it before your competitors do!
Sammy Mansourpour is managing director at integrated agency AgencyUK .