Iteration is something that has been at the core of how Fight works from the beginning. While on one level it’s simply a process, its philosophical underpinnings say a lot about how we think about things and how we approach our work. It’s something I’ve spoken about, written about and been called out on. Listening to Jared Spool speak Wednesday night about his research on Amazon’s successes, failures, and their approach to risk, it all seemed to perfectly demonstrate how iteration can be a powerful way to not just succeed, but to succeed in ways that aren’t possible working in one shot campaigns.

A key point in his presentation was that taking risks is critical to success for business. In his words – “risk averse companies produce crap.”

This was a point made over and over: Amazon is successful in large part because they take risks. They try things. They experiment. Equally important – they’re not afraid to try something and have it fail. But they’re not reckless, and thats critical. In fact, they’re incredibly effective at targeting their risk taking, and mitigating their exposure to potential failure. To illustrate this, Jared told the story of Amazon’s 2007 change from their well known (if slightly notorious) tabbed navigation to their current side navigation. Over the course of 12 weeks Amazon rolled the new system in four phases. First, only 5000 non-cookied visitors would see it. These are considered the least important customers to Amazon, so failure here wouldn’t be devastating. Next they rolled the new navigation out to 1 in 5 non-cookied visitors. Then 5000 cookied visitors. Finally, 1 in 5 cookied visitors saw the navigation. At each step Amazon would gather data on the performance of the navigation, and make adjustments based on this. The full release cycle took 12 weeks. Making fundamental changes to the navigation system on a site that millions of visits a day presents a huge risk for Amazon. But keep this in mind: the average sale for Amazon is $35. If they raise that $5 to $40 it would bring in an extra $875,000,000. Thats a lot of motivation for them to try things, nearly anything, to see what works.

I think this story demonstrated perfectly the difference between risk taking and recklessness. Between having a goal, and running experiments to discover new insights, and just tossing things out there to see what works. Marketing on the web can be the perfect opportunity to work iteratively and to experiment; but I think too often projects are treated as “all-or-nothing” endeavors where you get just one try to get something right.

By setting clear goals at the start of the project, and approaching it iteratively, even when an experiment doesn’t go as planned, it doesn’t have to be a failure. This is what I find troubling about the notion of “failing harder” that seems to affect so much of the marketing culture now. It sets up a false dichotomy that says to succeed big, you need to fail big. I’d rather succeed big, and fail small, and I think Jared’s insights into Amazon show this is entirely possible to do.