Best Practices: How to Harness the Chaos of Experimentation for Marketing Breakthroughs
As marketers, we are master tinkerers. We tweak this and modify that as we follow our customers in their individual digital journeys. The ad-tech revolution has given us many new tools in our toolbox to delight any hardcore marketer.
Ironically though, as tools got more sophisticated, learning became more obscure -- veiled in a faux "big data certainty culture" that leaves little room for the serendipitous learning achieved only through experimentation.
"I have not failed. I've just found 10,000 ways that won't work." -- Thomas Edison.
Experimentation done right is messy business, which makes it foreign to the "predictable," data-driven ethos of corporate culture. So we are faced with a maddeningly simple and seemingly irreconcilable conundrum -- how does one balance the need for predictable performance against the recognition that breakthroughs happen in the messy, unpredictable world of experiments?
I posed this question to some of my colleagues at forward-thinking organizations like AT&T, Siemens and Kimberly-Clark. In listening to their stories, I realized that these marketers are writing new playbooks about how to embrace the messy, failure-prone business of experimentation and still have a business win.
Here are four recurring themes that provide us with a roadmap on how to integrate experimentation successfully into marketing organizations:
1. Be a true guardian of the experimentation mission.
Technology and data dominate marketing, bringing along with it a very binary outlook to learning. There is either a winner ("1") or loser ("0") that overshadows the more subtle human-based learning buried beneath all that data.
Therefore, these organizations resist the temptation to pre-declare what a "winning" outcome would look like -- freeing them to see more clearly the unexpected connections that happen serendipitously.
This level of marketing experimentation requires senior advocates with deep experience who can judiciously use big data balanced with intuition to guide how results of experiments are interpreted.
2. Get everyone in on the act.
Some companies are baking experimentation into the agency-client relationship by rewarding experimentation -- financially -- for its own sake. Others are introducing the experimentation concept into the pitch process -- asking agencies to present ideas about proposed experimentation learning paths. Some advertisers even ask agencies to share their worst experimentation results.
Marketers are starting to realize that only through experimentation can the value of ad-tech fully reveal itself. As this is a new frontier, marketers are placing high value on agencies that can be true partners in the experimentation game -- the good, the bad and the bizarre.
3. Make small decisions more often.
Everyone talks the talk of experimentation, but in private, the fear of failure hangs heavy in marketing corridors because marketing risk is already high, dampening everyone's appetite.
Yet innovative companies understand that evolution requires abandoning "big testing/big learning" and migrating to practices where "small testing/small learning" is used to get at "big answers."
This approach offers less risk and better answers that can be developed over time, as Mayur Gupta, global head of marketing technology and innovation at Kimberly-Clark, explained to me. This liberating insight allows operating teams and business leaders to embark on a continuous learning path despite the unpredictable nature of experimentation.
4. Agile marketing is about mastering "less is more" thinking.
In the current ad-tech deluge, one is tempted to spend more on tech or the latest platform. But the "less is more" approach to experimentation is being put into practice at these companies. For instance, Bill Stabile, executive director-branding, advertising and sponsorships at Siemens, said he uses different types of technology to assess the changing interplay between core marketing functions. More generally, he said experimentation can be applied to great effect in:
Creative: Understanding how platforms can help marketers move from a process of "big creative" with "big nine-month campaigns" to smaller campaigns with fast iteration for testing, learning and optimization.
Content: Developing new types of "always-on" content that can be deployed based on data and insights instead of creating expensive "big and perfect content."
Media: Testing alternate platforms and networks for lots of smaller learnings that can roll up into a cohesive "big idea."
The tech and data deluge has temporarily blinded us to the fact that done right, experimentation teaches us how to merge the "unpredictable human element" and science side of marketing. It's time to not only destigmatize failure, but to even celebrate experiments gone wildly wrong.