Data-driven Decisions vs. Experimentation

"The first leap is on gut, but the data helps you refine as you go." - Michael DelRossi

This week I attended the Marketing in Context conference at Boston University. Attendees hailed from Tufts, Brown, Boston College, Brandeis, Center for Digital Imaging Arts, and Boston University. While the conference focused on storytelling, a theme that stood out to me was the need for both data-driven decisions and experimentation in telling better stories. There are several approaches to finding that balance between testing new paths and relying on data. 

“The first leap is on gut"

Michael DelRossi of Analog Devices acknowledges that you won’t always have data when you start. You will have to do some qualitative analysis and take a first leap based on your gut. “You might do an A/B test … you’re using that to refine the next message, and the next message after that. At each iteration, you’re using the data.” He remarked, “the first leap is on gut, but the data helps you refine as you go.” 

Emerging channels

The topic of emerging channels was addressed by Sean Joseph Zepps of Deep Focus, who said that some channels are so new that not many brands are on them. You’ll have to take an organic approach because a paid approach isn’t always the best approach on these channels. Figure out your budget of money and resources for established channels, then look at the emerging channels separately and decide where you’re willing to test.  

Marketing in Context conference at Boston University, March 25, 2016

Influencer experimentation

Mike Proulx and Mike Platco agreed: let influencers tell your brand story. It is a leap of faith for companies to let go; “many of them aren’t willing to do it,” acknowledged Proulx. Platco, who works for Snapchat, pointed out that when you don’t have content on Snapchat, you don’t exist, because Snapchat only shows content that is less than 24 hours old. Slow and steady growth can work. He shared how Coca Cola started hiring influencers for 6-, 8-, even 10-week stints and they would share two stories per week. Platco says some brands let creators do whatever they want, with approval, on the account. 

Data as a new industry

Platco explains that a platform like Snapchat, which has very little data, has resulted in new companies emerging that extrapolate the data and sell that data to brands. “It’s not that they can extrapolate anything that anyone can’t do themselves, but just the ability to put those into an algorithm … retention and drop-off rate … brands can’t do it themselves and that has become an industry.” 

Reserving time for experiments

Jaime Mahoney of Keurig Green Mountain claims that from a brand manager’s standpoint, you don’t have an answer to a lot of questions so you have to go on experiments. “If it doesn’t work, move on.” Proulx pointed out that “technology allows brands to test different ways to express their story and reserve some of their energy, time, and focus for experimentation."

How much to measure

Anish Kattukaran of Google explains, "Sometimes you want to act on as much data as you can get. Other times, you want to act quickly on as much data as you have.”

Data-driven or experimental?

The key is to use both data-driven approaches and experimentation to your advantage. Some emerging platforms lack the robust analytics offered by long-standing platforms, and experimentation can be a way to gather the data necessary to drive future decisions.