As part of my research on the year ahead, I embarked on a series, 20.17 Big Ideas for 2017, to ask a number of my favorite award-winning marketing experts, authors, and other thought leaders-- as well as some of Firebrand Group's own digital strategy and branding experts--to recommend one "Big Idea" that companies can take advantage of to get ahead in 2017 (find the whole book here).

One of the individuals I was fortunate enough to interview for this series was Abby Whitmer, Director of Global Digital Marketing for Mary Kay. Whitmer, a longtime beauty insider, was added to the beauty giant to enhance its digital innovation, so it's no surprise that her big idea for 2017 involves Machine Learning. Here's our conversation:

How are you using machine learning in your day to day?

I'm trying to focus on how machine learning can be used now - as opposed to a very pie-in-the-sky, long-term way. Can we ask one question, like how to reduce customer churn? And let the data tell a story; let the algorithms point out the anomalies instead of us trying to manipulate the answer.

What's the best way to roll out machine learning within a startup?

In 2017, the leaders who will win with machine learning will be taking it one step at a time. I feel like those iterative steps will be huge for a number of different aspects of any business: for customer service, product forecasting, predicting consumer's going to be all about utilizing these new predictive analytics tools. For Machine Learning, it has the potential to be transformative in so many parts of the enterprise, so you simply can't tackle everything at once. For example, in a company likes ours, you can't tackle predicting every potential customer crises and also try to understand all emerging beauty micro-trends in Asia Pacific.

You have to start somewhere, but start small. Have clear ownership and avoid "too many hands in the pot" syndrome. Right now everyone is talking about machine learning, but not everyone has the skill set to utilize it within your company. Those skills are still largely in academia or in big consulting houses. So if you're leading the charge, take the time to educate while you test and learn.

Should we be cautious about wading into the waters of machine learning?

People are talking about it like it's here (and it is), but they're still a little scared and don't quite know how to employ it properly. Right now, you do need a sophisticated tool to utilize machine learning. By definition, it has to be able to teach itself, and if you have the right objectives and questions you can also help guide it. It requires practice and strategic Q&A and a lot of self teaching.

Use a tool like IBM Watson that you can play around with. You can pull in your own data. Pull in your own Twitter or LinkedIn or cell phone records, and start asking it questions and see what it comes up with. Because of advancements in natural language processing, you can talk to the machine like you would a person. You don't have to write in code to find a data field or sql query... you just ask and can see what it comes up with. Play with it on a personal level.

We may have the tools, but not every organization has the will. How does one sell this into their organization?

For me, you have to start everywhere...but not necessarily at the top. Instead, I've been going to every department's team meetings and showing how this little known tool, with just a minor hit to their budget, can help impact their business. From Legal to Protective Services, Events, Sales, Customer Service, Brand, Quality Assurance, Marketing, Finance and forecasting and beyond.... what does it mean for them? Develop use cases for each part of the organization. You don't have to tackle them all right now, but get people inspired.

Show that the data speaks their language and power users will crop up from all over the organizations. Then you don't need to sell it. Pretty soon, it will be used in too many places for anyone to try to roll it back. And then the idea trickles up. The C-suite now wants to understand it, wants to fund it, wants it to move faster and add resources because their hearing the impact stories from everywhere.

Here's wishing you lots of success with embracing machine learning in 2017. It's not as scary as it sounds!