Are your models annoying or delighting your consumers? Marketers have been using machine learning for years to help create advanced models, but that is not going to cut it in the future. Too many personalization efforts today rely on algorithms that are turning people off. And our attempts at automated personalization engines are forcing consumer audiences to simply tune brands out.
In this episode, we explore the connection between models, analytics, and even Eric Clapton with our guest, Stephen Yu. As a songwriter, musician, and master modeler, Stephen brings a unique perspective to data science and a cautionary message as well: Those who rely solely on automated analytic solutions will be the first to be replaced by the machine!
Ruth Stevens: What should our listeners know about analytics and modeling so they can prepare for the future?
Stephen Yu: The idea of applying modeling techniques to marketing and targeting is an old idea. The major difference now is that it is much faster. There is real-time personalization now. If you’re not fast, it is already too late. Also, it is not optional now like it used to be. The data is so abundant now, and without modeling, it's hard to dissect through the customer affinity or how likely they are to do something. In the old days, you had to have a degree in statistics to even think about modeling. There are a lot of easy tools out there now.
Ruth Stevens: Consumers are getting annoyed by some of the things marketers are doing as a result of what they’ve learned through modeling and analytics personalization. We are bombarding people with irrelevant or offensive messaging. How can we think differently about analytics and modeling so that we don't make these terrible marketing decisions?
Stephen Yu: I summarize this as humanizing the data. It is not just about numbers and figures, it's about a person. There are people behind the data. You don’t know somebody just because you know some tidbits about a last click. Just because you know something, don't brag about the knowledge. Turn it into a gentle nudge instead of saying, “I know you're looking for this new phone because you clicked something.” Don't say you're tracking every move that they make. Instead, at the right moment, give them a gentle nudge – “would you be interested in something like this?” Known, explicit data is hard to come by. Let’s say you are running a cable company, and you know somebody is cheering for a football team, next time he shows up on a page, you show him that team, but what if you don’t know anything about that person. You don’t want to show them the same thing over and over again just because they clicked it once. You’re only using a fraction of that data, and then you act like you know that person, and it is annoying. Models can help explain movement, but they aren’t the hard truth. They give a likelihood of something – such as you’re likely to be an early adopter, you’re likely to be a family person, etc.
Ruth Stevens: Is modeling and analytics going to replace us marketers in the future?
Stephen Yu: I’m worried about folks who just buy some personalization engine and think that the job is done for them. That's exactly how you get replaced by machines. We should strive to give purposes to machines. If you run a team of statistician or modelers or data scientists, you never just let them go, you’re involved. You have to have the same relationship with the machines as well.
Cyndi Greenglass: What are the skills and attributes that an aspiring data informed marketer should have in the future?
Stephen Yu: The math and programming skills go without saying. If you can’t go through mounds of data at ease, then the end result won’t be good. You have to be logical. If you had to build models, even as an exercise in school, that goes a long way. Also, read about the philosophy and history of marketing and travel and see what transactions are like all over the world.
Cyndi Greenglass: That's the key—think like a human, not like a machine. Be curious, empathetic and broad minded.
Key Takeaways/Three Little Piggies
- The models are only part of the truth about a consumer and they're certainly not the hard truth. They're speculation.
- It is important for marketers to be empathetic and to look at the world and be curious. We have to be broad thinkers. Models and analytics can help us solve problems, but we must bring a broader understanding of the world to solving that problem.
- Marketers have to be humble. We have to go into the market with an understanding that we don't know the truth. We can’t assume that consumers are behaving the way the models have predicted so our messaging needs to be softer or at least less conclusive.
Marketing Communications Today presents Horizons, it’s forward-thinking, looking ahead, through the front windshield and beyond, into the marketing future. Join Cyndi and Ruth bi-weekly for new ideas, technologies, tools and strategies that are emerging to help marketers navigate over the marketing horizon.
Meet our guest
Stephen H. Yu is a world-class database marketer with a proven track record in comprehensive strategic planning and complete tactical execution, from data collection and database design to model-based targeting and delivery, effectively bridging the gap between the marketing and technology world with a balanced view obtained from over 30 years of experience in best practices of database marketing.
Currently he is President & Chief Consultant at Willow Data Strategy. Previously, he served as Practice Head, Advanced Analytics & Insights for eClerx, and as Vice President, Data Strategy & Analytics at Data Axle (formerly Infogroup), leading the analytical services and product development teams for both. He was also a co-founder, visionary and the principal technology architect of I-Behavior, a premiere co-op database company that was the first to fully incorporate SKU-level data into the targeting process, forever raising the price of entry for the entire list industry.
His extensive database and consulting assignments have included major financial institutions, service providers, telecommunication and utility companies, publishers, multi-channel retailers and non-profit organizations, for both B-to-C and B-to-B applications. He is an active speaker in conference circuits and a regular contributor of articles in industry publications.
Meet the hosts
Cyndi W. Greenglass is a founding partner and president at Livingston Strategies, a data-informed, strategic consulting firm that helps clients develop, execute, and measure their customer communications with a close focus on results. Cyndi has razor-sharp strategic skills matched by impeccable on-the-ground savvy and tactical abilities. She is an Adjunct Instructor in the Data Marketing Communications online master's degree program from WVU.
Greenglass has twice been named into the Top 100 Influential BTB Marketers by Crain’s BtoB Magazine and was the 2012 CADM Chicago Direct Marketer of the Year. She is a member of the Board of Advisors for BRAND United and has taught, trained and presented at over 50 conferences throughout the world.
Ruth P. Stevens consults on customer acquisition and retention, for business-to-business clients. Ruth serves on the boards of directors of the HIMMS Media Group, and the Business Information Industry Association. She is a trustee of Princeton-In-Asia, past chair of the Business-to-Business Council of the DMA, and past president of the Direct Marketing Club of New York.
Ruth was named one of the 100 Most Influential People in Business Marketing by Crain’s BtoB magazine, and one of 20 Women to Watch by the Sales Lead Management Association. She serves as a mentor to fledgling companies at the ERA business accelerator in New York City.
Interested in growing your career with a master's degree in Data, Digital or Integrated Marketing Communications? Request more information today!