elizabeth burke wang
hello, i'm elizabeth. i take myself slightly too seriously, because of YOLO reasons, and I'm also a storyteller.
i have spent 10 years in corporate america, half on consulting for media & entertainment research arms of McKinsey and Nielsen, and half on the client side in retail and financial technology marketing. i love customer insights, test design, modeling, forecasting, and planning. i like to work in multiple fields-- i care about how humans spend their time, and they do that across many fields. i have worked designing survey questions, querying thousands of databases, building automated reporting, building attribution pathways and methodology, working with data engineers to track new metrics, modeling data, establishing norms and context, and both forecasting and planning results. i've worked with datasets anywhere from 2 subject matter experts to 5 survey respondents to 80 million users, and as is the case in many professions, scale has been an amazing challenge.
the scope of my profession has shifted so dramatically in the past decade-- a wealth of brain power has been poured into the digital measurement space, and so much of american time use has shifted online from more traditional media channels. i believe that the technological frameworks i've worked on in my profession-- establishing attribution pathways for marketing channels, determining measurement frameworks that are effective-- are seeping into civic life. Misinformation, disinformation, algorithmic content recommenders and suppressors; these all have profound impact on civil discourse, our political discourse, and even our sense of reality.
as someone born into the age of information (and in redmond, to a Microsoft technical writer, no less!), i was raised with a sense of hope that more measurement and more information would help us create a better world. and for some, it really has. but for those who can't contextualize data or who don't have the skills to critique and understand the scope of large datasets, i fear we have done more harm than good. the tendency i have seen for leaders in my field to lean on Data Science and Machine Learning as a magic mirror is upsetting. many people far smarter than i have written at length about the dangers of the algorithms that have been introduced to us without proper feedback loops and evaluation. it is scary stuff, and it does not get the attention that it deserves.
the question i'm interested in, however is just this: Where is the Bright Line? this is the same question that has powered my career, which especially in CRM and customer insights is largely made up of exercises in high-ROI stereotyping. stereotyping is a necessary mental task in marketing; who are we talking to? what do they love, and what do they hate? i've helped companies move past deceptive concepts, like the mythical "average customer." but ultimately, there is no objective truth in the work that marketing analysts perform. there are only angles, and commonsense practices. especially when it comes to using historical data to predict future data. in fact, my biggest accomplishments at any company haven't been in proactively creating customer segmentation and frameworks for creative messaging-- the best things i've done for anyone have been to help them set up norms and processes to stop doing obviously stupid things. After 10 years, I've never found a Bright Line that wasn't "This thing we are doing is clearly bad for the customer" or "This person apparently died in front of their TV and we need to throw their data out." Along with the maxim, "Anything humans do, some of them will overdo."
if you have need of conscientious customer segmentation work, you can hire me to help you with customer segmentation research and frameworks. i can build end-to-end cluster segmentations, help you build marketing norms using predictive modeling (including but not limited to decision trees and random forests, time series models, OLS models, & boosting models). i can even help you build custom parametric models to help you plan marketing spend, email volume and frequency, and web traffic.
my education and skills
columbia college /08 - econ & psych
johns hopkins data science specialization /15
data science dojo /17 (this is incredibly fun and i recommend it)
languages (all intermediate to advanced)
sql (standard and legacy)
music: violin, guitar, piano, voice
styles: cape breton, irish, scottish, bluegrass, folk
drawing: digital art, photoshop, and i'm learning animation
sound engineering: logic, protools