From IT to Product Management to HR, data has become an essential component of successful problem solving. At Movéo, we know that this is true for our clients, but as part of our Fourth Evolution, we decided it was time to take a step back and make sure that we were utilizing data in the most effective way possible in our own business operations. To do this, we recently held a “data bootcamp” in order to give our employees a new way of thinking about data and an understanding of what resources are available to them to help them solve business problems. Here are some key lessons we learned:
We began with a session on causality and experimental design. In this session, we learned the difference between correlation and causation/attribution, and learned strategies on how to critically think about attribution of relevant events. We went through an example of a simple but widely used experiment type—the A/B test. We then discussed other examples of marketing tests and how they can influence decision-making in marketing and improve efficacy of our campaigns.
Afterward, we held a session on the iterative design process and data visualization, where participants got to experience how to design data visualizations firsthand. Participants took part in key elements of the design cycle and experienced the differences between what is “interesting”, “pretty” and “useful”. We also touched on some general principles of data visualization design, including things like ink density and Occam’s Razor. Importantly, we were pushed to think of other data sets that might be relevant to a problem at hand, and realized data visualization is not about visualizing the data you have in front of you, but rather about helping someone solve a problem.
Our next session focused on metrics and transparency. We discussed ways in which certain metrics (i.e. click-through-rates, online 5-star-based reviews, etc.) can be used as tools to drive decision-making. We discussed how to be as transparent as possible in creating metrics, providing context and using relevant data to draw sound conclusions. Team members came away with the confidence to draw their own conclusions about quantitative results, the ability to critically interpret the meaning behind various metrics, and the desire to create metrics that transparently describe the business situation at hand.
Finally, we reviewed models and algorithms, giving a high-level description of what a model is, providing examples of some relevant models, and going over how to read and interpret these models. We discussed the trade-off between model simplicity and accuracy, and described the benefits and limitations of thinking about real-world phenomena in terms of models.
This was a great opportunity for us to revisit data strategy and prepare ourselves for a 2015 completely guided by data-driven marketing. Let us know: what parts of your business could you use a bootcamp on?
Photo Credit: Waag Society via Flickr Creative Commons
(Click here to expand)
