Monica Rogati grew her career at LinkedIn and Jawbone learning how to 'tame big data' and now successfully advises other companies on how to mine the messy abundance of data they collect into 'products, actionable insights, and meaningful stories.' At LinkedIn she helped create a Talent Match system that matched jobs to candidates and at Jawbone, developed data products to help promote healthier lifestyles based on data from the UP wristband. Monica, who was listed on Fortune Magazine Big Data All-Stars (2014) and Fast Company's 100 most creative people, spoke to Talking Cranes about women in tech finding champions and ROI in the data science industry.
Hi Monica, as an industry leader and champion of data science, could you tell us a little bit about your role and background?
Sure… so I am one of the data science advisors (at Data Collective & Insight Data Science). My background is in applied machine learning and statistical text mining. I’ve done my PhD at Carnegie Mellon.
Then I spent 5 years at LinkedIn building data products, which means recommender systems – anything that has to do with improving the user experience using data, and at the same time I found all kinds of interesting patterns in the data (in the LinkedIn data) I was able to share with the rest of the world – that was very, very interesting and lots of fun.
And then I moved on to Jawbone where I built the data team from scratch…and actually built the data products infrastructure, and again, found a lot of interesting patterns in how people move and what influences activity in sleep – the effects of sleep on happiness and so on, while at the same time improving people’s lives through data products and so on.
And more recently I’m a data science advisor where I help a lot of companies make the most of their data – to think about what data they should be collecting to build better user experiences.
What do you think that organizations can do differently to really move the needle for women in tech?
That’s a tough one because I think largely it goes beyond the organizations themselves. I think there’s a lot of factors that are outside of an organization’s control around culture and how things are at the wider level.
One thing that I’ve seen companies do is offer classes around pitching your work and how to showcase your work in very brief sentences…so that you develop these relationships and people who know what your working on and people who become your advocates.
Right? Because mentors is one thing, but being your champion when you’re not there, having somebody at the leadership level be your champion –they can’t do that unless they know what you’re working on. So being able to quickly tell them what you’re working on and why that’s important, what impact you’re having – when you see them around at coffee or the water cooler – that’s something that’s important that women can benefit from.
You’ve grown in your career, you’re a leader in your field, and as you said, from a technology standpoint you’re a champion for what your technology is about and what your research is about– so what was effective for you in growing in your career?
It’s hard to say right? Because you know we don’t have an A/B test…we don’t have a control model when we do certain things so that we can compare who got further. So it’s hard to pinpoint particular things. Being very strong technically has helped a lot because you can actually have those technical conversations and have an informed viewpoint - so that’s on one side – being technical, and then on the other hand, understanding the impact - the business side of what you are working on. Right? Because then that will allow you to bridge the gap between the business side and the technical side.
You said in data science, ROI is everything good divided by everything bad. What exactly do you mean by that?
Sure – so the question was with limited resources, which is always the case in data science, how do you decide what to work on? How do you decide what to spend resources on? And so, my answer to that is to evaluate the impact or attempt to evaluate the impact before you even start. Right? So think about, well if I work on this particular problem, what at the end of it, what’s the best case of what’s going to happen? Well this problem is going to be solved or I’m going to find this insight. What’s the likely case? Well..more likely, this is going to happen. What’s the worst case? Maybe I don’t find anything. And so you put that in perspective so that’s the impact and everything good.
And then you think about, okay, what does it take to get there? What’s the effort? What do I have to spend in terms of resources, social capital in terms of effort and team morale, or basically there’s (less everything bad) than what’s the cost – everything comes at a cost. You have to think through the very varied costs of going into that project.
Right? So that’s why I was saying that it’s everything good divided by everything bad, because you have to evaluate the impact in the context of what’s the potential reward and what’s the effort you have to put in it.