Q&A: Enova’s DeCosmo on Analytics Tech Trends

Chief Marketer recently talked with Joe DeCosmo, the new chief analytics officer at online consumer lender Enova, to get his take on trends in data and analytics.

CHIEF MARKETER: What are the big trends when it comes to tech solutions for analytics?

DECOSMO: Data sources. As we can identify new and different data sources that we can tap into on an ongoing basis to drive our models, we can make better product decisions. We work with the major credit reporting agencies and others that bring specialty data sources, as well. Everything for us is driven by the customer experience so we don’t do a ton of outbound marketing, we’re primarily doing things like search and pay-per-click campaigns, but for us it’s all about once the customer comes to our website, how can we most efficiently get them through the process of applying for and either receiving a loan or not. For us it’s about real-time data processing, look-ups. We leverage open source tools pretty regularly for that, and they allow us to constantly improve our ability to handle more data more quickly—both on the front end and back end for scoring and decision-making.

CM: What is it that you and your team like about the flexibility of open source analytics tools?

DECOSMO: To be able to iterate really quickly on a new model that we want to test, or if we want to run some experiments on the website to see how customers react, the newer development tools allow us to iterate very quickly on those kinds of things. It’s been a big thing for us. I’d say the amount of data that we’re dealing with isn’t extraordinary when you’re look at what most companies are dealing with these days—it’s really the speed that we can integrate and make a decision that’s crucial to us, because we don’t want to leave that customer hanging.

CM: Is it a challenge for you and your team to keep on top of new analytics tech solutions, or have you found the sweet spot?

DECOSMO: From an analytics perspective, we keep on top of techniques and algorithms and methodologies and data sources. Those are the things we watch very closely, and make sure we’re making time to research and experiment with some of these solutions. The tools themselves? We don’t spend a lot of time on those. We have a short list of tools that we like and that our team is comfortable working in, like R and Python and SAS, to some extent. We don’t pay much attention to what the latest and greatest modeling tool is out there.

I’ve been in analytics for 27 years now, and from my standpoint there is such a variety of tools out there now that I think it’s easy to find a couple that you can build around and for your team to focus on. That’s why for us it’s a couple of open source tools, one or two licensed solutions and that’s kind of our toolbox. The interesting thing is that as we hire folks out of college, they all know the open source tools and it’s actually the minority that have experience in the more traditional licensed software tools. That’s much different than it was 10 years ago.

CM: What’s next? A continued focus on speed?

DECOSMO: We haven’t invested a ton into the front end of marketing technology, an that’s an area of opportunity for us, so I hope a year from now I’ll be able to talk more about how we’ve been able to move the needle on the front end, as well, to make our overall marketing spend more efficient than it is today. Flexibility of development and speed of decision-making is a constant theme for us, because data will always be there.