Marketing to Millennials (Steven Dupree, VP Marketing of SoFi) : Interview with Steven Dupree

Glenn: Hi, everyone. I’m Glenn Gow, Founder & Advisor of Crimson Marketing. Welcome to Moneyball for Marketing where we talk about the incredible changes happening in marketing organizations around big data and marketing technology. We feature marketing technology insights from the top marketers in the world. The reference to Moneyball is from the story of how the Oakland A’s baseball team were able to win and win and win because they figured out how to use data and technology to their advantage. If you’d like to learn about how to use big data and marketing technology and marketing to help you win visit us at or email us at And now on to our podcast.
Today I am very pleased to welcome Steven Dupree, VP of Marketing at SoFi. Steven Leads direct marketing and growth at So Fi. His responsibilities include overseeing cross channel marketing such as referral marketing, email, search, display, mobile, direct mail, and out of home campaigns across products. So how does SoFi describe themselves? So Fi is a leader in marketplace lending with over 6 billion dollars in loans issued. So Fi is transforming financial services for financially responsible consumers with student loan refinancing, mortgages, mortgage refinancing, and personal loans. Steven it’s a real pleasure to have you as a guest.

Steven: Thank you Glenn it’s a pleasure to be here.

Glenn: You and I were speaking a while ago about how data is a key tool for you in your marketing efforts. And since that’s one of our favorite topics on here at Moneyball for Marketing, tell us your thoughts about how you use data and how you view it as a key tool.
Steven: Sure. So in terms of web optimization data is really good for us because we’ll do a lot of testing on our marketing site to see which versions of the pages are going to get the most people through. We use Optimizely it’s a very good tool for that. And the way that it will work is we’ll put together a hypothesis, often times it will come from product marketing, or from the executive team, or even from employees, and we’ll build out a test plan and run multiple versions of a page. And we’ll measure from one funnel point to another, and run it out to significance ad find the version of the page that works best to run. It’s been even more interesting more recently because we’re starting to really define what our brand guidelines look like. It’s parameterized in a test in a very special way. So for instance we may decide to change the button or the call to action on the page, but we won’t touch something like the logo. So that’s how we might do it for web.

Glenn: You had mentioned you also use data to decide what audiences to go after. I’m fascinated by that topic; can you elaborate on that?

Steven: Sure. We use data a lot in determining which populations we want to reach out to both to online and offline means. There’s the set of data that we get from people who have expressed interest or intent on our site so we’ll use tools such as Google Analytics and out own data warehouse to track people that have come to and have tried to register or apply to one of our products. So we have a lot of profile attributes that inform which publications they’re reading, which websites they’re visiting, and we’ll target those. We’ll use click based tracking so we can see if people click on an ad that we put out there. Not only can we see how effectively they’re getting to the funnel, but we’ll tie it back to spend in order to see where we should be investing more heavily and where we should be cutting.

Glenn: So you talked about determining what site they visit or what publications they read. How do you gather that information do you ask them or do you append information about them?

Steven: That’s a great question. So there’s 3 different ways we get at that information. We use the click based tracking so that’s dropping a cookie when someone clicks on an advertising link that we’ve provided somewhere, and we leave that back in our data warehouse. A second way is, many of the partners we work with such as display service providers and ad exchanges, and also social media vendors like Facebook and even sites like Yahoo, Yahoo has a native app called Gemini that allows them some sort of tagging capabilities so we’re able to place some information on our registration, and when people register we can actually feed that information (no personal information of course that goes without saying) but just send a beacon that a person that came form that a particular source hit that particular point. So we can use that for tracking. The third way that we gather the data is actually through self-reporting. We’ll ask people when they register how did they hear about us, and we’ll let them choose multiple options. And that’s a valuable way to get tracking on sources that are notoriously difficult to place technology around.

Glenn: Ok. Yes, we like to call it following the digital footprint. So you have a lot of different products you can offer. So if I come to your website, how do you customize the website in some way so that you’re providing just the right offer for me?

Steven: That’s a great question. So when people come into our website we have a menu of products, it’s one of the things that sets us apart as a financial services provider. We don’t offer just one product. I should be more specific, it’s one of the things that sets us apart as sort of a marketplace lender. We’re really focused on multiple product offerings not just one thing. We used to do just student loan refinancing, but we’ve branched into other products, and we’re going to be crossing and have products launched in 2016 that really round out our offering. And so in order to most effectively make the customer happy, we will change things that are are back end based on for instance whether or not somebody has student loan debt. We might then offer them a cross-sell opportunity from a personal loan for instance to a student loan refinancing but we know enough not to do so if they don’t have any student loans. So that’s something that we’re working on. We want to make sure that our web experience is really well suited for our customer which is sort of this older millennial, has invested a lot in education, has a lot of work experience, so we’re trying to keep everything clean, mobile friendly, and sort of keep response times to a minimum when people reach out to us.

Glenn: Good so that brings up a topic we talked about earlier as well, user experience, and your focus on making that as optimal as possible. Share with us some of the things you do there.

Steven: Yeah, so on the user experience front, we’re doing a lot of things differently than traditional financial services. So one thing we do for example is we offer unemployment protection to all our members. So with most of our loan types, if you take out a loan usually you have to be employed, but if you lose your job we don’t call you up for collections. We call you up to help you find another job within our SoFi member network which really sets us apart, and it’s really great for the end-user. So that’s something that we do, and we have done for a couple of years now. Another thing that we do that sets us apart, we really make it easy to apply onboard through mobile, so you can upload photos from your mobile phone. You don’t need to go scan or fax things like you might for traditional banks, so we really try to make that seamless.

Glenn: I’ll bet that’s a big deal for user’s experience to optimize for mobile.

Steven: Yeah, we’ve taken in a lot of feedback about that. One of the advantages of building a company that’s really new is you’re not encumbered by some of the legacy technologies and processes. So we can do things like that to really move it forward. I’d say one last thing that’s really good for our users too is, we really want them to feel like they’re part of a network. So I already mentioned the unemployment protection, we will give deferment to qualified entrepreneurs who want to start a business which is a cool feature, and we have pretty good incentives for people to refer their friends to the student loan or personal loan products, and that’s something we’ve honed over time. Because what we find is a SoFi member is a particular type of person that’s gone to school, often grad school, and knows a lot of other people that have say student loan debt or need to get their debt down, so they’re able to qualify their friends. And since trust is really critical, they can get them over that trust hurdle.

Glenn: That’s actually an interesting topic. So it sounds like referrals is a big part of your acquisition strategy. Tell us a bit more about what you do for customer acquisition.

Steven: Sure. The way that we do it is it’s a referral link that people get and they share with their friends. And one of the things we learned early on is that it’s great to have a publicly available referral program, but we also like to offer the referral program and hand it to people once they become members, because let’s face it, people are generally lazy, and so that’s something we’ve done for a long time. We have a symmetrical referral program, so both the referrer and the borrower receive some sort of payout for a successful referral, and that’s a really nice way to do it. Because the referrer doesn’t have to feel bad that they benefitted from helping someone, and that the borrower doesn’t have to feel bad about taking that person’s time.

Glenn: Interesting approach. So tell us some of the other things that you do for acquisition.

Steven: Sure. We do a lot of search marketing because it’s a great way to capture people who’ve already identified a need for our product or for student loan refinancing, or personal loan, mortgages more generally. We’ve done a lot to try to capture real estate on the relevant searches, and also build out things like site links to sort of drive volume. We use some geographic targeting, that’s a really useful tool particularly with something like mortgage where we’re not licensed in every state, so it really helps us to target and to also cater towards particular markets regulatory and otherwise.

Glenn: Sure ok. And you also say you use direct mail which is not common for the guests that I have on this show. Tell us a little bit about that, and particularly tell us how you tie the results of direct mail back into everything else you’re doing.

Steven: Yeah I’m happy to talk about direct mail. It doesn’t get a lot of love. We’re actually somewhat unique in that we don’t do a lot of direct mail compared to other financial service marketplace lenders. It’s a small part of our overall mix, but as a paid channel it’s an important part of our business. I think that when you work in financial services, the high acquisition cost corresponds to high lifetime values enables it as a channel. And in particular, there’s a qualification step that we can get with direct mail that you don’t necessarily get with other channels. For instance, we can screen for people that have never had delinquencies, or that they’re current on their loan payments. And we can look at people’s work experience, those really help us to target the right people. We actually announced yesterday in fact that we’re not using FICO in our underwriting. So we don’t actually need to look at the credit score, but we still look at other criteria, and direct mail allows us to be really targeted with whom we send to.

Glenn: And tell us about the data portion, I’ll call it direct mail response, how do you track the efficacy.

Steven: So direct mail is actually one of the easier methods to track and it’s been around a long time. The reason is because you can be dictatorial about where you want your customer to go. And so we provide vanity urls that are customized per piece. And so we’ll run cells of tens of thousands of names, and each cell will have a particular message, test or offer test or creative test, and we’ll look at our data warehouse. It’ll take a few weeks or months for the campaign to propagate through, but we’ll look at the response curve and figure out which variants are working best. We’ll take the best ones and make those our controls going forward, and over time we’ve become a lot more sophisticated. Not just in terms of the messaging and stuff but also to some of the audience segmentation and making sure that people that are never going to take a loan, we’re not sending them a piece of mail every month because it’s a sunk cost and they don’t like it either. We’ve actually been building a model that scores the likelihood of response, and using that we can actually get a lot more efficient with that channel.

Glenn: Excellent. You also used the phrase with me data warehouse, and we’re big fans here at Crimson of the concept of a data warehouse, not necessarily that it has to be a data warehouse itself but more importantly that marketers have a central repository for prospect and customer information that they can trust. I’d really like to hear you talk about what you’re doing to capture and manage that data, and then how it ties together with your various systems.

Steven: Thank you for asking about that. The data warehouse is different to every marketer, and different to every company in terms of its scope and its importance. It’s really important here at SoFi. It’s our source of truth because so much of our technology is part of onboarding and getting loans issued and customers signed up and funded, and so a lot of the information that relates to those activities is proprietary. We store it in a SQL warehouse that we’re able to query. We use Tableau a lot, that’s a really great tool that I’ve been a fan of for decades for data visualization and pulling. So we do use other sources, Google analytics is really important for we trafficking, and as I already mentioned some of the ad networks that we work with have data sources too that we pull from. But the source of truth is what’s going to allow us to tie all funded loans back to where people are coming from. And one of the things that makes a financial services company more challenging in terms of tracing is that it’s a higher consideration purchase. It takes somebody a little bit longer from the time of coming to your website before they’re fully funded. Hopefully just a couple days, but compare that with buying a t-shirt on amazon which you do in about two minutes. And a lot of those funnel steps such as moving someone through the review process actually occur offline, so there are things that you wouldn’t be able to track easily with web analytics anyway. So that’s why the data warehouse becomes so important for us.

Glenn: And you mentioned using Tableau for reporting and attribution. I’m particularly interested in how you use attribution, especially because you have both online and offline communications.

Steven: There’s a lot of ways to do attribution. I talked earlier about the real crude ways of asking people where they came from. Of course we’re more sophisticated than that. There’s click based attribution which is tied to cookies, and then there is I guess more recently multi touch attribution where you’re able to weigh all the touchpoints of somebody seeing and clicking on digital ads or footprints throughout their process, and weight them appropriately. You can also look at things like radio point and time type investments. So I really believe heavily in multi touch attribution. We’re somewhere in between those latter two. There’s a lot of places we don’t track because of security reasons. We’re very careful about what code we place inside, anywhere near or around any sensitive financial data. So for that reason it limits some of the tools that one would use to do multi-touch. However, we have a really robust way to get at the big base, and we have some things that we’ve put in place to layer on understanding of other touch points to give us an adjusted model that works really well for us.

Glenn: Alright excellent thank you. Yes, attribution is a very tricky topic but one that is of great interest to our audience. So Steven as we wrap up here let me give you a chance to just think ahead for a moment for the next year. How do you see the future as it relates to data and technology, and what might we be looking for if we were to go to your site in the future? What kind of improvements would we see?

Steven: I think the future is bright in terms of data and technology. One of the things that we’re exploring that I think a lot of companies are trying to solve in 2016 is how better to understand the impact of brand on direct response. Some of those really big investments that you make to raise awareness and drive consideration which ultimately lowers the acquisition cost across all channels in the long term. And we’re making some big investments on the brand side. I think there are some tools that I’m looking forward to reaching and particularly implementing, I’m not going to name them but I think as a whole we’re not the only ones doing this. You see a resurgence in tech companies investing in things like out of home and television. They’ve realized it’s more than just getting people who are on Google to come to your site instead of the competition, it really is building trust and a relationship with the customer. So I think what you’re going to see next year hopefully are some good tools and technologies that support brand building and merge it a little bit better with the acquisition side of the world, because a lot of companies have that problem.

Glenn: That’s very interesting if we can measure the impact of building trust successfully across multiple channels, that would be enormous.

Steven: I would love that.

Glenn: Alright well listen we are actually out of time, so I really appreciate what you’ve shared with us today. I’ve learned a great deal, and thank you so much.

Steven: Thank you very much Glenn it was a pleasure.

Glenn: Alright talk to you soon.

If you like this podcast, please subscribe and rate us on iTunes and tell your friends about us. You can also go to our website,, and sign up for our free monthly newsletter featuring the very best of our marketing insights, featured Moneyball for Marketing podcasts, and one of our favorite features called, “Bad Marketing,” or email me at Thanks for listening to Moneyball for Marketing from Crimson Marketing. Have a great week and let us know if we can help you in any way.