Redesigning the mortgage application experience for Capital One.
TEAM | Chris Ryan, Nicki Lee, Chad Poorman, Jarred Brooks
ROLE | Experience design, service design, research strategy
TOOLS | Axure, Sketch, InVision, Mural.ly
METHODS | research strategy, service design, competitive analysis, market research, UX flows, service blueprinting, service modeling, stakeholder modeling
With Capital One, I worked on designing a completely digital mortgage experience that seamlessly combines the pre-qualification, application, and closing stages of getting a mortgage (shown below).
Initially, this project was presented to me as a conversation problem. How could we keep borrowers in the funnel, reducing drop-off between each phase of the experience? However, after reviewing the existing research and examining the competitive landscape, I realized that a more in-depth look at the entire application experience would be necessary to drive business growth (and higher NPS scores). My team agreed, and we began a more holistic redesign process during my time at Capital One.
We first documented the current application and then began iterating on the flow.
I conducted my own informal research with realtors in my network to validate the need for realtor input.
formal interview analysis
I analyzed small group interviews with homebuyers and individual interviews with realtors.
regulations and credit decisioning
We studied the regulations and internal policy decisions behind the application inputs.
We looked at direct and indirect competitors to see how they differentiated their experiences.
I looked at the broader real estate landscape to determine if there are future opportunities that we should build room for now.
I worked with tech and product partners to imagine an ideal future across physical and digital touchpoints.
I visualized the relationships between different actors and touchpoints that influence borrower's mortgage decisions.
To help establish shared priorities for product and design, I synthesized the customer journey, our experience goals, and the features that would accomplish those goals.
current + Future-State UX Flow
The application that was currently available to users had been designed with loan officers in mind. While not dissimilar from the applications of other major banks, it was simply a digital form, requiring all the same inputs, and in much the same format, as a physical application.
We began by creating a UX flow of all inputs for the current application. We then put everything on the whiteboard and starting clustering questions, removing questions as we eliminated them.
Finally, we put it into Axure, where we continued to iterate on the digital version.
During this stage, we found that many of the inputs in our mortgage application are required solely because a legacy system asks for it.
To support putting in the tech effort to reform these backend systems, we carefully documented each input we could cut from the application. Once we reached a critical mass, we began consulting with our tech partners to create a roadmap for the backend changes.
Guerilla interviews |Realtors
Early in the project, I began talking to realtors in my network about what advice they gave their clients when it came to selecting a mortgage company. While this was my own private research, it sounded the alarm on several key issues that came up again during formal research.
"I steer my clients away from big banks. Ideally, they work with someone local, so we can get to the root of any issues quickly."
Interview analysis | prospective homebuyers
Early, in the project, I began reviewing Capital One's existing research on the homebuying experience. We had a library of small group interviews with people currently looking to buy a home. Even though these interviews originally were designed to test the website, not the application, they yielded serveral key insights that drove our design process.
Borrowers who have been through the process before are happy to provide more information if it will make the process easier.
This was a particularly important insight for some of our digital data initiative, which asked borrowers for their bank logins in order to instantly fill and validate the assets part of the application. Based on this insight, we recommended a similar initiative for income verification.
Borrowers appreciate the small additional services we could offer them, things like a transparent look at their credit scores and reports.
While learning about credit decisions, we learned a lot of things that an ordinary borrower would appreciate knowing. We decided that if we could give borrowers useful insights, we should. As one borrower said, "You're giving me all this stuff for free, which helps me trust you."
Borrowers need to learn as they go, but they're also easily overwhelmed by too much information.
We decided that having the right balance of educational information and minimal design was the best way to help borrowers feel comfortable in the experience. Ideally, it should feel like a loan officer is there on screen, ready to answer their next question as it comes up.
interview analysis | realtors
Near the end of my time at Capital One, I learned that there were two days of marketing interviews with realtors scheduled to take place in the usability lab. I got permission to observe and from those sessions learned that a number of our earlier insights appeared to be true.
"With big banks, it's almost like you apply for the mortgage twice. Once with the application itself, and then again a week before your closing date when they suddenly want all this additional paperwork."
After listening to realtor's experiences, we found that handoffs resulted in large delays, frustrations, and occasionally even caused borrowers to lose the home they we trying to buy.
regulations and credit decisioning
Throughout the project, my team and I dove into the federal regulations and internal credit-decisioning guidelines that determined much of the mortgage application's content. By looking closely at the regulations themselves, we were able to determine why many application questions were asked and think about creative alternatives for getting that information.
A little something called the 1003 determines a lot of what is asked on a mortgage application.
The 1003 form is used to sell mortgages to the federal government and, since most loans are sold, it was imperative that we have answers for all required 1003 questions. With a little digging, though, we found that many fields on the 1003 are not required and could be eliminated from our application.
We went through application experiences of direct and indirect competitors, looking at banks, non-traditional lenders, SaaS providers, and lender aggregates to determine the common design patterns and service promises used in these experiences. We then had critical conversations with the entire design team about which of these service offerings and design patterns best fit Capital One's mission and users.
had traditional applications, with lots of inputs on each page.
were starting to offer banks an easy way to streamline their appication experience.
offered a marketedly differeent experience, with simple interfaces that gave borrowers extra guidance.
I researched the mortgage and homebuying experience at large, considering both horizontal and vertical components. I considered how we could build the current applicadtion to support possible future moves into other parts of the homebuying process, like search and selection.
There's movement within the industry to work across the homebuying experience. I saw particular opportunity for Capital One in partnering with with a brokerage service like Redfin.
In addition to thinking about the application itself, we wanted to think about how the experience would play out across touchpoints. I began by blueprinting with product and tech partners using the traditional categories, but soon zeroed in on current Capital One customers as our target for MVP. I created additional swimlanes for their experience, since there would be different tech solutions available for current Capital One customers than for other borrowers.
I visualized the relationships between different actors and touchpoints that influence borrower's mortgage decisions and quickly realize two things that would guide our design.
Many borrowers will use a mortgage broker or will automatically apply for a mortgage with their bank. This makes other bank's customers less accessible, but shifted our focus onto existing Capital One customers.
We realized that there were a lot of handoffs for a single application within Capital One. Could this be causing the problems the realtors were talking about?
To synthesize what our experience goals were for the product, what painpoints these goals addressed, and what features would be needed to meet those goals, I created an experience model to serve as a design roadmap. It included features to be included in MVP, and features to be developed for later releases, along with representative quotes from borrowers.
When I finished my time at Capital One, this project was well underway. Blueprinting was ongoing (I handed it over to Nicki Lee), but most of our research had been wrapped up and synthesized. We knew what kind of experience we wanted, and the questions going forward were largely ones of technical capability and prioritization.
Personally, I found this project fascinating because if offered the chance to improve an experience that is fundamentally unwanted, which posed unique design challenges. It was, as my dad would put it, a target-rich environment, which offered lots of opportunities for meaningful change. Three of the fundamental pain points that we decided to address were:
Applying for a mortgage is really hard. The paperwork is overwhelming, and can seem to go on forever.
The loan application process is incredibly opaque. Borrowers don't get to know what determines if they'll get approved, and even after they're approved they don't know what's going on with their loan.
Mortgage processing is often inefficient. Borrowers often have to scan, fax, or even bring documents in person. These documents are then faxed elsewhere, where a mortgage specialist catalogues them, and then moved again to underwriting, which can lead to requests for more documents.
I'm absolutely excited to see what the phenomenal team at Capital One rolls out!