Role: senior Product designer
Product Selector for Lenders is a loan product and pricing feature for Roostify, a digital mortgage banking solution. Roostify's mortgage lending web platform allows borrowers to easily apply for a mortgage, upload documents, and communicate with loan officers securely. Loan officers can manage their pipeline, create tasks for borrowers, and share product and pricing information with their borrowers.
The Product Selector for Lenders (PSFL) is a digital lending platform experience allowing loan officers to access, model and share product and pricing options to borrowers, leading to increased overall loan conversions.
As design lead for this project, I worked closely with Product Managers and Engineering leads to participate in this service design initiative, conducted research, and synthesized the findings that led to this feature.
We wanted to make sure that our tool could be accessed at any point in a borrower's journey: whether they had completed the application or not, so it was important to have personas that could reflect different borrower needs and situations.
Our goals were to produce additional features for the platform that would increase the overall speed of the process. Research determined 3 areas of opportunities, which became separate design initiatives: a faster automatic underwriting system (AUS), a product and pricing feature with the ability to include fees and closing cost calculations, and some self-driving features for borrowers that would save loan officers time (credit pulls, pre-approval letters).
How might we facilitate an evolving financial decision-making process around product selection and fees for home mortgages that positions the lender as a trusted advisor for their borrowers?
Insights from on-site user research showed us that loan officers, processors and underwriters use many tools to qualify borrowers for specific sets of loans. On site observations and interviews revealed the heavy overload of software tools they relied on to gather their customer's data, run conditioning, surface loan products, and make certain rate or pricing calculations.
Understanding borrower goals helped present relevant products, such as for instance, how long they planned to live in their home. Focusing on the minimum amount of information necessary, I sketched out simple user flows for potential conversations between loan officers and their borrowers.
Ideating Phase: Sketches
Rough 2 minute sketches to come up with as many ideas for the sections as possible.
Compiling User Scenarios
By roleplaying and taking from what was learned about the process, multiple possible scenarios were compiled. The goal here was to account for varied borrower conversation flows and touchpoints of data exchange.
Multiple potential user flows were consolidated to 3 main different scenarios: just browsing, or shopping for rates, making changes during the data gathering process, and situations with more than one borrower, where group decisions had to be made. We called them Shopping, Changes, and Decisions. A service blueprint was applied to each one.
Once the touchpoints were established, I created wireframes to show the minimum required data inputs necessary for running a DU (Desktop Underwriter, an automated underwriting engine) to meet the lender's conditions for issuing loans. Pulling from the 1003 Roostify inputs, I organized the data into groups and then sections (below), keeping in consideration the ability for Loan Officers to edit some of this data.
Loan officer component sketches for sharing information to a borrower.
We learned from our customer advisory group that forms should be dynamically generated (as opposed to the legacy UX for the Borrower application flow of progressive disclosures) as some of the criteria entered would surface different form fields. Secondly, in cases where a loan officer was conducting a product search for an existing lead, it would be greatly valued to have the borrower's basic information pre-filled into the search criteria, to eliminate redundant keying. Based on that, we made the decision to allow the loan officer to start a product search directly from their loan applications list pipeline, by including it in a set of actions for each borrower in their list. This is in addition to starting a new search from within a left navigation menu. The form is also dynamic, and responsive.
Based on user feedback as well as technical and compliancy requirements, I separated the inputs from the results onto separate screens. The input form is dynamic, and the results are surfaced on a grouped, selectable data table with expanding rows to show more detailed fee information, as well as further actions to take on individual loan products.
Input and product search results screen comps.
The Product Selector was successfully launched as an integrated product & pricing feature, and decreased the average time it would take for a loan officer to close on a loan, from several weeks to about 7-10 days. Implementation of the feature contributed to an increase in loan conversions (30% to 100+%) for our customers. This resulted in an increase in new customer contracts as well as extended renewals for existing one.