Buying your first property should be an exciting adventure, but all too often it’s plagued with stress and anxiety, thanks to an old-fashioned ways of doing things.
By creating an ecosystem of tools and services that engage property buyers, sellers, estate agents, brokers and conveyancers, OneDome brings these various separate journeys together into one larger and actually exciting customer adventure.
My brief was to help fulfil this business strategy by creating the final piece of the ecosystem: the online mortgage application service. This would be a brand new end-to-end mortgage application experience to be embedded into their wider platform.
I was hired as a full-stack designer, to undertake all design tasks for the mortgage application service, from research, to prototyping, to final UI design and delivery.
The product owner and I made an excellent team. We worked closely together throughout the project, always ensuring the interests of the users, management and other internal teams were represented appropriately.
How we got there
The lay of the land
It was clear that we had two primary actors that would be interacting with each other: the buyers and the brokers.
It had already been established that for MVP OneDome would hire one initial broker who would operate the broker’s side of the application service. This broker would guide the applicants through the process, receive and process their information, and issue them with documentation at key stages.
To keep things simple for the MVP, we decided to focus on first-time buyers. This was done by looking at the value of the market segments, and the added complexity that would be required to cater to other types of buyers. Our target market:
- Existing OneDome customers emerging from the Estate Agent journey (more on that later)
- New customers coming in from search and ads
All actors need a stage, and our stage was the OneDome ecosystem. This consisted of several existing property-related services and journeys, and we knew that we’d need to map these all out clearly so that we could understand how to create a new organism within it.
We broke the project down into the following phases:
One of my favourite things to do is create a space on a wall somewhere visible to the rest of the company so that anyone can get an idea of what’s going on. It’s amazing how many useful and enjoyable conversations get started this way with unexpected people that the project might not have touched otherwise.
Let's get down to it
The first part of the process would be establishing the constraints on the brokers’ side of things. This essentially boiled down to financial tools and regulations. The cold hard reality was that we would not be able to step outside these constraints, so we decided that we needed to get to know them well before we even looked at buyers.
Examining broker tools
We contacted several brokers and asked them to take us through the entire process for handling a customer’s mortgage application. This gave us insight into:
- What tools were used
- Which questions were required
- How to work around various obstacles
- The challenges that the brokers had with obtaining certain pieces of information from applicants.
This was our first major constraint: brokers would need answers to about 50 specific questions, in a specific order, in specific stages, and in specific formats.
Before dealing with those 50 broker questions, we wanted to find out more about the pain-points of the property-buying journey. To do this we recruited a number of people who had bought their first property in the last 12 months, and we conducted structured interviews with them.
- Lack of choice
- Pushy sales pitches
- Feeling like a commodity
- Disconnection from the process
- Lack of transparency and visibility
Armed with this knowledge, we already had ideas about how we could create a much more trustworthy and efficient experience.
Questions about questions (about questions)
Now, about those 50 questions. We needed to find out:
- Which questions were actually required
- What people felt and thought about the questions
- How to identify and handle sensitive or difficult questions
- How to group and sequence the questions in the best way
- How to support the customer as much as possible with each question
- How to incentivise the customer along this journey
Study 1: How does this make you feel?
We recruited 6 carefully screened participants and asked them to participate in an exercise designed to uncover their thoughts and feelings about the mortgage application questions.
Each participant sorted through the questions, placing each one into one of three groups: happy, medium and sad.
For background music, we put on How Does It Make You Feel from Air's 2001 album 10 000 Hz Legend 🎶.
We spent the second half of the exercise digging into why the participant was unhappy with the question. This led to all kinds of explanations - sometimes the problem is just semantic, other times the question is suspiciously irrelevant, and in the worst cases it is just too personal.
At the end of this we knew which questions the customers hated (yes, actually hated), which ones should be dropped if possible, which ones needed to be rephrased and which ones needed extra supporting content, like definitions or justifications.
Study 2: Answer Time
Not knowing the answer to a question is a critical roadblock in the application process—it can lead the user to submit incorrect information or to abandon the application entirely.
We asked each participant to review all of the questions and to point out which ones they couldn’t answer. For each of these tricky questions we asked the participant:
- What would they need to do to find the answer?
- How long would it take?
This was very useful, and led to several features to help users when they got stuck, such as being able to set a reminder to come back to the application the following day.
Study 3: Classic Sort
Now that we’d explored the emotional dimensions of the questions, we could turn to the more technical task of figuring out how to group and order the questions. For this we could turn to the classic open card sort, a more straightforward exercise.
We got the same participants to separate the cards into any number of groups and then give each group a name. The results of the exercise showed a clear general pattern, and gave us confidence in knowing how users would expect the questions to be grouped.
This would turn out to be critically useful, because the MIDAS tool did not group the questions in the same way.
Truth and trust
Zooming out a bit, we saw some key human themes emerging from these studies: truth and trust.
We saw that the participants were almost always asking “why do they need this information?”. Combine that angst with concerns about overt privacy invasion, and you have a recipe for a terrible customer experience.
At worst you lose the customer and damage your brand, and at best you receive untruthful answers that lead to inaccurate calculations and inappropriate financial products being offered.
We had several participants tell us that they would lie to try to get a better offer.
This isn’t an exaggeration—we had several participants tell us that they would lie to try to get a better offer. It boiled down to this: If we could earn the customers trust, we could convert more users and increase the quality and appropriateness of the financial products being offered.
Mapping the service ecosystem
We needed to design the new mortgage service so that it would fit into the OneDome ecosystem, so we needed to ensure that customers would be able to move in and out of the various journeys in a way that made sense to them and to the business.
We started by interviewing the business leads of each of the ecosystem services to find out what their goals and strategies were, as well as their priorities and challenges.
For each service we completed thorough journey maps. These showed all touchpoints, entry and exit points, and how users would move out of one journey and into another. Using this information we could increase business value by ensuring that customers we funneled into the right places at the right times.
The audits also showed the flow of data through the journeys. This was especially important because knowing what data was available at which points, we could create features and UX enhancements to improve the effectiveness of the user experience.
And finally, we can prototype!
Another one of my favourite things to do is co-design. Grab a whiteboard and a bunch of interested people and just start drawing. It’s amazing what the collective can come up with. This is a part of design thinking.
Question flows and UI logic
The journey that an applicant takes through the application process has many forks along the way. Some answers lead to more questions, others let you skip past irrelevant sections. We mapped out every path and presented it in a way that the developers could easily translate into code.
From initial sketches to final UI
I usually start off with a set of initial sketches. These don’t take long and can be tested within the organisation and with recruited participants easily and cheaply. This is the start of the design iteration cycle. Evaluation is usually done using ordinary user testing techniques, such as the Think Aloud protocol.
Once I've settled on a direction we can clarify it further with more a detailed prototype:
With more user testing we can increase our confidence in the designs and starting bringing in the brand and adding a much more visual expression to the design:
I ❤️ mortgages!
This project was a great success with OneDome receiving many more leads than expected, so the issue of scaling the service became a pressing one immediately. Since then OneDome have scaled up the service largely thanks to a subseqent partnership with Mortgage Advice Bureau.
Customer feedback was very positive, with one customer famously shouting "I love mortgages!" to the broker when their mortgage was issued.