Keller Williams

Most mortgage companies typically serve only a transactional role in providing financing for a buyer to close once they've selected a home. This is a huge missed opportunity, squandering the chance to engage with buyers early in their process and help them consider their goals more holistically.

Given that most buyers tend to index narrowly on interest rate and pre-approval amounts, rather than looking at this complex transaction through the lens of their wider financial situation, they too often experience a string of rejection due to weak offers on unrealistic properties, or face the inability to close, buyer's remorse, and a host of other financial difficulties (moving costs, fees, difficult mortgage burdens, ongoing debt, and so on).

To guide our design problem-solving towards solving this, early on we pivoted our core success metric away from 'lead generation', to, instead, 'qualified leads'. A 'flakey lead' that aborts the process is no good to anyone: but a qualified lead is one who is engaged, invested, clear-minded, and confident in moving ahead. With this metric in place, our design task was about building rich relationships and illuminating lives: helping people achieve the dream of homeownership while, in many cases, dramatically improving their overall finances.

It turned out that, in this case, quality begat quantity. In other words, what is good for buyers is also good for business. "Product-led growth," at significant scale!

Our prototype-driven research made clear that, initially at least, people like to remain anonymous, and are very wary of being hounded and spammed if they provide any personal information. Given this, user-preference was strongly for a 'sandbox' version of the Total Cost Analysis tool, allowing them to kick the tires and do some basic scenario modelling themselves before speaking with an expert. When asked, they were generally willing to accept a margin of error of up to 30% in this, with the tool still being very highly valued, knowing that an expert would be able to straighten things out when they were ready.

Amazingly, even the MVP version (shown above) — effectively a static 'ad' for the concept, with minimal accordion-based interactivity — proved game-changing in helping prospective buyers transition away from a 'price' mindset (indexing narrowly on interest rates) towards an 'advice' mindset (a broadening awareness of the options, and an appetite for guidance). This transition represents a key milestone in qualified lead-conversion, being both very positive for the buyer's now more-informed path, as well as reflecting very positively on Keller Mortgage as a brand, and setting the Agent and Loan Officer up for collaborative next steps.

Beyond our core strategy of helping prospective buyers feel "Informed, empowered, successful" (wording which I came up with during a design critique, but which ultimately became the public facing slogan of Keller Mortgage), we explored design interventions at a more behavioral level, represented by the green lines in the diagram below.

Mapping prospective buyers' typical journeys and segmenting them by behavioral modalities such as 'Help Me Do It' and 'Let Me Do It', 'Experienced' and 'Inexperienced', we were able to identify moments to explore behavioral nudges and ideate around these via a Lean UX process of asking what we needed to learn, sketching what we needed in order to learn it, and then prototyping to evaluate these new design ideas with real users.

While user flows were non-linear and widely divergent, reflecting the massive variety of people looking for a mortgage, our design 'interventions' represented a distributed system of parts working collectively to reinforce messaging and provide multiple 'paths' to the same destination.

Tooltips on core information provides a progressively-revealed, well-paced means of education; a 'call and response' with the user.

Such an approach is almost diametrically opposed to the more common tendencies of aggressively 'encouraging' a user to 'Sign up!' and 'Apply Now!', coupled with (at best) passive chunks of long-form educational content. While perhaps seeming a subtle difference, in testing sessions with a widely diversified group of participants, such experiential differences were clearly recognized and appreciated.