Designing for trust

The challenges of building a health care product

Creating software always means asking for trust. Tap a button, we promise it’ll take you where you want to go. Sign up for an account, we promise to respect your privacy. There are a million ways a user interface can solidify the relationship between a product and the user — and a million more ways it can break it.

At Amino, we’re trying to help people outsmart health care by providing cost estimates and experience analyses for nearly every doctor in America. Before we launched, our data on doctors and procedures had never been available to consumers. All this new information makes Amino a better way to find a doctor than the usual approach of asking friends or searching Yelp reviews.

But we face 3 pretty serious challenges to success with users:

  1. Health care is notorious for pay-to-play promos, cold and confusing interfaces, and slippery language.
  2. It’s hard to convince people to change the way they’ve been looking for doctors their whole lives.
  3. We’re introducing brand new decision-making criteria for choosing which doctor to see.

Building trust with our users is essential if we’re going to have any chance of tackling these challenges. There are many ways to foster trust, but design is one of the tools we can tune the most. And we try to learn as much as we can through user testing to make sure our users feel comfortable taking these leaps with us.

Amino’s doctor search has been live for about a year, and we’ve been able to test assumptions, iterate, and introduce new ideas into the product. We’re a small design team — 4 designers, 3 front-end engineers — so we often move fast and fix later. But user feedback is always part of our process because we need to make sure we’re continually building a positive relationship with our users. We’re still adjusting our tactics with every new feature, but along the way we’ve learned a few lessons about building trust through design as we tackled these challenges.

Challenge #1: Facing health care problems can be unpleasant. Amino should feel as uplifting as possible.

Health care, like car tires and renter’s insurance, is a “grudge purchase.” People have to go to the doctor, but they’re not happy about it. In our industry, we can’t rely on the excitement that people bring to a shopping site when they’re looking for a new gadget, or a travel site when searching for the best vacation. People would rather avoid the hassle of finding a new doctor: the endless phone tag, the awkward first visits, and the incomprehensible medical bills. To help our users take the first step when they’d rather procrastinate, Amino’s design needs to stir up positive feelings and associations.

A lot of health sites use photos of doctors’ offices or people in pain to communicate the ways they can help. For us, it’s more important to let users know how using Amino will make them feel — light, confident, and optimistic about what comes next. That’s why we chose photos of kids and penguins for our homepage. These images are full of energy and forward momentum. They set the tone for Amino as something hopeful and, most importantly, different than the way they’ve previously experienced health care.

In our branding and our UI, we embrace warm, bright reds, purples and greens — colors and gradients you’re more likely to see on a culture site than in your insurance portal. Again, we’re trying to signal that we think about health care differently. In user testing, we’re thrilled when we hear words like “beautiful” and “a breath of fresh air” used to describe our interface. Right now, we’re working towards a more unified visual style across the site, but we want to make sure that we keep those voices in mind when we’re revisiting color and type.

Language is another minefield when working in health care. Specialties and conditions can be complicated phrases, and you’d need a PhD in health benefits administration to understand all the nuances of how your plan works. We have a Slack channel at Amino devoted to the process of choosing the right language to use in our product and messaging. It’s called, aptly, “Words Are Hard.”

Words will always be hard, but we’ve come up with a few guidelines in our quest for a great relationship with our users. We use the accepted medical terms when consistency is important — procedures, conditions, specialties — but otherwise, we use simple, positive words and personalize when we can. We also lean on action words throughout the experience — “compare network rates,” and “calculate what you’ll pay” — to remind users that they are in control. Our goal is to communicate that our users are in the driver’s seat, but they can trust us to guide them through the doctor-finding process.

Challenge #2: Changing behavior is an uphill climb. Let’s start with something familiar.

According to our research, the most common way people find a doctor right now is through friends and family. That’s a personal process based on real-life relationships. It’s hard to get people to break old habits and try something new. Marketing and partnerships can help pave the way, but on the product design team, we’ve found that adopting elements from consumer products in other industries helps ground Amino’s experience in friendly and familiar patterns.

Maps are a large part of our user experience at Amino and one way of bringing in patterns from other consumer products. Putting results in the context of a map makes searching for health care feel closer to searching for a restaurant on Yelp or Foursquare — a fun and familiar process. And we don’t have to cue users very much for them to know that they can navigate their options by tapping pins, or zoom the map in and out. In user testing, we’ve gotten very positive feedback about our use of maps, but users want even more of the functionality that they’re used to in other products, like automatically redoing a search when the map moves. So we’re working on building some of those features.

Our step-by-step search flow — what we call “the wizard” — was one of our first trust wins. We originally had a single form to collect a user’s info before displaying their results. One page, no clicking around. But in trying out the flow ourselves, we felt overwhelmed by the number of questions. We toyed with just asking for a user’s location and then letting them filter a long list. But we really wanted to communicate that our results are tailored to each user: it’s a short list they can trust.

We took inspiration from TurboTax, which asks the user a series of simple questions, and then delivers big time. When we’d used that product in the past, we were happy to answer as many questions as necessary to feed the wizard — as long as it took care of our taxes for us. So we tested our own wizard with users, a little nervous that people would get impatient and drop off after the first or second step. It was a big surprise when we got consistent and resounding feedback that users loved it.

The card flow felt more like a fun quiz than a diagnosis, and gave us the space and time to explain why we needed the info. Some users even suggested that we add more cards to the flow to get even more personalized results. It turned out that the friction of multiple cards and moments to opt-in gave them more confidence in the results that came afterwards.

Challenge #3: New data can be overwhelming. We’ll guide you towards what you need to know right now.

Too much new information is hard to process, but not enough can lead a user to feel like something’s being hidden. A trusting relationship is dependent on progressive disclosure and smart synthesis. We spend a lot of time trying to pull out the what’s and why’s of insurance and health care costs. It’s tough stuff, and involves writing, data visualization, type hierarchy, and many other tools. We’ve arrived at an approach that keeps us from going in circles: we try to make a thoughtful and user-focused decision about what to show at the top level, but we always allow the user to find out more if they’re interested.

We go to great lengths to provide a way to find helpful details whenever the user encounters a new concept. It could be in the form of a tool tip, a lengthy explainer in a modal, or a link to our methodology pages with a deep dive into our process. In testing, users really like that these options exist and read them more than we expected. But even users who don’t read through to the end are relieved that the answer can be found somewhere. In our experience, having the information available makes users feel safer — the right mindset for smart decision-making.

One area that has really tested our design mettle is our cost product. Before we launched procedure cost estimates on Amino, we knew that many people had never seen doctor-level cost data like this before. It was hard to know whether they would understand how to use it to make health care decisions. Our challenge with cost estimates on Amino has been to provide the data, but also offer education around what it means. That’s tough when the cost for any single health care experience is driven by a mind-boggling array of factors: the plan you’re on, where you live, the doctor you choose, the time of year, your personal health history, and much more.

Amino’s cost estimates are “network rates,” the amount that you and your health insurance might pay for a procedure. When we launched our cost product, we worried that users would think that this is the amount they’re personally responsible for. So everywhere that network rates appear, we’ve tried to clearly direct users to our out-of-pocket cost calculator to help them understand what portion they’ll pay. But since insurance plan details and deductible status can affect out-of-pocket costs — and we don’t have access to either one right now — the number we provide is only an estimate.

Unfortunately, user testing has revealed that we need to do a much better job; users often think that network rates refer to what they’ll pay out of pocket, or assume that our calculator is much “smarter” than it really is. To make things a little trickier, we’ve found that users really like the calculator. They just don’t know how to interpret the number it gives them. Improving users’ understanding requires us to educate them about even more convoluted concepts than what’s currently in the product — and there are many trust pitfalls along the way. We’re working to address our users’ misconceptions, but improving this particular feature is not for the faint of heart.

User testing is the key to moving forward.

As we continue to build new features and improve our product, we’re working hard to deepen the trust we’ve established with our users so far. But health care consumers will only get savvier, and our challenges are sure to change. As more Americans move to high deductible insurance plans, users may demand more detail on cost estimates, new information about health care facilities, or features like comparing doctors side-by-side.

User testing is key to keeping in touch with those changing user needs. Right now, we use consumer surveys to gauge sentiment on concepts, for in-depth user testing, and some A/B testing on our live site. But part of the challenge of testing is knowing what’s signal and what’s noise. As we’re developing our product, we’re also working on that process of listening to users and translating their feedback into features and improvements.

Maintaining a trusting relationship with our users — and with anyone who encounters our product or brand — means always being open to feedback. You can let us know what you think at

Thanks to Sumul Shah, Maudie Shah, Carrie Phillips, and Hannah Levy for contributing to this piece.

This piece was originally published on Amino’s blog on September 21, 2016.