The Setup

Traditional Life Insurance is a business built on data. Insurance companies collect data about a client, build a risk profile, and assign a price to that person based on how likely they are to die.

The problem with this is that the process of collecting this data can be very time consuming, lasting anywhere from 1 to 5 months. Unsurprisingly, this can result in major pain on the client experience side, waiting for months to get their product.

‍

‍

While most traditional companies have leaned into the “difficult but cheap” model of Life Insurance, a recent trend has seen companies going the other way toward “easy but expensive” with an all-digital life insurance process.

‍

‍

This process is a significant speed boost, taking 30 minutes to 2 weeks to complete. The pitfall— company can’t collect the kind of invasive data (such as an in-person medical exam), thus leading a less complete risk profile. The company must then protect itself and assume high risk in the voids left by these important data points which result in significantly higher rates.

‍

‍

Health IQ, seeing the customer pain & wanting to learn more about this burgeoning market, decided it was time for its first all-digital experience — Health IQ Express. The following is the process and current status of that project.

I believe deep product knowledge is essential to building the best solutions in any product. To illustrate this knowledge, I’ve written up a more extensive intro to the strategic and business setup, where you can find all you never wanted to know about Life Insurance.

‍

The Process

Goals & Metrics

At the outset, we defined several goals for the process, which was distinct by stakeholder and perspective.

For the User

• Looking for fast & easy
• Target user: healthy, late 20s/early 30s so low likelihood of manifesting medical conditions = fastest time-to-approval & better rates
• Mid to Upper Class and above — steady income, more willing to pay more for ease

For the Business

• Spin up a new channel
• Gain competitive intel on niche industry
• Support mission — new way to reward the health conscious

For the Carrier

• Vested interest in mitigating risk
• Keep client pool high quality, low risk

For the Designer

• Weave together stakeholder perspectives into a human experience
• Inject empathy, elevate moments
• Keep both business goals and user goals in your sights

For success metrics, part of the process involved figuring out what success meant, especially as we now were looking at controlling the full funnel within this condensed time window and digital space.

• For the Marketing portion — we want to test if our initial impression/offering is an attractive one and gets visitors to continue past the 1st page. We also want to see if the initial, qualifying lead questions results in Drop Off. Then we want to see how many of those people ultimately continue to the Application step
• For the Application portion–this is a lot of questions, so we’re interested in where and why we have drop off, how we can move the ultimate Sale Rate at the end, and finally getting people to submit their payment and accept the offer.
• On the Approval Type — we also need to understand the client pool we’re bringing in and subset our data. Ideally, they are perfectly healthy and truthful, leading to an Instant Approval — a lot of those means a job well done and high-fives all around. General Approvals are those who aren’t outright rejected but are referred to further underwriting and still receive an offer.

Research

Existing Patterns

‍

My next step in approaching a problem is to research existing patterns. This means doing deep dives on competitors (like Ethos and Ladder Life), looking at our own design learnings that can inform this project (like our Website and Landing Pages), and other related products where we can infer some lessons and take inspiration (such as TurboTax — a complex, all-digital financial experience).

We took a number ideas forward to test, but generally we found that all these products have the same goal — simplifying the complex. And that comes across in the UI & design, which is kept largely stark, simple and straightforward. Especially when the process is long and complicated, you don’t need bells and whistles to make it even more dense.

Another key finding was seeing how companies have gotten creative with pricing and quoting. In our research we came across creative approaches and tests which indicated to us that this was an important part of the process.

Another interesting thing we found in our digital competitors was the “Honesty Pledge” — a screen which basically just says “hey, we operate on the honor system, you’re not gonna lie right?” Our Legal team said there was nothing binding or required about it, but it stuck in our brains and influenced later iterations of the product.

‍

Agent & Ops Team Interviews

‍

‍

The next part of my research process is getting time with the experts. In this case, this involved sitting with Sales Agents and Ops team members to absorb their knowledge and get their deep experience coaching people through the life insurance process.

Take Aways:

• Quote-time is key — you have to strike a difficult balance between appealing and realistic. When Marketing is out advertising too aggressively on an unrealistic quote, it can make the Sales discussion much more difficult.
• Carrier Name vs. Price — While some clients are all about price, other fixate on company name recognition
• Family/personal health history questions — people lie, they get offended/defensive
• Questions about their lifestyle choices/travel — people don’t understand how their risky activities/travel can get them rated. Sometimes application questions will ask about “intended travel” which will trip them up — they’ll overshare and say “idk I guess one day,” you have to explain to them that it’s only what’s booked — if you don’t have that flight you’d one day like to take to the to the Middle East booked then you don’t need to report it. It’s only going to raise unnecessary red flags further down the line.
• Height/weight — this one is important for the quote because of BMI, but a lot of people lie on it. It’s not always because they’re trying to be deceptive, it’s that most people don’t frequently measure or weigh themselves and genuinely don’t know and will just give you their last known… which can be a while ago.
• Doctor’s information — frequent stalling point because people rarely have this info on hand or remember.
• Payment — it always goes better if you collect payment in advance — saving it for post-signature decreases Accept Rates.

‍

Carrier Discussions/Restrictions

• Application — there are certain portions of the application that must be word-for-word
• Coverage Cap — because of the higher risk involved and the new, unproven partnership, we were capped on coverage amount — an unfortunate feature given our success finding lower risk, higher income clients on the traditional term life side of the business.
• Make it harder — they rang this bell many-a-time. As the financially responsible party, they want more data points, more restrictive targeting, more scary and legally all-encompassing language.

‍

Initial MVP Launch

All these learnings were done simultaneously with the implementation of an MVP. The MVP was done with an older digital partner, never meant to be a forever solution due to the constraints that working with an existing model brought. It didn’t give us the flexibility or the innovativeness that we sought.

The MVP launch consisted of a pretty sparse initial offering, as an MVP should — no bells and whistles, just putting out there something to collect some first-hand intel. It went as expected—not great — but gave us the opportunity to dry run and understand the kind of things we should be looking at.

‍

‍

It was exciting to see and be largely in control of the full lead-to-sale funnel for once. We got to see question by question drop-off along the funnel and validate what questions were the problem spots. Mostly, it was as we had predicted — drop-off on the 1st page, the quote page, the SSN, medical conditions, etc.

First Test

The first boiler plate test we had lined up at the starting gate was to see if the learnings that we’d taken from the course of optimizing our multistep landing page experience could improve the digital product experience as well. This included things like using a card format to focus users, security/trust elements throughout, etc.

‍

‍

The result was interesting. Unsurprisingly, it improved initial drop-off during the marketing portion, but that improved conversion dwindled to 0% change on ultimate conversion despite a relatively large effect early in the funnel.

Sadly, this was our first and last test on the MVP as the product was shuttered not long after. The carrier backing the product was working with a competitor at a much higher volume. We learned that this carrier was giving them a critical mass of high risk clients and they were hoping that our health-focused marketing would help offset the risk and save the product. Alas, we were too late. With a 0.5% conversion rate (14.5% lower conversion rate than we were were lead to expect…) and a cautious initial marketing spend was never going to cut it.

Health IQ Digital 2.0

Undeterred by the initial failure, we forged on to a larger project — building and filing our own product from the ground up. This would give us control over the application wording, allow us to insert newer technology, and introduce elements such as Lab Check & the Health IQ Quiz which can drive down price.

Control App Questions

I inserted myself early on in the meetings with actuaries while building the application we would file. Decisions made in that room were going to have significant ramifications and limitations down the line. Together, we would pick through application questions, break down every word.

‍

It was great being involved at that stage, because we were going to be legally bound to many of those questions and their instinct was toward the most secure legal-ese that they could get. My job was to encourage flexibility to allow us to test later down the line, calling for early hypothesizing on potential future tests. For example, we worked on multiple wordings and styling options for each question to vary tone, format, and length.

Leverage Data & Tech for Better UX & Pricing

Another key benefit of getting to build a more modern, built-for-digital application was being able to introduce modern technology early on. For example, including features like the ability to scan your drivers license or credit card with your phone. Also the ability to include modern payment methods like Apple Pay.

‍

‍

These kinds of introductions were key for 2 primary reasons. First, it was important to us to leverage every possible technology to potentially improve the UX of the product. Second, it was marketable to insurance partners and re-insurers who help host and financially back our product. This meant design could have a direct influence on price, which was a very empowering position to be in.

More Ways to Share Data, Better Prices

The most pivotal part of our new product were the Lab and Script Check features.

LabCheck is an integration with labs like QuestDiagnostic which allowed us to use a client’s information to search for past blood and urine test results. If a client had one within the last 2 years, we could use that as a stand-in for their Medical Exam which would allow us to drastically reduce price.

ScriptCheck is very similar, allowing us to pull information from past prescriptions.

‍

‍

Both Lab and ScriptCheck gave us the ability to gain more data about a client to get a more complete risk assessment, driving down the all-digital price to something much closer to an in-person price.

We also found more integrations that would allow us to collect more information, since Lab and ScriptChecks were only going to cover about 60% of the population. So we worked on additional integrations that would allow users to share information directly from their healthcare provider or even from their health/fitness apps.

‍

‍

But we didn’t want to only use this information in the background to drive down price. We wanted to leverage this data to make hard questions easier. For example, for legal purposes the medical questions could be quite dense — asking “have you had any of these” and listing very long lists of medical conditions. We wanted to leverage our background data and look at their medications and procedures to try to recommend answers, so a quick skim is all they need to find the conditions that pertain to them. And, if they were thinking about lying well, they get the sense the jig is up.

‍

‍

We also wanted to get creative with the ScriptCheck data and use it to ease another question we know can be difficult or time consuming — the doctor lookup. We built in search technology but, using the physician name from prescription data, we could offer a UI that allows the user to more easily find and select their doctors.

Improving the User Experience

Next, we were interested in improving the User Experience in a number of ways. First, we wanted to introduce a Virtual Agent and Agent Tips, to try to re-introduce a human element and bridge the gap over an important piece that’s lost with the over-the-phone experience.

‍

‍

Using the tips we learned from the agents, we had identified key points in the process where clarification or encouragement would be necessary. Now, the Virtual Agent could pop in to offer guidance, just like the might on the phone.

It’s key to introduce the Health IQ Quiz which is so central to our core business. While this likely won’t be a part of the initial launch of v2, we’re keeping it in our sights in service of the mission

Further more, we know that quotes is a key part of the process, and for do-it-yourself-ers, there can be a lot of ambiguity in this coverage-term-price decision. To help, we want to get creative with how we present quote — making the decision making by automatically suggesting 3 tiered packages based on a price rather than coverage and using Virtual Agent tips to offer helpful tidbits

‍

‍

Finally, it’s also important to me to try to elevate moments, especially with “celebration” being such an integral part of Health IQ’s mission. Sometimes we just need to spend the additional time and effort adding moments of celebration and encouragement, even if it appears frivolous or cheesy on the surface, it can be the difference between a bland and forgettable experience and one that you walk away feeling good about.

Employ Behavioral Economic Theory & Psychology

Given my background in Psychology, I was excited by this one. I got to exercise old brain muscles and huddle with Behavioral Economists to think even deeper about the way we presented material to elicit the most truth. This is important as lying is much easier on a digital product, so assuaging our financial backers worries by leveraging the power of Psychological tweaks was a powerful addition to the product.

One way we approached this was using a “please verify” model. With all the background information we were collecting, there were questions we already had the answers to and, but strategically placing those “verify” questions right before important questions we don’t know the answer to, we hypothesized we could increase the veracity of these answers.

‍

‍

In the example above, Height/Weight are a key factor, so we wanted to increase our confidence in these. This is hard to test in the wild, so I hope to one day be able to run a clinical psych experiment to test this hypothesis in which we can verify answer vs. reality among participants.

The Virtual Agent also played into this approach. We hypothesized that a human face in the experience, even though users obviously know it’s not real, would make it more engaging and difficult to outright lie.

We also applied a tried and true concept of Behavioral Economics — anchoring. This means setting the bar excessively low so that even negative but more truthful answers come more freely. For example, building a UI for the question of weight that allows us to prefill with a weight that is very high for their height. This way, even if they’re wait is high, a user thinks “Oh, I’m heavy, but I’m not that heavy,” and puts a more accurate weight vs. giving themselves the benefit of a little weight trimming on their answers. This was central to posing two critical questions: weight and nicotine use.

‍

Conclusion

Sadly, this project was shelved while the entire company put the petal to the metal to launch Medicare in time for Annual Enrollment 2020. I hope to take it up again soon. First, because I put in a lot of work here and am itching to see how it does. Second, because this is one of my favorite projects I’ve ever worked on, for several reasons very important to me:

• Exemplifies the value of the work I did to acquire deep product knowledge
• It emphasized a high-tech approach, leveraging interesting technologies to aid in user experience
• It heavily values and relies upon behavioral psychology knowledge which, second to design, is my other passion
• Finally, getting to control so many pieces allowed me to introduce human elements like moments of celebration which, especially in Silicon Valley, is an important facet that designers should look to advocate for.

‍

‍