I am a tree-shaped researcher.

Excellent for reaching a wide audience at scale. I have become wildly fond of branching surveys that incorporate both closed, quantitative measurement and open, qualitative exploration.

At Omada, I sent countless surveys, often using them as a way to both get a gut sense of a topic and then recruit from the respondents for a follow up phone interview to get into more depth.

At Everlane, I designed a survey system to ping either 6 days or 6 weeks after purchase — 6 days to see if the product was as expected, and 6 weeks to see if the product held up to early use. A subset of unhappy responses fed directly into the CX teams queue for proactive outreach to make things right.

To mine customer survey data and map it against transactional behaviors, I’ve learned to love SQL and friendly data analysis tools like Tableau.

The product satisfaction dashboard I created at Everlane married product rating data with returns data — each are measures of customer satisfaction. The resulting “worst offenders” become targets for deeper research followed by prototyping and A/B testing potential solutions.

In-person interviews provide the richest data — they best allow you to compare what people say and what they actually do. They’re also the most costly in terms of time and travel expenses. They are most important when you don’t yet know what is important and/or when context is critical.

To design a new device that could take a photo of printed text and read it aloud for people who are blind, low vision, or dyslexic, we needed to do extensive in-person research to understand the needs, tools, coping strategies, behaviors, and wants of each group ... and to see the text they’d expect the device to be able to read.

Usability testing can happen remotely, but there is a richness from in-person testing, and when your customers are not tech-savvy, remote testing setups can be an intimidating way to start the process. In person you can see the nuances of their gestures, their faces, their almost-actions.

Throughout the design of this product for people who are blind, low vision, and dyslexic, we had five rounds of usability testing, culminating with a full out of the box through advance use test. A young girl with dyslexia breezed through. An older women with macular degeneration struggled but persevered to be able to read her books again. In person, we could see the hiccups we might have missed remotely.

Focus groups have a notoriously bad reputation for a good reason: They’re hard to do well. Good ones balance well designed conversation with individual activities to draw out each person’s thoughts outside of “group think.”

For a mega consumer packaged goods company in the home goods space, I designed a flow that alternated from asking women to (1) quietly imagine their ideal homes, (2) share frustrations in a lively discussion of their current household products, and (3) participate in a co-creation session...

Co-creation gives research participants the opportunity to show us their ideal product, service, or experience by combining prepared design elements. It’s a particularly useful addition to focus groups, but it can be used in any research setting.

In these same focus groups, we asked the women to envision their perfect product with material, product concept, and emotional stimuli. The big insight: they cared more about the LOOK of materials than about the actual material. This was a HUGE win for us in terms of costs, safety, and manufacturing!

Expert interviews, particularly from analogous fields, can lend huge insight into framing new opportunities for solutions ... especially when your customers don’t yet exist.

In a car safety technology project, we were designing a heads up display for consumer cars, but the average consumer has no experience with HUDs. To accelerate our learning, we talked to a race car designer and driver duo. They had tested all manner of HUDs, and they talked confidently about the type of information you want at any given time.

Card sorts help evaluate the relative importance of information or features. The best card sorts include not only priority rating but contextual framing.

For the heads up display project, we asked drivers what safety information they wanted when — the information you want in city vs freeway driving is very different! The big insight: No one actually cares about the speed limit or how fast they’re going — they only care if they are 7mph OVER the speed limit on the freeway or 5mph OVER in the city.

Critical to bring research to life, I tend toward archetypes over personas to help teams focus on strong user patterns that are essential for TK TK.

Archetypes are constructed from key patterns among user groups. Different from marketing demographic segmentation, archetypes are based on behaviors. They’re most helpful for making customers or users tangible to business, design, and engineering stakeholders.

Archteypes were crucial for the project designing a device that needed to serve blind, low vision, and dyslexic people — VERY different groups. We used our archetypes to vet concepts — they had to satisfy both non-tech savvy, low vision retirees and young students with dyslexia who would wouldn’t be caught dead with an “aide.”

The same research yielded design principles — guiding criteria that can frame design and concept evaluation.

The more that the full team can understand and agree on principles, the better chance the product will truly satisfy and delight users.

In the case of this project, our very diverse potential users all wanted a device that was accurate, convenient, and discreet. These principles guided physical and digital design decisions. Each group had more specific core needs in terms of reading activities — these were the basis for the key modes and features.

Customer Journey maps detail the important actions, objects, and people in an experience, shown over time. When you layer on emotion, the customer journey map lets you both pinpoint opportunities to fix problems and increase customer love.

For a genetic sequencing company, Illumina,

we mapped the sequencing flow, based on our interviews, from prep to analysis. They wanted to target a new, naive user, and these flows helped us find the most likely errors. These translated into a new device — MiSeq — with specially designed features based on our insights.