From brain science to quantitative UX
Career lessons and observations for starting off in the UX field
This July, I celebrated four years working at Pinterest. Before transitioning to a career in quantitative user experience (quant UX) research, I had been doing neuroscience research in academia for over a decade. In my current work, I help make data-driven, user-centered product decisions that improve end-user experience on Pinterest. This means I survey people to understand their motivation, comprehension, and feelings about using Pinterest. I query log data to understand product usage patterns, run analysis, build statistical models and correlate user sentiments (like “my home feed is confusing”) with their on-site behavior (like if they saved any Pins while browsing their home feed).
As a neuroscientist, it took my colleagues and I four years to publish one paper in Nature or Neuron. At Pinterest, research projects and decision-making are happening at a much faster pace. If you’re also looking to try out a career in quant UX (or have recently made the switch), here are five big lessons I’ve learned so far.
Lesson 1: Choose the right research problem to solve
In science, research motivation can sometimes be driven by pure curiosity — the pursuit of knowledge.
In tech, not so much. At Pinterest, quant UXRs are usually staffed across horizontal teams, which means you’ll likely receive multiple requests from different teams with competing priorities. It’s critical for a researcher to determine “what’s important to know” over “what’s interesting to know,” and prioritize projects accordingly in order to maximize business impact while maintaining a healthy work-life balance.
I consider two criteria when assessing whether to take on a research project:
1) Will the research solve a critical business problem?
2) Will I learn a new technical skill set, build a new relationship, or learn about a new product area?
A good research project addresses a critical business need and furthers your career growth.
When you’re looking to understand if your research is business-critical (remember, this isn’t the academic science lab anymore), you can start by asking the person who requested the research:
- What do you plan to do with my research findings?
- What’s the timeline?
- Will the research help solve a specific user’s needs?
- Will it help make a short-term product decision, or set the foundation that guides product development for years?
Asking these questions will help you assess the importance and urgency of the research request, and prioritize it accordingly.
Lesson 2: Let the research question determine the methodology — and not the other way around
As a neuroscientist, my research focused on understanding what happens in our brain during rapid eye movements — a model system to understand the neural basis of human motor decision making. We made sure our experimental design used the right methodologies: Scientific rigor means researchers not only have to deeply understand different research methods but, more critically, know when to choose an appropriate method to solve a problem. This is also true for quant UX.
I once had a product manager ask me to run a tracking survey on a new product — that’s where you measure user sentiment over time to assess a product’s strength or weakness. The issue with this tracking survey was that the PM requested the research method without a specific question in mind — he just wanted to know how much people liked his new product (or not). But without an existing baseline about user sentiment or knowledge about users’ past behaviors, how would we know what questions to include in a tracking survey, or what success and failure would look like?
Yup: I was asked to research with a specific method, instead of letting the type of research project dictate the method. So instead of starting with the tracking survey as prescribed, I piloted a 10-question targeted survey to understand not only how users feel and perceive the new product, but also why. We combined the survey data with a behavioral log analysis to understand what user actions could contribute to these sentiments. For example, could the number of Pins created, searched, or saved contribute to how respondents felt?
After extracting the learning from the targeted survey, we trimmed the number of questions down to two and implemented them in the tracker. We also collected the survey data in parallel with the team’s core behavioral metrics, so we can measure both the sentiment and usage change over time. Had we implemented just the tracking study instead of adding a targeted survey or behavioral analysis beforehand, we might have wasted a few months of data collection and still have little understanding of how people truly think of the new product.
Lesson 3: Build a relationship with engineers before asking them to build anything
When I was doing neuroscience research, we established an in-scanner eye-tracking system to simultaneously collect participants’ eye movements data and brain signals. My advisor (and our research institute!) was supportive, and our lab had the funding to pay for the equipment and setup. Although there were other hurdles, the team had a shared vision that made it easier to overcome our obstacles.
At Pinterest, quant UX work often requires collaborating with engineers in order to get things done (sometimes, it means changing production code) — they’re like the technicians that built the lab set-up in our scientific research days. But naturally, there’s an engineering resource constraint. Engineers are busy coding up new product features or investigating bug issues. Your priority may not be their priority.
In order to make progress, you need to build relationships with your engineering stakeholders. (Just like you would with any other cross-functional partner!) Tell engineers about your role, the nature of your work, and discuss how you’d like to work together. This way, when it comes time to make an official request, you will already have laid the groundwork to get the eng support more easily and quickly.
When you need specific engineering support, include a brief description of the research project and what you’re trying to achieve when you reach out. Don’t assume they’ll be able to help; instead, build a mutual understanding of how their work and your research both contribute to reaching the broader team or company goals. When you make an effort to understand their priorities and constraints and offer help to unblock them with your research, you’ll increase the likeliness of quick eng support and team buy-in.
(And don’t forget to show your sincere appreciation for their help afterward. Everyone loves a thank-you note.)
Lesson 4: Collaborate with other quantitative and qualitative experts
Good quant UX work doesn’t exist in isolation but, rather, lies in its connection with other disciplines and methodologies. As a neuroscientist, I combined eye movements data from both humans and primates, and whole-brain imaging data with single neuron’s action potential data to investigate what happens in our brain when making a motoric decision; similarly, quant UX data is just part of telling a bigger-picture story.
On the Pinterest Research team, I collaborate with market researchers and qualitative researchers. Outside of the Research team, I also work with product analysts and data scientists. Each discipline has its own strength and weakness, but drawing data from multiple sources and across methodologies helps tell a compelling and convincing story to our cross-functional partners.
For example:
- I might partner with a market researcher to learn how people perceive the Pinterest brand, against competitors, in the US and international markets.
- I might add the stories, quotes, and emotional themes a qualitative researcher shares about new Pinterest users to numbers and charts from my quant report, to help designers and engineers get a deeper understanding of people using Pinterest for the first time.
- Lastly, I might partner with a data scientist to evaluate how people think of the quality of the content that was recommended to them, in addition to the data scientist’s analysis that showed how people interact with the content.
Lesson 5: Use data to tell a story…instead of telling a story about your data
Being able to tell a compelling story using data is the most important skill I’m still learning as a researcher — both in academia and at Pinterest.
When developing your story, think about your audience first. Structure your story around the major themes (three is a good number) you’re trying to convey, and then tailor it to their needs and interests. Is your primary audience made of coworkers with a quantitative background, like data scientists and analysts, who are very interested in learning your technical methods? If so, you might want to spend time upfront to explain the steps you took to analyze the data. Are they product managers, designers, or engineers who are mostly interested in actionable recommendations? If so, you might want to spend more time interpreting the research finding and its significance than explaining how to read each chart. Craft your share-out for each audience, and edit mercilessly. When you have a diverse group of audience, tailor the story to your primary audience. Don’t be obsessed with every detail (this isn’t scientific academia, after all). Just because you analyzed data doesn’t mean you need to show it in your presentation. There’s an appendix for a reason.
If you’re having trouble identifying the most important information for compelling storytelling, I find it helpful to print out all your worksheets and stick them to a blackboard (this can also be done digitally). This way, you’ll have a bird’s-eye view of the entire analysis, and can move the print-outs around to organize the information in a more story-friendly way.
To close up…
It’s been 4 years since I switched to quant UX research so that I could learn and grow in a fast-paced environment like the technology industry, and apply the skills learned from doing scientific research into solving real-world problems. I loved doing neuroscience research, just as I love quant UX: A researcher is always learning and discovering. After four years and five big lessons, I’m still finding new ways to solve ambiguous business problems, take on more mentorship responsibilities, and increase the visibility of quant UX so more people can see the impact that quantitative UX researchers can have in their organizations.
And if you’ve got more learnings to share, I’m all ears.
Illustrated by: Altay Sendil, Head of Research @Pinterest; Special thanks to Ximena Vengoechea and Molly Marriner for editing.
