How I wound up being a Quantitative UX Researcher

Not exactly Data Scientist, but close enough you can barely tell the difference. It’s a fairly obscure position. This is, in broad strokes…

Not exactly Data Scientist, but close enough you can barely tell the difference. It’s a fairly obscure position. This is, in broad strokes, how I wound up where I am in the data world. Current as of Q1 2019, no idea what the future will bring.

Out of the myriad ways of coming to wear a “data scientist” hat, I firmly came down the “data analyst” side of things. I also fell into this during the era when DS as a field was being forged, so the bar of entry is massively different from where it is today. But who knows, someone out there might find my odd path interesting.

That mandarin duck in Central Park from 2018. Because why not?

I spent much of my days as an internal facing resource, providing insight and support to internal teams, often helping project managers make product decisions. That’s left its mark in how I approach many problems: I’ve said it before, but I consider myself a force multiplier, in that I help people do their jobs even better with the help of data.

Jump forward some years and now I’m a “Quantitative UX Researcher”, a title that doesn’t seem to exist in many places. The only real mention of it I can besides an occasional job posting is this article by some folks at Facebook. Where they distinguish qUXR as having significant overlap with a data scientist, but with a more user-centric focus.

To sum up a large and complex field like User Experience Research in a handful of paragraphs is rather futile, but it ultimately boils down to understanding the relationship between people and a product, how and why people use something, where it all fits in their life and the world. It’s a multi-disciplinary field that draws on aspects of design, writing, the social sciences (such as psychology, sociology, anthropology, HCI, etc). On average it carries a more qualitative/design-y public image than the mathy hard-science image of the Data Science field today.

While the field of UX Research has always made use of various quantitative methods, such as surveys and experiments, as Big Data became more important to understanding users a niche started to open up where a more traditional qualitative researcher didn’t have the engineering and quantitative skills to make full use of the available data. From there, talented folk who would have done well as a data scientist or analyst, but wanted to focus more on understanding users moved to fill the niche. That’s how I found my way there.

How I wound up here

I came to data science essentially by leaving the social sciences with a master’s degree. Afterwards I did a giant tour of a number of analyst positions across a bunch of industries. The one thread throughout is that I’ve largely been in smaller startups (20–150 people) because I like wearing multiple hats, and often wound up in a role of deriving insights from data to help other teams.

For undergrad I had studied philosophy (not just any philosophy, Continental Philosphy!) and business administration (operations management, decision support systems). I then got a social science masters in Communications. There I did a mix of computational linguistics while getting social science theory, methodology and philosophy of science drilled into my head. This is probably why I still talk about “doing science” so much even now.

Out of school I joined an interior design firm that redesigned offices. They used a relatively novel method of having people walk around offices regularly, collecting quantitative data on space utilization to prove their case on how many meeting rooms and desks people needed. I did a lot of survey and data analysis there and learned to read the unspoken issues people had with their workspace from tables of space quality ratings and free responses.

I had a very brief stint at an ad-tech firm during the height of the market crash. It was a tiny 25 person shop and was a toxic revolving door. But it paid the bills and I learned SQL there (on the first hour of the job, on a production server). I also learned the extent of stress I was willing to take. We’ll leave it at that.

I later moved to Meetup where I spent my time supporting product development teams, working closely with project managers and helping teams size out potential feature impact, set up and run tests, and build a bazillion reports. At the same time I supported the entire company with data, eventually helping out everyone from legal to customer support.

On top of all that I was best friends with the qualitative researcher on my team and the two of us would combine powers to take over the world. Working together, we would fluidly move between qual and quant methods to understand users while becoming familiar with the methods used by the other side.

Afterwards, I moved to Bitly and got a taste of what businesses with enterprise customers are like (Meetup’s organizer subscription model was essentially a consumer model). The dynamics and questions are quite different, with a heavier emphasis on direct feedback from customers and a more focused scope. I also got my first chance to play with legitimate Big Data™ — running raw map/reduce jobs on a Hadoop stack, sometimes doing silly things like sending grep jobs over the Hadoop streaming protocol. As usual, I was the lone data analyst on staff, but we had a couple of data scientists building new product and a very strong data-driven culture put in place by previous staff.

Next came Primary, a very small (< 30 person at the time) company that designed, manufactured and sold kids clothes online. It’s very rare to work at a small startup that actually has a legitimate supply chain to deal with. They’d actually be buying cloth, moving it to factories and finally to a warehouse shipping product to customer’s doors. Inventory management and cost calculations get really complex. It was the first time I was the lone data engineer helping build out infrastructure and data culture at the same time.

Finally, Google offered me a deal I couldn’t pass up, not as a data scientist (often called an analyst in Google’s job classification) but a Quant UXR. I’ve actually had chats with analysts and we were amused to find that our base skill sets are largely the same. The main difference really does boil down to how close I sit in relation to product teams and my preference for using tools yield interpretable results.

What’s the job like?

Without bogging down into specifics, it’s a researcher role, so it’s ultimately searching for (actionable) insight. Depending on the needs, things are tactical and strategic. Expect to switch gears to handle any random situation that comes up, and run things in parallel.

You could be supporting a brand new product, at which point you’re dealing with all traditional startup issues like tiny sample sizes, questions about market fit, understanding initial users, ramping up adoption and sales, running tests on your initial crappy funnel page, etc.

Or you could be working with a more mature product with an established user base, years of historic data, actual instrumentation, with a focus on more esoteric segments of the user population, handling big data, and dealing with legacy systems and issues.

Tools and methods are the same as a typical DS, SQL in various forms, Hadoop, Python, R, Jupyter notebooks, etc. Spreadsheets and slide decks are still unavoidable. You’ll still be dealing with arcane tech stacks that stretch from front to back end, so you best have your technical skills in order. End of the day, a lot of the job is still Counting Things Correctly™.

The biggest part is likely the constant communication with cross functional teams that’s. You’ll be part of strategy meetings where you’ll get question that need answering with data. Then presentation of findings. On top of that, general research to confirm (or challenge) decisions coming down the pipeline. Plus working with people to instrument things, as well as smaller tactical work.

All typical sounding stuff.

Where next?

As of this writing in early 2019, nowhere yet. I’ve always enjoyed figuring out what users are doing by looking at walls of data. There’s always going to be a place in the world for that kind of work, no matter what title it carries.

As usual, feel free to hit me up on twitter if you’ve got questions.

Here’s another duck because it’s cute and I have too many of these photos.