FAQ for academic social scientists interested in tech

I’ve talked to a lot of grad students (and a few faculty) about transitioning from academia/research to a job in tech. Mostly the questions I get are similar, so I decided it was finally time to do the lazy programmer thing and stop repeating myself. Warning: unless this actually something you are considering, this will be really boring to read. That’s why I put a table of contents in. So you wouldn’t have to read any more than was absolutely necessary. I wrote as little as I could to spare you.

How does a product team work?

Since most jobs in tech consist of working on a product team, it might be helpful to understand them. Generally a product team is organized around goals (e.g. a growth team might focus on product changes that increase inbound traffic via improved search engine ranking) or an area of a product to improve (e.g. a mobile team trying to improve purchase rates on an e-commerce app). These teams are relatively small (12 people-ish) and typically consist of engineers (7), a designer (1), data scientist (1), product manager (1), engineering manager or tech lead  (1) and a user experience researcher (1). The team is accountable for improving some high level (often metrics-measured) outcome via changes to their product area. Often (but not always) those improvements are measured via online experiments.

What are the kinds of roles that academics can get?

User Experience Researcher

This is one of the more common tech roles – folks who do this role are traditionally people with a human-computer interaction background (HCI) but increasingly you see former academics / social scientists in this role. In this job you are trying to answer research questions for a specific product team in order to understand how best to improve the product. You’ll use surveys, and lots interviews / walkthroughs with small samples of customers in order to identify a product’s problem areas and points of confusion. This role requires some quantitative expertise and research methods savvy (anthropologists and sociologists can excel), and little to no programming.

Data Scientist

This is THE trendy job title these days (Google, btw doesn’t have data scientists — they have Quantitative User Experience Researchers) and needs the least explaining. The big technical skills you need other than stats/experiment design are R and SQL. You are at the right skill level if you can do cross joins, self-joins, analytic functions in SQL and the equivalent dplyr transformations in R, do your modeling in R (python is fine too) and write up your results in some markdown with nice ggplot graphs, and then log into a remote server in order to check in your results into a github repository. The competition in this space is pretty fierce these days, though so is the demand for talent.

Quantitative Analyst / FP&A

Though less known, this role involves forecasting and trying to attribute the impact of changes to the business. This often involves analysis of marketing actions (e.g. the ROI of different ad placement actions), or identifying strategic areas of opportunity through analysis of customer data. Facility in Excel and time series forecasting are key for this kind of role. Often you see people from finance backgrounds do this, but a very quantitative social scientist can do it (I know someone in this role who had previously been a grad student in neuroscience).

People Analyst

These jobs are desirable but hard to get since there are so few companies that have the resources to fund a large team of people doing this or the scale to carry out meaningful research on its employees. This job involves things like running internal company surveys, experimenting with different kinds of workplace changes to maximize team effectiveness, researching the hiring process (sometimes including diversity related work). There are also external firms (like Paradigm) that consult with companies to apply HR insights.

Research Scientist

This role involves higher level research for larger companies. Examples include an ad-tech company that needs a very good ad bidding algorithm that is fast and optimizes some function under constraints. Or a large e-commerce company that wants to improve its product recommendation engine. Or an advanced-technologies / skunkworks type group for a company trying to innovate and develop new products. Some folks called data scientists (like at Facebook for instance) do this too. At Google they are often called research scientists. This often involves some specialty skill — you probably know if you do something that would make you competitive for these jobs.

Product Manager

This is a role to aspire to (IMHO) – it’s pretty rare for someone with only an academic background to jump into product management. I’ve written about this role and it’s relationship to what academics do before.

What is your day to day like as a ____ ?

For most of these roles, the day to day involves a meeting or two (team meetings are a lot like lab meetings) and then lots of time to work on projects with some snacks in between. Projects span anywhere from weeks to months, depending on the role. The lifecycle is definitely shorter than most academics, and the speed is something that can take getting used to. Done is better than perfect, as they say. If you do well and decide to go into management (e.g. managing a team of researchers, or all the researchers in a product area), you become more like a coach. Like Andy Grove says, all of management is motivating and training.

What kind of programming background do I need?

See above. Depends on the kind of job you want. If you want to do product research in tech, the more programming the better, for sure. The more you can do on your own without needing an engineer’s help, the better. Kill what you eat, etc.

Should I do a bootcamp?

That depends on the bootcamp. Insights does seem to lead to people getting tech jobs from people I’ve spoken to. I’m not sure I’d give a full throated endorsement vs trying to get a job straight away, and it does cost a not insubstantial sum (Correction: It’s free, aside from the opportunity cost of your time — I was mistaken). I think they benefit from being somewhat choosy about who they admit. It can also help as a way to get interviewed if that’s proving to be a challenge.

How do I break into the industry?

Networking turns out to be key. Getting a referral from someone, especially at a top company, is pretty essential for getting through resume screening, even if your resume is reasonably good. I blindly sent Google a fairly academic resume in my 6th year as a graduate student and I never heard back from them. I had a referral from someone a few years later and that made a big difference. If you can’t find a connection in your extended network, you can try working an alumni connection (people do this with me on linkedin all the time). I hope that companies can ultimately find a way to make their search processes more inclusive so that people with more diverse backgrounds can break into tech roles. This is a really hard problem that I don’t know the answer to. Not that we live in a meritocracy (or would want to even if it were possible).

What should my resume look like?

Here’s what mine looked like pre-Google in 2011. Notice that the skills I highlight are ones you probably have as someone with graduate training.

What’s the hardest part about transitioning from academia to industry?

Most jobs will involve applications of your technical skills but many (especially ones in a more analytics role, like the Quant Analyst role I described above) won’t apply your social scientific knowledge directly. You might need to get some of your intellectual stimulation elsewhere.

More trying perhaps, you are for sure going to die forgotten and unknown in the annals of history.I was really sad when I left academia. It was part of my identity to be a psychologist, descended in academic lineage directly from William James. I got over it. Eventually.

But don’t worry! Human contributions ultimately won’t outlast the universe itself, which could end in, e.g. in heat death. So nothing we do is enduring against that backdrop! Anyway, in industry you work on a team (something I personally prefer and find liberating), so it might take some getting used to that It’s Not Always About You.

4 thoughts on “FAQ for academic social scientists interested in tech

  1. Pingback: Friday links: Muppet academia, Anscombe’s dinosaur, and more | Dynamic Ecology

  2. Aleksandr Sinayev

    Great post! As an academic social scientists turned data scientist, I would add that knowing Python is preferable to knowing R for increasing your adds to land you a job as a data scientist, since most companies (outside of Google, Facebook and a handful of others) seem to have little interest in doing traditional statistics (statistics outside of machine learning).

Leave a Reply

Your email address will not be published. Required fields are marked *