Analysts should have portfolios
Replace boot camps with slide decks.
If I left Mode tomorrow, would you interview me to be an analyst on your data team?
Perhaps you like this blog, and think you might. Or maybe you hate it, and you’d rather put your business in the hands of Google's "I'm Feeling Lucky" button, Miss Cleo, or the call-in segment of Car Talk than risk it with me.1
No matter your feelings, though, you don’t have much to judge my capacity as an analyst on. Despite me blowing more than my fair share of hot air into the public square, it's pretty hard to say whether or not I'm actually any good at "doing the work." I could be a lousy analyst who's just loud.
For other candidates—the healthy ones who have better things to do with their time than write a weekly blog about data—the information gap is even worse. Unlike sales people, we can’t share our quota attainment records because our work is often impossible to quantify. Unlike engineers, we can’t show people the technology we’ve built or point them to clever bits of public work on Github, because analytical artifacts like SQL queries and Python scripts typically only have meaning in the context in which they were created. Unlike executives, we can’t use our organization’s success as evidence of our own, because there are too many steps between good analysis and good business results. Instead, the only places we have to showcase our work are blogs and Twitter, which favor those who want a public profile, and Kaggle, which favors those who want to brag about their Mensa membership and Klout score without having to go to more than one website.
In the absence of better signals, most of our fates as job candidates are determined by our resumes and LinkedIn profiles. But resumes tend to only capture two things: The pedigree of your prior experiences, and how good you are at writing a resume. Unfortunately, in the latter case, career counselors and Microsoft Word templates seem to have taught us that resumes should be lengthy lists of skills2 and artificially quantified accomplishments,3 rather than useful descriptions of what we’ve done and what we like to do. The only talent we can show off on this format of resume is restraint.
And so, as hiring managers, we apply crude and problematic filters when screening candidates to interview. Have I heard of their prior employer? Does their list of technical skills include the arbitrary things that I prefer? Is their resume too long to bother reading? Were they referred by someone I know? Do I like their name?4
Putting aside issues of nepotism5 and bias, this approach leaves a lot of talented people on the outside looking in. It creates a variant of the experience catch-22—I can’t get a job without experience, but I can’t get experience without a job—in which analysts can best demonstrate their abilities in an interview, but can’t get an interview without demonstrating their abilities. Unlike the experience trap, however, this applies to mid-level and senior analysts. Consider: If you have a decade of great work behind you, but no brand names on your LinkedIn profile, how do you get noticed? Call me a cynic, but adding numbers and action verbs to the three bullets under each job probably won’t do the trick.
As a result, companies lose out on a massive amount of talent, and talented analysts lose out on a massive amount of opportunity. Analysts from blue chip tech companies and top ten colleges bounce around from prestige job to prestige job, while those on the outside fight tooth and nail to get so much as a phone screen. Which is a tragedy—some of the best analysts I know came from famous companies and fancy colleges, but others came from decades-old retailers, unremarkable colleges, and the bowels of Charles Schwab’s audit department. The only difference between the two groups is how hard they had to work to get the interview.
I think we can solve this though—analysts should create portfolios.
The eight slide resume
When hiring analysts, there is one thing that catches my attention every time I see it on a resume: A personal blog of analysis on public data. It’s not that the existence of a blog is a signal; it’s that blogs give you an immediate glimpse of the type of analyst their author is. Are they drawn to machine learning or statistical projects? Do they focus on data collection and cleanup? Are they curious problem solvers, and see data as an interrogative means to an inquisitive end? Blogs also expose how clever they are as an analyst, how crisp they are as a communicator, and a bit about their style and personality along the way. If I could choose one way to screen candidates—a resume review, a take-home coding test, a traditional interview, or a look at their blog—I’d choose the blog.
Blogs, however, have problems. Like contributing to open-source software, writing a blog isn’t something everyone has the time or interest to do. And blogs introduce favoritism, where I might prefer someone if they write about things that personally interest me. For example, I’m predisposed to like a candidate if I find their blog about sports analytics than if they write about the Westminster dog show—both because of affinity for the subject, and because it gives me material to lazily wander through a biased “beer test” interview.
Portfolios keep the baby and throw out the bathwater.
Like design portfolios, an analytical portfolio would be a short exhibition of professional work. It would walk through a few projects the analyst completed, and share the context behind different problems, how they solved them, and the impact of their efforts. It could explain their thought process and their analytical wanderings. They could include partially redacted charts, slides, and dashboards that they built.
Though a portfolio could also include code, the point isn’t to share SQL snippets or a few Jupyter notebooks; it’s to talk about the thinking around the code. As an interviewer, I want to understand how candidates translate a vague business problem into a quantitative one, and how they navigate the inevitable frustrations of their first efforts not working out. I want to hear about what unexpected things they uncovered, and how they found them. I want to learn how they explain problems and tell stories. And as a candidate, I want to share a bit about myself—the projects I’ve enjoyed, the things I’m proud of, and the work I want to do.
A simple deck—the sort of thing for which a first draft could be made in a few hours—could accomplish all of this. Inspired by the overly prescriptive spirit of a five-paragraph essay,6 here’s how I’d do it in eight slides:
Slide 1: A cover slide. Your name, some punchy catch phrase, and cool hero image that you found on a Google image search for “ambition photograph nature.”
Slide 2: A brief bio. A timeline of your career, with notes about each role. Your LinkedIn profile, basically.
Slide 3: A big problem you worked on. Introduce one of your biggest or most impactful projects from a recent job. What was the problem?
Slide 4: Your solution. How did you solve the problem in the prior slide? Include screenshots of charts, reports, or architectural diagrams of what you built.
Slides 5 and 6: A project that you’re proud of. In the same two-slide format, talk through a problem you’re particularly proud of solving. Instead of a “big” one that you’d typically put on your resume, choose something that demonstrates what type of work you find exciting and motivating.
Slides 7 and 8: A project that differentiates you. In two slides again, talk through something you’ve worked on that captures who you are as an analyst. What would make people remember you, and differentiates you from other candidates?
Look, I made an example!
A common app for analysts
Imagine a world where analysts are taught that their ticket to their dream job isn’t a three-month coding boot camp, but a great portfolio that tells the story of their interests, their work, and their career. Imagine if companies used this material, even if it was just a short slide deck, as the most important piece of a job application.
First, this would reorient analytical jobs away from technical skills and towards the creative reasoning skills that they actually require. Whether we like it or not, job candidates are going to teach themselves to the test: So long as we use take-home SQL assignments and whiteboard coding exercises in our interview processes, that’s what people will try to learn. Throwing those out in favor of a more creative portfolio proves to candidates that we’re serious about certain technical skills being “nice to haves.” Just as design portfolios push designers’ choice of tools into the background in favor of highlighting what they actually create, an analytical portfolio puts analysis in front of the languages with which it was performed.
Second, portfolios are an equal opportunity asset. Recent grads can show off school work and personal projects. People looking to make career transitions can include side projects that show their potential as an analyst. Moreover, because portfolios let people tell the story of how they did their work, small projects can be just as compelling as big ones.
Third, the structure of the portfolio itself—the narrative frame around the analytical pictures—is another useful way for candidates to differentiate themselves. Analysts aren't just actuaries who take in and spit out numbers; they're storytellers who need to be engaging and persuasive. Developing and presenting a portfolio allows for exactly that.
Fourth, interviews would be better. Portfolios could replace controversial take-home tests, which turn some candidates away, with material that could be repurposed for every interview (analyst portfolios, the common app for data jobs). Portfolios also give potential employers more context about applicants, allowing interviewers to cover more ground in their conversations with candidates.
Finally, portfolios offer the benefits of a personal blog with less potential for the same bias. By normalizing a focus on work projects over personal hobbies, portfolios would be more likely to be judged on their quality instead of their choice of subject matter. They would take less time to maintain than a blog. And because they don't need to be public, portfolios don't favor those who are comfortable maintaining a public presence over those who'd rather be more private.
Diamonds in the rough
One of the best product designers I know didn’t start her career as a designer. She was working for a fashion company as a project manager—hardly the type of resume that would catch the eye of most tech design teams. But she dabbled in design, and happened to have a chance to share her work with one of the most respected designers at one of Silicon Valley’s most design-oriented companies. Immediately impressed with her talent, they offered her not only a job, but an entirely new career.
I firmly believe there are similar analytical diamonds in the rough7 out there. There are thousands of great hires to be had and careers to be made—but as long as the best way to get people to pay attention to you is by being obnoxious on the internet, they’ll remain undiscovered. Analytical portfolios could finally give them a chance to shine.
Or, given that one of the Car Talk guys is dead, you might need to ask Miss Cleo to ask them. Update: I have been informed that Miss Cleo is also dead.
Accomplishments: While I was an analyst at GameSpot, I conducted frequent analyses to help guide marketing operations and strategy. Results I helped drive: CPC improved by an average of 12%; paid pipeline volume increased by 19% y/y; GameStop stock rose by 2,658%.
But make yours in your own way, because five-paragraph essays are “dysfunctional…off-putting, infantilizing and intellectually arid.”
Ok, so I haven’t seen Aladdin in a while, and this opening scene is really something. A sinister sorcerer meets a thief in the desert. The sorcerer asks the thief for the thing he paid him to steal; the thief says he has it, but had to “slit a few throats” to get it. The sorcerer steals it from him and uses it to conjure a giant tiger head in the sand. The sorcerer commands the thief to walk into its mouth. The tiger head tells the thief not to; the thief becomes understandably scared; the sorcerer tells him to go in anyway. The tiger head kills the thief. The mildly irritated sorcerer flatly says the thief wasn’t worthy; his parrot sidekick makes a sarcastic crack about it.