by Saman Arefi and Philipp Wiedemann
Ever since that now infamous Harvard Business Review article from nearly ten years ago that proclaimed Data Scientist as the sexiest job of the 21st century, companies worldwide had to deal with an oversupply of junior talent; a trend nurtured by the overall big data hype and the countless offerings of three-month bootcamps.
While modern businesses have correctly assessed the importance of data in their operations and adjusted sights accordingly, the underlying hiring practices haven’t.
Arguably, some of this is due to a lack of understanding of what data people actually do: after all, data analyst, data scientist, and data engineer are still roles that are being used interchangeably, resulting in some of the more creative workplaces simply calling everyone ‘data rockstars’ or similar to avoid this pitfall altogether.
It is precisely this gap in understanding between applicants and employers, particularly in the market for junior talent, that causes frustration for all parties involved — to put it bluntly: data is yet another victim of (flawed) hiring in tech.
As two people who had the pleasure to look for a job during COVID and have since also been on the other side of the lever, we’d like to offer some insights into what we’ve experienced and learnt so far.
Data-driven doesn’t mean ticking off boxes in a list
When recruiters receive several hundred applicants per role, naturally you have to filter somehow, and the approaches here differ wildly. Some hiring managers ask supposedly standard questions to test whether applicants are familiar with certain technologies or tools, and to what degree. Others ask for recitation of textbook definitions of common algorithms, akin to a high school recital of a classical play. On the other end of the spectrum, the American tech giants paved the way for an entire industry around “cracking the Google-style tech screening process,” with questions ranging from IQ tests to leetcode tasks. All of these approaches do have their merits — to some extent.
There’s an argument to be made however that simply studying for a specific style of interview doesn’t necessarily mean you will be a good hire. Rather, it may boil down to you being good at this style of interview, with applicants sometimes boasting how they got the job after going through the entire funnel three times.
…and we get it: hiring is hard and COVID has shown us how hard it really can be, with some countries having gone through hiring freezes.
However, we’d like to press for a more holistic approach to selecting candidates.
Essentially, we believe in three pillars when it comes to hiring our potential future colleagues, which we’d like to expand on in the following.
Don’t crack the code, but do get cracking
It took us both twelve days to go from our first chat with someone from Project A to getting an offer. What happened within these twelve days? We had a screening call with someone from the team as well as with TA, did a technical and/or a management interview, presented a case study in front of the team, chatted to some potential future colleagues over coffee and had a final chat with management.
This many steps in such a short time are far from the norm. But if you think about it — they all have to happen, so the question is, when.
Timeliness both respects your candidate’s time and also ensures a quick turnaround internally, maintaining a lean pipeline, while also keeping costs low. Keeping a candidate waiting increases the chance of them taking another offer, resulting in you effectively having wasted all your resources thus far. Sourcing talent is time-intensive, expensive, and does become highly competitive further down the funnel. Don’t keep your candidates waiting, they’re probably talking to other companies as well.
Value (y)our values
At Project A we try to live by a set of values. For us, cultural fit goes beyond us just “liking” you. We appreciate and encourage some edge (off-topic: don’t forget to sign up for this year’s PAKCon about The Art of Dissent!), as being honest is also what’s in our ventures’ best interest. We embrace diversity and look for people who are eager to shape their (working) environment. Our work is collaborative — as such, we put a lot of effort into ensuring that our potential future colleagues are great to work with. ‘Happy to help’ should not be just a phrase.
As previously mentioned, we do make this part of our process, by inviting candidates to informal coffee chats. In this more relaxed atmosphere, we are able to gauge banter and openness that, like it or not, do play an important role in stakeholder management when working with our venture partners.
Do keep in mind that no company is exactly alike to any other, so your company values should help shape your approach.
Knowledge gaps are okay
When doing our technical screening and our case studies, we don’t necessarily look for candidates who will know the answer to every single question. In fact, we appreciate an honest and humble “I don’t know” over something made up on the spot, as we differentiate between knowledge and potential.
Technical interviews shouldn’t be treated like college entrance exams, wherein you judge candidates on a binary correct/incorrect scale. We are not trying to eliminate people based on some arbitrary threshold, but want to understand their way of thinking.
If you, as an interviewer, have little to no understanding of technology and you’re conducting the interview strictly from a pre-written questionnaire from your tech team, you won’t be able to accurately evaluate potentially promising candidates since you’re essentially eliminating the range of possible answers a priori.
We’ve had great success with doing the technical screening interviews (done by our senior engineers/analysts) as the first step in the pipeline versus doing the HR screening interviews first. Again, we do this to keep the timeline lean whilst also relieving our recruiter colleagues from interviewing possible ill-suited candidates.
The case study then bookends the technical part of our interview process, where we again emphasize the importance of keeping it lean: it’s designed to take only about two hours, at the end of which the candidate presents their results to the entire team. Pre-COVID this took place at the wonderful Project A tower. Now — well, you know…
We believe this provides us with great insights into people’s approaches and working styles. At the same time, it keeps the footprint low. If a candidate is still studying, writing a thesis or working somewhere else, tackling a multi-day/week take-home task may not always be feasible.
Disclaimer: We’re not HR people
We want to be really open about this last paragraph: this piece is intended as an opinion on hiring from the perspective of recent hirees who found themselves on the other side of the table right after starting their jobs. We’ve gone through various funnels ourselves and exchanged thoughts on this very subject with many of our peers.
After now having been involved with both getting hired and hiring, the above reflects our thoughts and lessons learnt:
- Timeliness: Be quick, or you’ll be too slow. Data is a competitive field, and that should be reflected in your hiring decisions as well.
- Cultural fit should not be underestimated: you and your colleagues will spend a substantial amount of their waking hours with this person. Ideally, this time should be enjoyable, no?
- Take a holistic approach to assessing knowledge and potential: somebody who aces a questionnaire might not have the potential to grow and improve, whereas someone who might not know all the answers will be that much more eager to constantly be better.
And if you don’t trust our opinion, then trust the numbers: the Data & Analytics team at Project A has a net promoter score of 100, with an overall company score of 77, which is much higher than those of FANGAMs (avg. of 33; includes non-tech teams), indicating that happy candidates do indeed lead to happy companies.
Saman Arefi is a Data Scientist and Philipp “Pippo” Wiedemann is a Data and Analytics Manager. They’re both part of the Data and Analytics team at Project A.