Playfair’s Hiring Process Explained


Our effort to build a fair recruitment engine and identify the best candidates

We’re hiring!

This post is meant to shine a light on Playfair’s process for assessing applicants for roles on our investment team. Our approach isn’t perfect, but we put a lot of thought and effort into doing our best. I welcome feedback in the comments section below!

Shortly after joining Playfair seven years ago, I wrote a blog post directed at startup founders in which I advised:

1. Recruiting is by far one of the most important aspects of building a successful business;

2. Without assembling the right team, your startup has virtually no chance of success;

3. A massive effort is required to build your team; effort that you personally must carry out and which must be prioritised above almost all other activities.

Luckily, I followed my own advice (for once) as I set out to build our Fund 2 team four years ago. At the time, I dedicated an entire six months to doing nothing else but finding and hiring Chris and then Henrik. Once they came on board, we continued to over-index on hiring in order to assemble our top performing VC team (not just my opinion, we’ve got the numbers to prove it!) with incredibly complementary skill-sets.

As we prepare for the launch of our third fund, I’ve been relentlessly sharing my view with this team that the recruitment process we’re running this summer — to find our next two colleagues — is the highest leverage thing any of us will do over the life of the fund.

Here’s why.

What’s the Point of a Recruitment Process?

It may seem like an obvious question, but we’re into first principles thinking here at Playfair. The first and obvious reason for a recruitment process is that demand will massively outstrip supply for our two open roles. I’m predicting we’ll receive over 1,000 applications, and so we’ll need to figure out a way to pick just two of them.

One option would be to just hire the first two applicants or, once the deadline for applications has passed, to hire two at random. Crazy to consider at first, but random selection has its merits. For one, it wouldn’t be much of a resource drain on the team. It would also do away with the biassed reasoning each of us interviewers are sure to employ when assessing candidates at interview stage.

The problem with random selection, however, is that it’s a suboptimal strategy when we consider the outcome we’re looking for from the process: to hire the two people who will perform best in the role and as members of the Payfair team.

No two people are alike and neither will any two people’s performances in a complex role such as that of a VC analyst . Since we believe that we’re on an important mission here at Playfair (to efficiently allocate capital to the entrepreneurs who will meaningfully advance humanity by building useful things), our aim is to maximise our success, and to do this we need the best available performers on our team.

This is why we run a recruitment process at Playfair: to find the candidates who will perform best in the role.

Fine, but how can we know which two of our ~1,000 applicants will perform best in these roles? Well, we can’t. Not definitively anyway. But we can do a lot better than random selection.

How? First by understanding, from both our own personal experiences and the scientific literature, that certain human attributes are very good (albeit imperfect) predictors of success in a role like the one we’re hiring for (that is, a knowledge role). Second, by understanding that certain assessment methods can be used to (again, imperfectly) predict which candidates possess those specific attributes, and thus are likely to perform best amongst all applicants once in the role.

Now let’s discuss what those attributes are and how we structure our recruitment process to find the candidates who possess them. But first, an important note on a unique aspect of life as an investor at Playfair.

Straight in at the Deep End

Fede, Chris and I strongly believe that the only way for those who are new to the VC industry to experience a steep learning curve is to learn by doing and learn by failing. It’s how Fede learned his craft as an angel. It’s how I learned mine when I was the only investor on the team for a while. Chris’ first job in VC was as Playfair’s managing partner. Obviously we each brought a great deal of relevant prior experience to our roles, but we still had to figure out many of the fundamentals over the first few years, often making mistakes, then recognising, fixing and finally learning from them. In this sense, it’s somewhat a similar experience to that of a first-time founder (although what they do is 10x harder).

Without a doubt, the most impressive learn by doing journey I’ve witnessed at Playfair — or anywhere for that matter — has been Jeevs’. What an absolute beast she’s been over the past two years. She had some big setbacks early on, but bounced back from them with such an impressive determination to figure out what went wrong and what needs to happen differently next time similar situations arise. When I unexpectedly took a three month sabbatical earlier this year, she was thrust into working solo with some of our founders and, showing an incredible level of resilience and on-the-fly problem solving, helped a bunch of them overcome very difficult challenges to building their businesses (and now they all want to work with her instead of me 😢).

What do the best candidates look like?

Anyone who knows me knows I’m obsessed with figuring out how companies can hire the best candidates. They also know that I hate arguments from anecdote and faulty generalisations, and that I believe in the scientific method as the best strategy for attaining the most objective, most accurate, and deepest truths about facts of any kind. That’s why the recruitment process I’ve built at Playfair is grounded in science.

Part of my collection of academic papers on the science of personnel selection, which numbers around 200 unique articles.

And the science is clear that there is a single best predictor of success across every knowledge role, including the one we’re recruiting for a Playfair: general cognitive ability.

GCA is essentially the ability to learn. It is not synonymous with intelligence, which implies genetic potential. Rather, it refers to “developed ability”, or a person’s capacity for learning things through processes such as abstraction, logic, reasoning, planning, critical thinking, problem-solving and creativity. It has a large genetic component, but is also built and strengthened through an individual’s formative years.

These skills are absolutely critical in venture capital investing, especially on a team like ours, where investors need to deploy all of the capabilities above on a daily basis to solve novel and complex problems.

Next up as a strong predictor of performance is conscientiousness, a personality trait of being disciplined, focused and responsible. Crucially, conscientiousness implies a desire to perform tasks well and to take obligations to others seriously (our steadfast duty to our founders is one of our core principles).

The lists of predictors of job performance isn’t exhaustive and I’d have to write a collection of books to list all of the ones we currently know about. For those curious to do some more research, have a look at things like: the need for achievement, possessing an internal locus of control, emotional stability, empathy and tolerance to ambiguity.

However, it is worth mentioning a few others that we’ve identified — more from personal experience than the literature — as being important to succeed as an investor at Playfair. These are a genuine interest in startups and technology, intellectual curiosity, the ability and desire to build productive networks, and humility.

As a team, we spent the first few weeks of summer doing our research, brainstorming and compiling a list of the attributes we want to assess each candidate against in our recruitment process.

Ultimately, the best candidates will be the ones who possess these traits in the highest quantity, or in the best combination. It’s really that simple. Our entire goal in this process is, once we get to the end of the recruitment process, to be left with the two individuals who are the smartest, most conscientious, most passionate about startups and technology, etc. etc.

The final question is, of course, now that we have our 1,000+ candidates and our list of criteria against which to assess them to find the best, how do we actually do it? The assessment process, of course, which looks like this:

For a rationale on this choice, we need to revisit the science (yay!). During a recruitment process, you need to figure out which characteristics predict performance, but also which methods of assessment have the highest predictive validity over those characteristics. Here’s what the science says are the best predictors:

From Schmidt (2016)

Over the course of the assessment process, we’ll use each of these selection procedures. But before we get stuck in, a note on referrals. Simple put, we don’t accept them. Anyone who wants to be considered for a role at Playfair must apply the normal way like everyone else. As you’ll see, the entire point of the massive effort we put into our recruitment process is to level the playing field for all applicants.

Stage 1: Resume and two basic questions

The first stage involves me manually reviewing every application (don’t feel sorry for me, I actually love this part). This goal here isn’t to identify the strongest (say top 5% of) candidates, because that would be impossible to do from looking at resumes alone, but rather to filter out those who are least likely to be amongst the strongest, as well as the more irrelevant ones (by which I mean, for example, people who live on the other side of the planet, since we’re not in a position to sponsor someone for a visa for a new grad role). By “least likely to be amongst the strongest,” I mean those who clearly don’t meet our base criteria of having a genuine interest in startups and technology or a high level of conscientiousness.

I’ve conducted over 250 interviews for investor roles at Playfair and can tell you unequivocally that if a person doesn’t already love and want to work with startups or cutting-edge technology, they’ll be out-prepared and outcompeted in the interview process by those who do. Ones lacking conscientiousness don’t put much effort or enthusiasm into their applications and probably wouldn’t, we assume, put much effort or enthusiasm into the job.

You can spot these candidates a few ways. For example, some won’t bother to answer the very basic questions we ask as part of the application process:

Because it’s so easy to apply for a job these days, we’ll see a lot of applicants who don’t really know much about VC and aren’t primarily interested in finding a job in this industry. For example, some will have resumes that are clearly written to secure them a role in a different industry (e.g. the Summary section will explicitly say something like “Recent graduate interested in roles in fund administration”).

About 60% of candidates will be disqualified at this stage.

Stage 2: Psychometric assessment

Because we can’t humanly interview the remaining ~400 candidates, and because we want to maintain objectivity in the process at this stage, candidates will automatically be sent a request to complete an online psychometric assessment, which has three parts:

  • Abstract Reasoning, which assesses the ability to think logically and solve abstract problems.
  • Numerical Comprehension, which assesses the ability to analyse numerical data and make data-driven decisions.
  • Attention & Focus, which examines the ability to focus on individual elements to solve problems.

This general cognitive ability test uses a methodology for scoring called Item Response Theory. Candidates are compared against the average performance of all candidates that have completed the tests.

In order to present results, a T-Score is used. Unlike a percentage score that one might receive on a school exam, a T-score is based on a grading model where average performance is always “50”. Most T-scores will range between 30 and 70. A score of 30 is a very low score, a performance similar to the lowest 2% of all candidates globally. A score of 70 marks a performance better or equal to 98% of all candidates.

When all of the results are in, we’ll see them plotted on a chart showing a normal distribution. Here’s an example of what a score for a candidate will look like (not a real candidate):

I’ve done some testing over the last few recruitment processes we’ve run at Playfair, where I randomly picked candidates from the left three quartiles to progress to the next stage of the interview process. These cohorts always perform significantly worse than those in the right quartile and no individual from within them ever got to the final interview stage.

At this stage of the process, most companies would progress the top ten performing candidates to the next stage of the process. But we want to give more people the opportunity to show us what they’ve got and to interact (if even very briefly) with a member of the Playfair team. As such, we tend to progress the top quartile of psychometric results — about 100 (!!) candidates in total — to the interview stage.

The sheer volume of interviewing Jeevs and Sheff will do at the next stage means they’ll have to substantially clear their schedules for three entire weeks in September just to get through all of the candidates.

Finally, it’s worth noting that hereafter we don’t review atheny candidates’ psychometric scores, as we feel everyone who progresses beyond this stage has sufficient brain power to perform to the standard required by the role. At this stage, the difference maker will be other important characteristics.

Stage 3: 30-min video interview with Jeevs or Sheff

Here’s where candidates get to meet a member of the team and ask questions about what life at Playfair is like.

We use structured interviews at this stage. When I blogged about structured interviews back in 2016, the science was clear on their superior predictive validity versus unstructured interviews. However, the application of a new, more accurate testing method has changed this conclusion. As you can see from the table above, structured and unstructured interviews have an identical operational validity, with structure interviews being slightly better predictors when paired with tests of GCA.

Notwithstanding, I still prefer structured interviewing at this stage because in my experience they are better at controlling for interviewer biases.

Jeevs and Sheff will interview 50 candidates each at this stage, asking candidates a number of randomly selected questions from a question bank we’ve prepared. The questions will be based on the specifics of the role itself and will generally take the following form: “can you give me an example of a time when you’ve done X?”

The interviewers are required to take detailed notes of each candidate’s answers, as well as to provide an analysis of the candidate’s performance and a score based on a similar rating scale we use to review pitches.

The reason for detailed note taking is that, once the interviews are concluded, Jeevs, Sheff and I will sit down together and review each application, looking in detail at each candidate’s resume, their answers to the written questions and the detailed written feedback. This provides an opportunity to spot any biassed analysis and scoring by the interviewers, given that 2/3rds of the reviewers will not actually have met the candidate yet.

Stage 4: 30-min video interview with Jeevs or Sheff

We’ll be down to ~30 candidates at this stage, each of whom will interview with whoever of Jeevs and Sheff they didn’t interview with at the immediately preceding stage.

The same review process occurs to whittle down the candidates pool to our ten finalists.

First onsite interview: “The Deck Review” with Henrik + a 30-min interview with me

Here’s where the final candidates come to our office and co-working space in Clerkenwell to meet the entire team in person. We’ll show them around and answer any burning questions they have about the fund or us as individuals.

Because we value the importance of grounding our interviews in what the daily life of an analyst on our team will be like, the next stage will involve candidates collaboratively reviewing a bunch of real-life pitch decks with Henrik with a view to answering one specific question: should we take a call with this founding team? At Playfair, we receive hundreds of pitch decks every month but, as is the case in our recruitment process, we only have the capacity to speak with a few of the founding teams. One of the main responsibilities of an analyst at Playfair is to work closely with Sheff to screen all of our inbound deals and decide, based both on objective criteria and their unique subjective assessment, which ones we should take a call with.

After this, the candidates will have a 30-min interview with me on a topic that I can’t disclose because that’s part of the challenge 🙂

Stage 5: Final onsite interview: “The Memo Task” + a 1-hour unstructured interview with Chris

The last stage of the process involves a take home task. Here we’ll present the candidates with a pitch deck (that a founder friend of mine has kindly agreed to let us use), a few days before coming back to our office, and ask them to write one or two sections of a short-form investment memo. Gathering information from disparate sources and summarising and analysing it in an elegant manner is another important duty of an investor at Playfair.

This stage is also about seeing how well a candidate works independently and without much guidance. When the candidates visit us again, they’ll take part in a mock investment committee meeting where we’ll discuss the work they’ve done and ask them to elaborate on certain aspects of it.

And finally, each candidate will have an interview with Chris with a focus on their career plans and working style.

After this, we’ll convene a final hiring committee meeting and decide which two candidates we’ll extend employment offers to.

Go to Publisher: Playfair Capital Blog – Medium
Author: Joe Thornton