Shortlisting Process

Introduction

Let's try to understand:
- the purpose of this process
- the steps involved
- the overall formula that describes it most simply

Purpose of Shortlisting

It's easy to assume that the purpose of shortlisting ~50 students is to identify the top 50 students and to interview them. This is close to the correct answer, but not quite. Because of this small difference, there are always a few shockers, that make one think that the firms' shortlisting processes are broken / unfair.
The firm's objective at the end of the day is to recruit 10-20 good enough candidates (not necessarily the best), and to therefore shortlist the 50 candidates that give them the best shot at recruiting this number. Let's breakdown what this means:

- Good enough -

As an associate or analyst, you will not make much of a dent in the firm's business. You need to be analytically sharp, a good communicator and a fast learner. If you're better than average, that's nice for the firm, but if you're below average, that's pretty problematic. Keep in mind that consulting firms make money by selling new projects to clients, which partners do. One associate growing one client's revenue by 15%, instead of 10% isn't doing as much for the firm as it might feel like. Therefore good enough, is where the RoI is best for them.

- Best shot -

The top 50 students as per just one set of dimensions, might not be the best set to look at. If every consultancy was looking at the same set of 50, one of them might end up with candidates 30-50, which might just be on the borderline of good enough. Most firms would agree on who the top 30 candidates are on paper, and everyone shortlists them, because recruiting some 10 of the top 30 is a safe bet. For the next 20, each company indexes on some less stereotypical dimensions, which are slightly unique to each firm. This hedges their bets against recruiting just one type of person, and ensures that they can more thoroughly vet those candidates since they aren't being interviewed by competition. This is how consultancies ensure that they get at least 10 good candidates, and possibly another 10 good candidates, if they have a good placement drive.

Steps Involved in Shortlisting

Each company might follow a marginally different process, with different types of people involved in slightly different steps. However, the below model will produce much the same result as whatever system is being used, and therefore remains a good mental model to think about how your resume will be looked at.

1. Automated filtering -

With 300-400 resumes coming in from a given school, it doesn't make sense for each and every one of them to be looked at. So a certain set of basic criteria are used to filter resumes. This may be actually automated, or may not be. But the idea is that this first round of filtering is very mechanical, and may as well be automated. Between fairly standardised fields such as undergraduate college, CGPA, competitive exams, company names in work experience and internships, years of experience, national standardized extra-curricular achievement fixed thresholds are put in place. Across all of these fields together, each selected resume needs to have at least one or two fields on which it crosses the threshold. There won't be a large number of resumes which get filtered out in this manner, but it may reduce the volume down to 50% or so. Keep in mind, that this process might not be in place in every firm, but for all practical purposes, you can imagine, that a resume which doesn't have anything strong across all standard fields, isn't going to be shortlisted anyway.

2. Manual scoring -

This is the most core step in the process. Once again, the specifics of how it's done may vary, but the below 2 models illustrate what you can expect the outcome to look like:
  1. Peaks method - In practise more consultancies might have moved to the more tech driven point-wise score model. However, I prefer thinking through the peaks method, because it's easier for a human (not privy to the scoring rules) to accurately visualise and arrive at a fair assessment. Each common subheading on your resume (eg. Job role / college degree + cgpa / internship) is treated as an achievement. Each of those achievements is then rated as excellent (peak), very good or average. Average achievements are essentially ignored. A resume with 3 peaks gets marked as highly recommended. One with 2 peaks and 2 very goods or any other such slightly lower combination will also be passed on without a strong recommendation. Resumes with just one peak or less aren't very likely to make it through. The exact threshold of how many peaks are required is largely a function of your batch, but will not vary very much from the example outlined above. The definition of what counts as excellent vs very good vs average for a given achievement area is mainly a function of what the average batch has looked like at your school for the past few years. Practically, you can assume that if your achievement puts you in the top 10% of your batch (these are fairly comparable, common subheadings) it's a peak.
  2. Point-wise scores - The effect of the point-wise scoring system is similar. Here each and every line (point) on the resume receives a score of 0, or 1. The sum totals of al resumes are compared. Here also, only common enough points will be given a score, similar to the peak method, and scoring will be based on similar principles. This is tougher to visualise, because if the same project is described in 3 lines or 5 lines, doesn't actually have an impact on the final score, but appears to if we score each of them individually.
Through some system like the above 2, all of the resumes are scored by analysts/associates/recruiting team. In doing so, a list of the top 75 or so resumes will be selected and passed on. Out of this list, some of them, maybe the top 25-30, will come highly recommended

3. Top level selection -

These top 75 odd resumes will now be subject to scrutiny of the partners or other senior consultants taking part in the recruitment process. For the most part, the top 25-30 resumes, which have come in highly recommended will be looked at and retained as a part of the shortlist. Out of the remaining 50 odd resumes, which average 2 peaks across either stereotypical or non-stereotypical dimensions, 25 must be selected. Here, the partners try to select the mix which appeals most to them. The core idea being that their experience should enable them to identify also candidates whose full potential might not be completely visible through just a resume. As a function of human nature, what happens is that the partners are highly likely to select now a mix of candidates who are similar to those in the first pick of 25, as well as candidates who have more unique peaks (authors, actors, sportspersons etc), which match the partners' experience of talented individuals. Over time, this is what builds into the noticed patterns of certain firms favouring candidates with a certain kind of experience, type of extra-curricular achievement etc.

Shortlisting Formula in a Nutshell

The reason that formulae and frameworks are so popular in consulting is that it simplifies a very complex world into a very solvable equation. They may not model all of the nuances and details, but they are still fairly good predictors of outcomes. Therefore, I will repeat once again, that each consultancies actual approach tp shortlisting might be marginally or apparently very different from what is outlined. It is also very important for them to oversell the idea that each shortlisted candidate was handpicked by partners. However, as distinct as their process might be at any step, the final shortlist can very nearly be predicted by applying the above explained process instead. To simplify the above approach further, you can always apply the following peak test to your resume:
1. What are the top 5 most impressive ideas on your resume (non niche)?
2. How many of these are peaks (objectively better than 90% of the batch)?
3. What are the other interesting things on your resume (unique / niche)?
4. Now you can order the batch first according to the number of peaks, and within a group order them according to whose unique points are more interesting (subjective).