Deloitte Tohmatsu Financial Advisory LLC
Accurately assessing data scientists, something hard to do based solely on resumes and interviews. Now hiring professionals with both strong analytical skills and consulting aptitude.
As a small, highly skilled team, DTFA (Deloitte Tohmatsu Financial Advisory LLC) seeks data scientists who have both the consulting mindset and the technical expertise to make an immediate impact. However, it was challenging to assess candidates’ fundamental data analysis skills during interviews alone. To address this, DTFA decided to implement Track Test, which allows both the company and candidates to confirm whether the required technical skills were in place before the first interview, creating a more effective and transparent hiring process.
The Challenge
In the past, in-house data scientists were responsible for screening resumes and conducting interviews themselves. However, it was hard to thoroughly evaluate candidates’ technical skills due to time and resource constraints.
DTFA focuses on three core areas:
- M&A Advisory - supporting clients through all stages of mergers and acquisitions, from strategy to post-merger integration.
- Crisis Management Business - assisting clients with corporate turnaround and fraud investigations.
- Innovation Business - helping clients build ecosystems and create new businesses.
These critical areas demand data scientists with practical experience who can apply analytics in key moments, drive better decision-making, and offer high-level advisory insights.
Given this context, our analytics team operates with about 20 members. About half of them are managers and new graduates, so when hiring mid-career professionals, we specifically seek immediate contributors. We simply do not have the resources to train people who are still at the “aspiring data scientist” level.
Initially, the first interview involved a 30-minute consulting case study, combined with an assessment of practical experience and technical skills by two data scientists. For the second interview, two more data scientists joined to thoroughly confirm the candidate’s hands-on experience and capabilities.
"However, we found that it was difficult to fully assess actual technical skills through interviews alone, which often led to a mismatch between expectations and reality after hiring."
In some cases, candidates’ prior data analysis work relied mostly on Excel, with little to no experience using tools like SQL or Python in practice, resulting in limited roles after joining and frustration on both sides.
To break this negative cycle, we sought a more effective hiring approach that would help us identify candidates whose skills closely matched our actual needs. That’s when we decided to introduce coding tests as a solution to objectively assess fundamental technical skills and close any gaps between expected and actual skill levels before joining the company.
The Solution
We started evaluating technical skills quickly and efficiently using a skill test. By exploring the test results in depth during interviews, we were also able to understand the level of training each candidate would need after joining the company.
We chose Track Test among several coding test services because it shows candidate rankings, making it easy to compare applicants and make faster hiring decisions.
Another key factor was the wide range of test domains available. At DTFA, we hire IT talent for various roles, including full-stack engineers. The challenge of measuring practical skills isn’t limited to data scientist positions, so having a solution that could be used across different job types was very appealing. We also saw potential in how Track Test allows us to tailor tests to different levels, such as “Junior Analyst,” “Analyst,” and “Senior Analyst,” and customize problems to match real projects.
After deciding to implement Track Test, our in-house data scientists created basic-level questions, and we revised the hiring process. Now, candidates take the coding test after passing the resume screening, and in the first interview, instead of a standard case study, we have them explain their approach to the test content. Even if their answers are incorrect, we believe we can still assess whether they have the technical foundation and growth potential to handle the job by understanding their thought process.
“So, we decided that everyone who took the test would proceed to the interview, but we started seeing more candidates than expected withdraw from the first interview if they hadn’t done well on the test.”
Honestly, this was a surprising outcome we hadn’t anticipated. From the perspective of candidates whose experience was mainly “doing data analysis in Excel,” the test itself probably helped them realize early on that their skills didn’t match the role, saving them from a poor fit after joining.
Candidates who scored reasonably well could then be assessed further during interviews, which helped prevent mismatches and made the hiring process much more efficient.
The Impact
The screening process finally enabled us to achieve a strong skill match.
In the first year, the quality of new hires became more consistent, and the team became more confident after the implementation.
To manage data-driven consulting projects, solid technical skills are essential. Until now, the real requirements were unclear to some candidates, but Track made them visible and fixed that gap.
Also, candidates interested only in pure data analysis are not a good fit since DTFA’s work also includes advisory services. The new process ensures this fit is confirmed during interviews.
The number of mid-career hires has stayed about the same compared to the previous year. However, the biggest change is that the gap between expected practical skills and the training effort required has been eliminated, resulting in a consistently higher quality of hires. The team now feels confident welcoming new members, and new hires can perform in line with their expectations, meaning our hiring process is finally working as intended.
That said, our use of Track Test is still evolving. There are still many things we’d like to explore, such as strategies for non-Japanese candidates, developing problems that measure consulting aptitude, and expanding the use of Track Test for other roles. These ideas are not yet fully defined, but we see great potential in Track, and we want to maximize its value by collaborating with other departments as well.
“Thanks to Track Test, we no longer see major gaps in candidates’ actual data science skills, which also helps us make informed project assignments.”
