This post covers groundwork and tactics to get the rest of you to follow your foot in the door.
This is the second of two posts to help you land the analytics job you want. Do check out the first about resume creation here.
Just as your resume is a golden document and not just something you churn out like the average to-do item, preparing for the interview involves more than showing up and winging it with your smarts. Even though you work day-in, day-out in your current analytics job, you need to put in some hard yards before the interview given that any or all of the below are possible:
- You work in an industry different from the prospective employer’s.
- The prospective employer’s firm is using a different programming language from the one your role required you to, or one you haven’t used in a long while.
- You implemented a given methodology/statistical technique to solve a problem over a year ago and the details have left your memory.
In short, it’s quite likely that you’ve got so stuck in your job routine that you’re either out of touch with or altogether lack the skills required to nail a given interview.
Failing an interview is not an indictment of your competence, nor is professional competence all that is required to clear one. Nailing an interview requires a combination of preparedness, sharpness and presentation to go along with what you know.
Do note — unless you’re interviewing for an entry-level role, you can be sure that companies are hunting for candidates who are ready, or can be groomed into the role. However, the extent of required grooming that is tolerated is business context and company culture. It’s non-negotiable that you be good at core skills — data pull and manipulation, statistical rigor and presentation.
Every company’s analytics hiring process is composed of different rounds. They test candidates on the below 5 themes (not necessarily one theme per round, or one round per theme):
- Industry knowledge: Bringing in industry expertise, or ability to learn through the interview
- Logical reasoning: Ability to think logically, basic numerical skills
- Data handling: Competence with data storage, data pull and data wrangling
- Business problem solving: Ability to logically break down business problems into factors and metrics, selection of statistical technique, depth of knowledge in applying and tweaking models, model validation
- Fitment: Cultural and behavioural fit
We will now cover each of these in detail.
You are not doomed if you don’t have exposure to the prospective employer’s industry, but there will be time prior to the interview that you can use for basic research about the industry and company. It would help you to be prepared with the answers to the below questions:
- What business is the company in? How is the company doing? Where is it placed with respect to its competition?
- What are the common problems faced by the industry in general, and the company in particular?
- What is the expert opinion on where the industry is headed?
There is no interview round dedicated solely to this, but you will look pretty good if you relate the questions you are asked back to this homework. At the very least, it’ll help you avoid wasting precious brainpower in the problem solving rounds, and give you ammo to respond to some of the fitment questions (more on this later).
It is not necessary that any or every question you face in data handling or problem solving will be industry-specific, but if it is, doing this will give you a leg up.
Several companies want to identify candidates who display logical thinking and numerical aptitude before they invest time in deeper discussions. This could be done through a phone screening or aptitude test, where the candidate can expect questions of broadly three categories:
- Quantitative puzzles: “If two people at either end of a line 100m apart start running towards each other at speed 10 kmph and speed 15 kmph respectively, which point in the line will they meet?”
- Data interpretation (Not likely over call): “Given the chart/table above, answer the questions below”
- Basic statistics and probability: “If you pick two cards from a deck without replacement, what is the probability of picking a king and a red card?”
- Analytical nous (Likely over call): “How would you explain regression model results to a person with no background in analytics?”
There are tons of these questions available online to help you prepare, and you’ve definitely spent several years tackling questions like these if you went through central board exams, national engineering entrances and national/international postgrad entrance exams in India.
Different companies do this differently — you have evaluation methods ranging from online programming tests (with ample anti-fraud measures that use your webcam) to face-to-face white boarding. The skills tested would be some or all of:
- Querying with SQL
- Data manipulation in Excel, R or Python
- Envisioning the optimal schema for a given business case
Companies generally will not hire candidates who do not demonstrate competence in all of these — Data storage, data pull and data wrangling are analytics bread and butter.
The interviewer can casually ask you short programming questions (“What’s the difference between having and where in SQL?”, “What are some ways of processing a dataset in R that are faster than a loop?”) before getting into the problems to code/whiteboard, so be sure to brush up your fundamentals and practice.
There is, of course, a chance that you’ve never worked on the exact tools that the employer is using, but you can just inform them about this, and explain your methodology on a platform you both know about, or do so theoretically. Different employers and roles require varying degrees of programming expertise — The more senior the role, the less negotiable it is that you be good at Python/R and SQL as a bare minimum.
The biggest let-down is when the interviewer goes ahead and quizzes you on a platform listed in your resume only to hear you to respond with “Well, I only did an online course on this”, or “I’m still learning it”. That sort of thing does not belong on your resume, then. Resist the temptation to put it in there or state that you know it.
Business problem solving
Any analytics project goes through the below stages:
Business problem -> Analytical problem -> Analytical solution -> Business recommendations
This section of the interview is usually designed to see how competent you are at every step of this journey. This might take place over one or multiple rounds.
To begin with, a good analyst should be able to turn a broadly stated business context into specific analytical problems, specifying the target metric(s) and possible influencing factors along with hypotheses around how those factors would impact the target metric(s).
For example, the context provided could be “How do I optimize retail stock?”, which breaks down into:
Analytical problems with target metrics: Maximize on-shelf availability, minimize wastage/dormant stock
Factors: Product rate of sale, Promotions, Seasonality, store size, store type etc.
After this, the interviewer could continue to quiz you on how you would see this through to an analytical solution (or pivot to a different context for this — a problem statement from your resume, for example). This is a non-exhaustive list of questions that can give you an idea of what you can expect:
- What statistical technique will you choose to solve this problem?
- Why did you choose this technique? Why not this other one?
- What are the assumptions implicit in your choice? What do you do if an assumption is violated?
- How would you validate your model?
- How do you establish causality?
Intimate knowledge of a technique is important to clear these hurdles. Please ensure you are familiar with all these details for projects on your CV, or state clearly the extent of your involvement in said project if you are not familiar (I once saw a candidate fall to pieces when a problem context was taken straight out of his CV, with a slight modification. That really shouldn’t have happened).
To polish this section off, you might be asked to turn your output into business recommendations. Coefficients and p-values might be your language of love, but business stakeholders speak English, and it is in English that you must communicate your results to them.
This round can seem open-ended and a lot of candidates complain about the subjectivity of it, but most employers are looking for the same things, which candidates struggle to broadcast given the opportunity. Things such as:
- Is this candidate passionate about this role?
- Is the candidate interested in joining the company?
- Is the candidate going to be difficult to work with?
To this end, there are several questions you could be prepared for heading-in. Many of these are commonly asked regardless of industry or role:
- What are your strengths and weaknesses?
- Why do you want to leave your previous employer?
- Why do you want to join this company? (Hello, industry research!)
- Where do you see yourself in ‘x’ years?
- Where do you think analytics is headed? What excites you about it?
- Can you describe an instance where you delivered under pressure/handled a difficult stakeholder?
I can’t give you a script here, your answers will be your own. I will list down some guidelines, however:
- Don’t bad-mouth your previous company, manager or teammates at any point. Even if the reason you’re quitting is because of low compensation/toxic culture/fallout with your team, you have to understand that these things are subjective, and stating them can make you come off as difficult to work with. You can put these in other ways that are just as true — career stagnation, lack of learning — but portray you as someone who just knows what they want.
- When you talk about your strengths, don’t brag, and don’t lie. Avoid cliches and buzzwords like “I am a hard worker” or “I am a team player”. You may have myriad strengths, but be prepared to talk in-depth about the top 3 that are job-relevant. Always have an example or two to back up every point you make.
- When you talk about your weaknesses, definitely mention how you are working on them. Don’t state <too-much-of-my-strengths> into weaknesses — everyone knows going overboard with anything is a negative, and it’s a weak response to do this. These, like strengths, need to be job-relevant (No, this won’t void your candidature unless they are deal-breakers like “not liking to work with other people”. We all struggle with something!). Bonus points if you can name ways you know the company you are interviewing for is known to help.
- Even if you’ve got plans to leave in a year for higher studies or relocate to another city, don’t say that. First, those aren’t goals, just methods you believe will get you to your actual goals (which is what you were asked). It could turn out that the company you are interviewing for can provide the means for you to achieve them from within. If you instead state these, you are broadcasting that you aren’t likely to be around very long, or are set on ways that involve your departure in order to achieve your goal — why would they hire you, then?
There are a couple of other questions you should be prepared for.
- “Tell me about yourself” — when asked this, spend a couple of minutes giving a brief professional introduction — the tenure of your experience in analytics, places you’ve worked, roles you have performed and which industries and tools you have exposure to (An expanded form of your resume’s professional summary, if you will). This isn’t, however, an invitation for you to deliver a verbal autobiography. Avoid mentioning personal details such as your age, marital status, hometown, hobbies and mother tongue. This is a professional question.
- “Do you have any questions for me?” — this is an opportunity for you to know the company more intimately. Feel free to ask specific things you are curious about the company or the job. Don’t put your questions forth in a generic manner (“Can you tell me more about this job?”). Ask them exactly what you want to know: What do they love and hate about working there? What does a typical week look like? What KPIs would you be judged on?
Please feel free to leave a comment with any questions or feedback.
Good luck to you in your job search!