Analytics Insight has collated a list a things that data scientists need to be aware of during interviews

It is quite an evident fact that "data scientist" as a career opportunity is the most sought after. With the world relying on data like never before, the demand for this profession is at an all-time high. So, if you are one of those gearing up to step into this profession, you already know what is expected out of you as a candidate. When preparing for the interview process, one area that seems to be stand out from the rest is that of "business case interviews". Here's everything you need to know about this so that you have an edge over the rest.

What are the categories of business cases?

In most cases, it is observed that the case type of questions are related to –

• Whether a particular product or service should be launched
• Diagnosing a particular problem
• How to improve the services or products
• Measuring success

Don't straight away jump to what you think about the case presented. Start by asking questions so that you get a better clarity about the same. Also, a point worth noting is that there's no right or wrong answer to the questions asked – what the interviewer checks is your "approach" in handling situations.

Now is the time to understand each of the categories in detail i.e. how to answer the questions pertaining to the categories mentioned above?

1. Launching new products / services – Needless to say, the agenda here is to test a product idea or whether to launch a product / feature. This is the most challenging of all the business cases as it requires an in-depth knowledge to present the facts and figures.

• Start by clarifying the goal.
• Talk about metrics to measure the success.
• Come up with an experimental design and include as many discussion points as you can. This helps in creating a good impression.
• Based on your experiment, come up with recommendations. This is the right way to end the case.

2. Diagnosing a problem –Here, you'll be presented with a case that require you to diagnose a problem. The interviewer will look at how systematic your approach Data scientists have to deal with this aspect every now and then, hence the importance. The following points come in handy –

• The first step is to engage with the interviewer / audience. For example, if you are been asked to identify the root cause behind the decreasing sales, you could start off by saying that 'the company has seen a decrease in sales. Is it because of less production or below average marketing strategies or decline in demand? Jumping straight to the approach doesn't seem appealing.
• Next up – talk about whether the metric has changed all of a sudden or has been gradually happening over a period.
• Then, talk about the demographics – whether a particular area is observing this problem or is it the case throughout the area of existence.
• Talk about the internal problems that might have led to this situation.
• Nothing can fall in place if there's no in-depth analysis involved. Without this, everything that you present stands weak. Go in depth as to what is required to solve the issue.
• Lastly, give a brief summary of the whole situation.

3. Improving the current products / services – This too requires good knowledge about the product or service that you are asked your views on. As a data scientist, you should be able to understand what areas need improvement to grow the business.

• Be specific about the features you'd want to focus on.
• Start by clarifying the goal.
• Then narrow down the scope of improvement by focusing on certain areas or features out of all.
• Explain what makes you recommend things that you just stated.

4. Measuring success– This is where the upcoming data scientists are asked to measure the success of a feature or a product. This is probably the best way to evaluate the candidate's capability of defining success metrics. The best way to deal with such business cases is to provide a maximum of 3 metrics, including 2 success metrics and one guardrail metric.

In the end, what matters is that you keep the whole interview process engaging so that the interviewer doesn't lose interest. Clearing this stage is one of the most critical ones in the whole data scientist interview process.