I applied through a referral I got from a friend who had been working at ZS for more than two years. Using his referral, I got a quick reply from ZS Associates with a test link and was a given a period of time to solve it.
Round 1: This was mostly based on aptitude questions, hosted on Talview. Though I am generally very good at aptitude questions, the timings of the rounds made it extra tricky and I believe that I missed some questions. While submitting, I had an issue with my internet, so Talview had a call centre I could call. My assessment was submitted successfully. So in the case, it happens to you, keeps their number handy.
Round 2: There was a short telephonic round that was conducted two months later. I was notified about one week before regarding the same by my recruiter. My interview was conducted by her and lasted about 15 minutes. She asked a question to talk about myself. Then she followed up with two guesstimate-based questions: determine the number of people living in Delhi wearing a red item at that time (ie a shirt, dress, jewellery, pants). This question confused me for a while. She gave me a minute to think as I gave an answer that was not so well-thought-out. For anyone new to guesstimates, it is crucial to think before you speak. I thought for about two minutes before I gave my answer. My logic was simple, I considered the fact it was a Tuesday (the day the interview was conducted). I split the population of Delhi into key demographics based on age: children, adults, and senior citizens taking a 30-50-20 split. I assumed since it was a working day and adults would most likely be wearing muted colours to attend video meetings (it was during the pandemic after all). I also assumed senior citizens would not be wearing such a colour. That left the children. I assumed children would wear bright colours as they do not have to worry about appearing in meetings. I assumed a set of popular colours: red, blue, green, yellow, white. I assumed the proportion of people wearing each of these colours would be same, 20% each. I also chose to ignore jewellery as this was during the pandemic and these are often worn during social events only which was pretty much nonexistent at the time. I calculated the answer and the interviewer responded with an okay.
The next question was the number of red lights (bulbs, tube lights, LED) in Delhi. After some careful thinking, I ignored LED by explaining that during the pandemic, no real festivities were happening. For the question of the tube light, since I knew the area of Delhi to be roughly 1500 square kilometres, I assumed the dimensions of Delhi as 30km x 50km with each road as roughly 500 meters long. I told the interviewer that the red bulbs would be used in traffic lights, so by calculating the number of roads, I could assume there was one per road. I gave her my answer. She followed up with a question regarding whether how many the bulbs would be functional. This part confused me again. I took some time and responded that a majority would be working, so I assumed 80% of the traffic lights were functioning fine and revised my answer accordingly. The interviewer accepted my answer.
Round 3: This was the case study round. I was given a large dataset of a pharma company to study. It was of a company that a leading market share of a product in 2015 which was reduced to a much smaller one due to the entrance of a new company with a similarly reliable product at a much more affordable price. There were no strategy-based questions as per the Case Interview Prep book.
It was all based on calculations. I was given two graphs: one was a graph of market share in 2015 and one was of the revenue share of the products in 2019. I had to estimate some key values using that information. Then I had to determine the three best ways to promote a product. I was given a list of methods with the fixed investment for each. Then there was another list containing the base revenue generated by each method. Then another list matrix is given which paired each method with another method as a follow-up. Each cell contained an offsetting revenue. Using this, I had to determine the best possible method to promote the product.
Profit = Base Revenue * 1.1 + Offset Revenue – (Cost of Method Chosen for Base Revenue + Cost of Method Chosen for Offset Revenue)
I had to choose the best three methods.
There was another question after this to find the best salesmen based on sales data. This one required you to find the best salesman, the worst one based on correlating multiple graphs together.
This also took a lot of calculations, so I suggest anyone appearing for this to be very thorough on your graph based questions. Even though I believe I answered these correctly, I was left with no time to revise my answers.
A follow-up round was held almost immediately, where I got to defend my solutions to an interviewer. The interviewer tried to prod me on details related to my answer. There were questions I was confident on, and she asked me repeatedly if I was sure which I was. The point of this round was to test your confidence basically.
Round 4: The final round was held by a manager. It began with the usual formalities, followed by one guesstimate: to find the number of cars plying through Gurgaon highway. I assumed the length as 20km of which half was five-lane and the other half was three lanes. I ignored bikes for my analysis as I was told to do so. I took the length of each of the cars as 2 m. I divided the period of high traffic: 7 am-11 am and 4 pm to 7 pm. For these, I assumed the traffic was bumper to bumper and it took a car for 1 hour.
(10000metres/2 )*(5lanes+3 lanes) * 2 (for two lanes)
I assumed during other periods traffic was half of this and added the two values together.
The interviewer accepted my answer and wished me luck.
I was rejected for the role.