Ribu M Jacob

Yes. So I began my career in 2010 as a presales consultant in Tata Consultancy Services. And I did my MBA from IIM Calcutta in 2013. I moved out in 2015, and in 2016, I started my career in start-ups, where I worked in hotels for a year. I was a business development manager, handling business in terms of adding more properties, hotels to the network. I did that for another year at Ojo Rooms till 2018. And in 2018 to mid 2022, I was working in Zomato, where I handled the business. I launched Zomato in Kerala. I first launched in Cochin and then went on to launch another 11 odd cities. Dovato is essentially a food delivery aggregator, and I handle the business for that, which will include all aspects of the business such as account management, marketing, local marketing, P&L management, and so on. I have a work experience of about 12 years, and that is my background.

The process of tailoring a data presentation to different audience types is yes. The first thing I would try to understand, obviously, is the kind of audience that I'm presenting to. For example, if I'm presenting to senior management, I would ensure that I keep the amount of data on the presentation to a minimum and make it very easy to understand, making very high-level abstracts, but with a few key performance indicators that are relevant to that stakeholder. So, the process would be to identify who is the stakeholder that I'm presenting to, understand what they are looking for from the presentation in terms of business objectives. Business objectives will differ from senior management to middle management; objectives do differ. So, I would keep those objectives in mind and then tailor my presentation according to their objectives. And if additional data is required during the presentation, I would be able to present it and justify my presentation to the user with the required data if asked for. Ultimately, the framework would be around understanding the audience, what their objectives are, and tailoring the presentation according to those objectives. Highlighting the data points that show how things are moving.

So, actions based on signal analysis, I would have a definite framework around it. My framework would be around a cost versus impact framework. I would provide recommendations in a descending order of significance, where the options that would provide maximum impact for a certain cost are presented first. The first option I would provide is the one that has maximum impact with the least cost. However, depending on the objectives of the stakeholder I'm presenting to, their choice would differ, but my framework would be based on a cost-benefit, cost-impact analysis.

How have you used your problem-solving abilities in past products related to marketing data analysis? Yes, in my time at Zomato, I used this a lot. So, for example, in Zomato, initially, we did not have market segmentation. User segmentation, and we developed that framework where we divided the market into, or rather, users into particular segments. We called them the less affluent users, the mid-market users, and the up-market users. Now, the LA users, the less affluent users, were classified on the basis of their average order value on the platform. So, essentially, the classifier used for segmentation was the average order value. And on the basis of this average order value and the frequency with which they ordered, etcetera, we defined these segments, and we tailored campaigns around these market segments. So, initially, when we started off, we did not have any segmentation, and we were using campaigns for all segments in one go, which did not provide the best return on investment. Our post-segmentation efforts allowed us to understand the behaviors of most of these users. For example, LA users cared more about discounts and more about campaigns. So, we tailored more campaigns, with a higher volume of campaigns, towards those segments of users. While for the up-market and mid-market users, we had slightly less expensive campaigns. And when we used segmentation and provided tailored campaigns, we not only saved a ton of money but also effectively bettered our performance. We had more loyal customers.

What approach do you use to analyze marketing trends and patterns within data? Essentially, I would look at the customer journey involved. So, for example, I look at a marketing funnel. Right? So customers come in through various sources, and I'll think Zomato is an example to explain this. Customers come in through various sources. They come on the app, they open the app, they open the restaurant menu, add their items to the cart, and then they place an order. This is a marketing funnel that's there. Now, I look at this funnel, I look at the progress of customers across this funnel, and I also look at the time periods when this happens. In terms of analyzing the volume through this funnel, I segment it based on time. And so, on a macro level, what I look at is how my volume looks at specific time periods. And within each specific time period, how does the volume of these customers move through the funnel? What are the funnel matrices? How does my customer's conversion rate change across each step in this funnel, across these different time periods? What is the conversion rate of each segment of my customer across this funnel, across these time periods? This is the approach that I would take. One thing I might have missed out mentioning is the source of my customers as well. I can have direct sources, or I could have sources from other web sources.

I have primarily used Excel and Google Sheets for a lot of the analysis. I get the data through tools that use SQL queries to extract data from our ETL processes. That data is then used in Excel, analyzed using Excel, and then represented using Google Data Studio or Excel for presenting proceedings data to stakeholders. But, primarily, we have our data infrastructure set up to capture large volumes of data, which is then analyzed. There are batch processes to put this data into the data structures, and then SQL queries are used to pull out data for specific use cases where the queries are customized to extract specific use data. The data stored in the database is clean, and the quality and integrity of the data are automatically ensured. We then pull this data in using Excel, using SQL queries, and use Excel sheets to do analysis on that data, and then present it using Google Data Studio.

I'm extremely proficient with Excel. I use vlookup and all the other tools within Excel to do analysis on the data. I'm also quite familiar with Tableau for data representation. I'm also familiar with using Python for analysis on extremely large datasets. I'm also familiar with SQL to extract data from large databases. However, for analysis, I use Excel to a great extent, but I'm also familiar with Python and I've used Python for extremely large dataset analysis. But Excel remains my go-to tool. Primarily, I've used VLOOKUP. I've also modified data to get it into a certain format, and so on. So I'm very proficient with Excel.

Data-driven decision-making processes. Absolutely. I'm ready. So, just to give you an example at my time in Zomato, I'll go back to my previous example where we didn't have data segmentation, user segmentation earlier, but we brought in user segmentation a year down the line where we divided the users into less affluent users, upmarket, mid-market users based on the average order value, the frequency with which they ordered and so on. So once this came in, we were able to bring a lot more insight into our decisions. So earlier, while we were analyzing the data as a whole, once we segmented the data, we were able to get more granular insights on each segment, and we were able to target these segments better, making their user experience a lot better once we analyzed them individually. Fundamentally, my process around using data is to first define what our key objective is. And then for taking decisions to achieve those key objectives, I process the data that I have. If the data is already segmented and sorted, then there's no problem. But otherwise, we have to figure out how to get the data and analyze it to help us make decisions that will help us make better ones.

So, I would have of it, I've been going to use GPT, large language models like GPT-4 to for data representation because it gives me an easy way to interact with Python, use Python to provide graphical representations. But when I started my career, I used Excel for data representation. Some time with Excel passed. I also realized that Google Data Studio and tools like Google Data Studio and Tableau are also extremely useful and way more flexible than Excel is. So, in the middle of my career, I started using tools like that. But right now, I use large language models and GPT-4, which makes it way more easier to help me produce these visualizations.

New experience, how have you used annotations to highlight key data points or trends in your data visualizations? I'm I have to say, I'm not really sure what you mean by annotations, but if what you mean is how I highlight data points in a representation, I make it very clear. For example, if I'm using a bar graph, I make it very clear that this background represents a, b, c. For example, maybe the background represents bugs, or maybe it's a pie chart. The pie chart represents our market share in a given geography, and this is our market share. I express that in percentage terms, and then I also provide a slight annotation aside to indicate that the percentage is out of a population of 25 lakhs, and so on. So, that kind of annotation I do provide to make it very clear to the stakeholder analyzing the visualization that what I'm showing you is the market share that we have. I ensure that I use annotations to help the stakeholder viewing my visualization understand what exactly I'm representing using the visualization. That has my visualization would leave no room for doubt to the key stakeholder about what it is exactly.