Tech
Feb 26, 2025
A conversation with Kshira Saagar, Wolt’s Head of International Analytics & Data Science
Meet Kshira, Wolt’s Head of International Analytics and Data Science. He started at DoorDash as Senior Director of Analytics & Data Science, where he built the international data science team that works across Canada, Australia, New Zealand, Germany, and Japan.
Today Kshira leads a team of 70+ data scientists at Wolt and has been driving a fundamental shift in the way data should support the business. Forget the endless “Hey, where’s my dashboard?” requests. Now it’s all about digital detective work, being close to the business and seeking “the truth” – how to best leverage data for the benefit of Wolt’s customers, courier partners, merchants, and employees.
Jack Nicholson once said, “You can’t handle the truth!”. Read on to see if Kshira – and Wolt – can.
Tell us about yourself and your career prior to joining DoorDash and Wolt.
“I studied maths, stats, and computer science in university and have always been fascinated with solving problems. Luckily for me, the analytics field was taking off just when I was starting my career, and that’s where I landed. Now I try to find solutions to problems that companies have.
I’ve always thought of myself as a “lost and found” guy. Data is one thing, but I mostly get excited about understanding why something is the way that it is. It’s high-tech detective work where I can connect the dots and solve problems for the benefit of our customers and the business.
Before joining DoorDash, I worked as Chief Data Officer at a bank and Director of Data Science for a big e-commerce company. In those positions, I learned a lot about leading and influencing people, but ended up being quite detached from the hands-on side of the data world in those roles. So when the opportunity at DoorDash came around and the challenge was “you have no people in your team, you’ll have to build everything: the team, vision, and processes from scratch, and there are a lot of problems to solve”, I got excited! Starting all over again was fun and I’ve learned tons along the way.”
You led the international data team at DoorDash first, then joined the Wolt side to lead our data organization. What have you learned so far?
“It’s truly been an incredible journey since joining Wolt 1.5 years ago! On the DoorDash side, our team has a great capability to hyperfocus on the huge market of the United States. At Wolt we have to focus on 28 markets all at once and ensure that solutions serve our business globally but can also be localized when needed. These two different perspectives have taught me so much!
What I’d say DoorDash and Wolt have in common is that our leaders have a really deep understanding of the business. At DoorDash we call this “operating at the lowest level of detail” - everyone from leadership to leads and team members truly aspire to know details so we can make progress on the things that are critical to our business and customers. I’m amazed at how well our leaders run our company strategically while still grasping what’s going on in different pockets of the business. This leadership style has inspired us to build our data team in a way that they get to be very close to the business and the “action” so they can see the tangible impact of their work.”
How would you describe the data organization’s vision at Wolt?
“We want to be true partners in truth-seeking, rather than assistants who pull numbers for our business teams. By true partners, we mean that we want to partner with our teams on a deeper level, rather than just provide data points or dashboards when requested. Together we can find out if the solutions we created really worked, why they worked, how they worked, and how much incremental impact they brought.
Truth-seeking means that we collaborate closely with different teams – and each other – to keep everyone honest. The fact is that we all make mistakes and sometimes what intuitively might seem to be the best decision isn’t necessarily the most optimal one. The “truth” only emerges afterward. And that’s okay. If you think about it, when we make mistakes in life not everyone will tell us, probably only our best friend or partner will feel comfortable enough to do so. We hope to partner with our teams so that we have that kind of two-way relationship where we keep each other honest and accountable. This inevitably leads to better results for the teams and the business overall.
We’re still a young team and we moved to this partnership model only a year ago, so we’re still early on this journey. This is why we’re looking for people who are willing to help drive this change in our ways of working.”
Find out more about the vision of the team in the podcast 👇
What is the role of a data scientist at Wolt?
“Well, they figure out how significantly the changes we make – be it for products, processes, or pricing – impact our business. We do it mainly in three ways: experiments, retrospective analyses, and finding out metric relationships. Experiments are the first and most commonly used way to figure out if the impact is truly incremental or not. However, not everything can be experimented with, so in those cases, we do retrospective analyses – we try something out and then see what happened based on the chosen actions.
The final method is metric relationships and heuristics. We give rules of thumb, or you could call them cheat sheets, to our teams. Based on our analyses we get lots of useful insights, so we package them in a digestible format for our teams. For instance, it could be something like “on average if you improve delivery time by one minute, you’ll be able to improve your topline by X%”. These cheat sheets can help our business teams prioritize certain projects and goals.”
In a company of Wolt’s size, I guess your team has their hands full with problems to solve?
“You bet, but again this is high-tech detective work! Our data scientists support five domains: the consumer side of our business, merchant side, courier side, new verticals (e.g. new business lines), and our support team. Each area has lots of interesting problems to solve!
On the consumer side, our data scientists are focused on optimizing the user experience on our Wolt app. Our app is used by over 45 million customers across 29 countries, so it’s critical that our team works on optimizing and personalizing the experience for various groups of audiences, each with their unique needs and wants.
We’re a delivery business at the end of the day, so our courier analytics team must understand how to optimize supply (the amount of courier partners doing deliveries) and demand (orders from our customers) at any given moment. As we operate in over 500 cities, and have over 260,000 courier partners, supply and demand is not an easy task to predict and optimize!
On the merchant side, we’re building what we call selection intelligence. It tells our sales folks which merchants we should go after in specific markets and estimates the business impact of these potential partnerships. This means our sales folks can spend less time searching for new prospects and more time strategizing how to close these potential partnerships.
Those are our big three audiences. But there’s more. One example is Wolt Market, our very own grocery store, which includes a completely new set of problems like: How should we price new items? What items go on discount and when? How do we ensure our inventory is well-balanced so products don’t go to waste?
And finally, we also support our own employees. How can we build capabilities and solutions so our support folks are best equipped to keep customers, courier partners, and merchant partners happy? We work on problems like, for example, what should a refund amount be if something goes wrong.”
What are the ingredients for a successful Wolt data scientist?
“Curiosity, feedback, and ownership. Firstly, we look for people who have innate curiosity. It’s incredibly important because it shows you’re interested in solving a problem, rather than just solving a problem because someone told you to.
Secondly, it’s important that our data scientists are willing to take on feedback and adapt accordingly. At Wolt, we have a strong feedback culture and even though we’re in the data team, we never just blindly look at data to give us direction. Our analysts need to always be ready to take on feedback and discuss the solutions with team members.
And finally, we want our data scientists to take ownership of the problems they solve. If you don’t feel like you truly own the area you’re working on, it’s easy to say it’s someone else’s problem.
Beyond these, everything else like business knowledge and technical capabilities can be learned on the job.”
What’s the most exciting part of a data scientist’s work at Wolt?
“We partner very closely with our business stakeholders and work on solving problems together, meaning our data scientists directly see how their work impacts the business. This kind of proximity is rare in businesses of our size.
You also get to hyperfocus on the problems you’re solving. In our data team, there are three areas: Data Science, Analytics Engineering, and BizOps. Data scientists are embedded in their respective business functions and get to hyperfocus on the key problems to solve in that business area. They’re supported by the Analytics Engineering team, who create “the source of truth”: the datasets, assets, and dashboards for our business to use. And finally, as we work so closely with our business teams, we’ve just been building up our new function called BizOps. In BizOps, our data analysts work closely with our operational teams to help solve for local business problems taking into account the local context of that market. Having this function enables us to make focused, local analyses based on the detailed knowledge of each country we operate in.”
We talk a lot about craft excellence at Wolt. What does that mean for our data teams?
“Craft excellence is about knowing what to build and what not to build.
Often people tend to focus on certain things that they really love, and want to solve problems in those areas. But it’s critical to sense-check ourselves on a regular basis - are we really solving the right problems? Is this really helping the business? When we focus on that, we build things that truly matter for the business and that helps us all move forward.”
What are you personally most excited about when thinking about the future of Wolt?
“I’m really excited about the growth that we’re experiencing - both in terms of our business growing, but also in terms of the problems we get to solve as a result of that. We’re already in 29 countries globally and keep expanding and launching new products and services. There’s still a lot of scope to grow and exciting things to build.
From the perspective of our data team, I’d say it’s a very exciting time to join. We’ve tripled the size of our team over the past year. Joining at this point means that you get to truly grow together with the team and the business. Our growth brings with it many opportunities for our people, and we have a lot of team members who’ve grown from individual contributors to become leaders, switched roles within the team, or even joined from elsewhere in the business.”
