Daniel Gebler, co-founder and CTO of Picnic, discusses building an online supermarket that has achieved profitability in the Netherlands and is scaling across Europe. He shares lessons from his e-commerce background, the importance of learning to unlearn when innovating radically, and how Picnic uses AI and robotics to simplify grocery shopping while expanding into recipes and meal planning.
▸ Read the transcript
If you want to do something really radically new, you need to learn to unlearn. So basically all the stuff that you have learned in the past, half of it is great. Also for the future, the other half you need to unlearn because something completely new will be relevant. If you do customer first AI first kind of innovation, then you're always a step ahead from the rest of the pack. We reached profitability for the entire Netherlands last year. Consumers are looking for a service that actually simplifies their life. So we are hiring now people that will be really effective in half a year or a year. So we need to now project a little bit how software engineering looks like in a year's time. think the most important mistakes that people make is that they get too late in action. Hello guys, here we are at TechEx. Again, thanks First Sight Media for the whole setup. And here we are with Daniel Gebler from Picnics CTO. Thanks for having me. Thank you, Daniel. So I wanted to ask some storytelling questions. So first of all, who is Daniel Gebler? Excellent first question. Thank you. So I'm Daniel, CTO of Picnic, an online supermarket that we started a couple of years ago here. I'm originally from Germany. I did my PhD here. Loved the people, loved this country, state. Perfect. You started your career at Fredhopper. And that brought you a lot of learning to Bingit. I was wondering, which was your experience there and which kind of learning do you have? Yeah. So after my studies, I got really excited about e-commerce and basically commerce online. And I wanted to work first in an environment where I could work with many, many kind of retail and other commerce businesses. So Fredhopper was a platform that allows recommendations and product suggestions to all kind of retailers around the world. So I learned a lot about retail. And basically I got three insights. The non-food world is already pretty much online, food not yet. So hence I got excited about food. The second one is real innovation is happening in logistics these days. So the commerce side is already pretty mature. And the third one is if you do customer third one is if you do customer first AI first kind of innovation, then you're always a step ahead from the rest of the pack. Then you did a PhD at the VU. Which kind of learners did you apply from academia to the entrepreneurship? Yeah, probably two interesting learning. So first of all, the first thing is first principle thinking. So any kind of meaningful deep problem can only be solved if you reason from first principles. And that is about logistics, about customer topics, but many, basically everything what we are doing now. And the second one is if you want to second one is if you want to do something really radically new, you need to learn to unlearn. So basically all the stuff that you have learned in the past, half of it is great. Also for the future, the other half you need to unlearn because something completely new will be relevant. And completely new will be relevant. And hence learning to unlearn was one of the key things I learned here. Later on, you co-founded Picnic with also some of the founders of Fredhopper. And the original team was five co-founders. I was wondering how do you manage these relationships, these co-founding relationships when there are more than two or three co-founders, which is? So if the ambition is large enough, there are basically two settings. Either you have a very dominant single person, that's how the entire strategy for everything can work. But actually you're falling short on many topics. Or you go to path that you have a couple of really strong domain experts. So somebody for commercial, somebody for logistics, somebody for tech that is really focusing on those specific kind of domains. In a way it really understands the cutting edge and goes beyond that. And then basically builds a new concept. We have opted for the second one. And so far it has worked out pretty well for us. So far Picnic, I think this year was a bit positive, but it's not being profitable. What does it mean to demonstrate that the business model works in Picnic? So you need a couple of steps. You need to go through a couple of steps when you're building such a business. So the first one is you need to show actual demand. So that's the first thing two customers like what you're offering. So this is where the first few years. And then you basically have a lot of growth. The second one is you're looking into can you make at some point money with it? And that goes in different stages. So the first step is can you make on a single city basis some profit. We made already very early profit on single cities. And then over time you go into regions and then an entire country. And exactly as you said, we country. And exactly as you said, we reached profitability for the entire Netherlands last year. And for the entire Netherlands last year. And the other countries are able to follow very soon. How do you de-risk such a consumer behavior disruption when you are innovating and also when you have different cultures on the making as well? Like you are expanding to France, you're expanding to Germany, How do you de-risk that? So when we have been thinking about building and especially scaling a business like Picnic, we were thought very carefully about what is the scalable core. So basically what are the kind of first line principles, first order principles that will apply across all countries. So one thing is consumers want to have great prices want to have fantastic service at the door want to have also always fresh products. So if you combine that, then you have already key elements of a proposition. The second one is also consumers are looking for a is also consumers are looking for a service that actually simplifies their life. So hence we move simplifies their life. So hence we move from a single product delivery proposition into a delivering recipes and meals and many other kind of things that actually take cognitive load away from deciding what to do ever to eat tomorrow or the day after tomorrow, etc. And if you combine that, then you end up in a concept like Picnic that consumers love and then also will expand very fast. What's something unexpected that you saw from expanding in other countries? There are many things that are actually the same between countries, but there are also very nice and subtle differences. For instance, in Netherlands, we started with a concept of a waiting list. So basically consumers couldn't immediately enter the store, but they had to wait a little bit until they could join the store. That worked in Netherlands pretty well. Consumers understood that we needed to do step by step scale up the service and it made it even a bit more attractive. In Germany, we tried the same, but what has happened in practical terms here is that Germans didn't like it too much. So they are very binary. Either you can immediately try it or you just don't have a second chance. So hence the waiting list concept didn't work that well in Germany. How is Picnic differentiating themselves from traditional supermarkets that are catching up on this online trend? So if you look a bit into who else is offering a service like Picnic, then obviously there are many other offerings that you have in the market. So physical supermarkets are providing something. You have online restaurant deliveries. Restaurants themselves are basically fulfilling the same need because the need is you need to feed basically your families. But if you look a bit into what is a very convenient solution that at a fantastic price point with a great service and frictionless fulfills the service, then actually you end up with not too many options. And we believe that we are probably one of the stronger options in the market. And our consumers are telling us this basically every day, also that we should further invest in that kind of proposition. And is Picnic acting towards new methods of product discoverability? Since you don't have like a traditional supermarket, the stands, you cannot see the product, how is Picnic trying to innovate in that? We have actually chosen a very simple for discovery of what are new items to list. If you go in a category in our shopping app and you don't find what you're looking for, then basically you have a small button where you can say, can I ask Picnic to also list that kind of item? The same happens also if you search for an item, you don't find it in the store, you simply ask Picnic, can you basically list this? And if there are enough customers asking for it, so typically three to five in a reach, then we are very happy to list this. And now we are in this trend of agentic AI. How is agentic AI changing the way Picnic operates? So agentic has many, many opportunities for business like Picnic. So obviously the entire product and software engineering goes just faster. So that is a fantastic first kind of dimension. A second dimension is that what we are offering as a product proposition, basically what you can do in the app, what can we do in the warehouses, is actually AI powered to a large extent, which makes it also more convenient, faster, more efficient, etc. But the actual trick is that we do also innovation on a lot of kind of business processes, for instance, to identify what are the next products to put in promotions, what is a nice price strategy, what is basically a nice visual to set, etc. So hence, we are innovating step by step on the proposition perspective. With respect to agentic commerce, that has been also discussed a lot these days, we don't see yet a lot of agents on our side. So there's obviously a few people are experimenting with it. But this is far from being a mainstream proposition. So we are certainly we are looking at a bit how this space develops, we have prepared ourselves that we can very quickly also launch the solutions here. But we are definitely not yet pushing this too much. Following up on physical AI, which are some examples of disruptions that happen in a picnic right now, or plan for the near future. So we started our order picking process with a manual fulfillment center. So basically, in a fulfillment center, you're picking them manually by hand, the products for customers. Step by step, we automated those processes with robots and other tools. And over time, we have also learned to use much more agentic methods to automate and robotize those kind of warehouses. So we have now quite a few picking robots, we have robots that are storing the products that are making quality control that are making counting of products, many, many things, where you see that most of the processes are now always a combination between human doing a task and the robot. And actually, when this is a bit of a surprise, most people like it. what are the challenges of beginning right now? Yes, there's a lot of kind of challenge that such businesses need to go through. One of the interesting kind of open topics that we are working on is how will especially agentic commerce develop further? Will that become a kind of another sales channel or this may be more kind of a borderline topic? But probably the most interesting and the biggest topic that we're working on is how can we make it even easier to organize the shopping for an entire week? And if you look about shopping for a week, then you look also about what is in the calendars of the kids or of the family, what you should consider, what are your nutritioning goals? Maybe what are your fitness goals? How you can combine with kind of your fitness workout routine, etc. Basically combining it as a kind of a real society app and not just a kind of a shopping app. How is keep hiring right now for Picnic now with this new agentic AI trend? How does Picnic select the new software engineers? The big challenge is that software engineering is developing so fast that the way how we develop now software is very different than a year ago and hence in a year most likely will also be a very very different. So we are hiring now different. So we are hiring now people that will be really effective in half a year or a year. So we need to now project a little bit how software engineering looks like in a year's time. And what we have in a year's time. And what we have focused on is basically two aspects. One is people with the right mentality that are willing to learn can learn and unlearn again. And the second one also highly analytical people that are able to really structure big problems and to divide and conquer. So to divide it in small problems that you can solve in an easy way and then compose the solutions to bigger solutions. And now some personal questions to end. If you had your younger self, I don't know maybe just before Fredhopper, which piece of advice could you tell him? I think the biggest advice that I would give everyone and also to my younger self is stay curious, stay bold and don't hesitate to take action. I think the most important mistakes think the most important mistakes that people make is that they get too late in action. And last late in action. And last question, if you could meet somebody either from the past or nowadays, who will this person be? I think meeting Steve Jobs would be pretty cool. This is just a kind of a hero in the way how you design actually products that shape an entire industry. Amazing. Thank you, Daniel. Thank you.
