Hagar Rips

(part 1/3) - 2015 Before launching Foodfox in 2015 (now known as Yandex.Eats – a hyper-local delivery service from restaurants), Sergey Polissar and I spoke with many very smart, very experienced, and highly relevant people to better assess the viability of the business model. All of them (literally all) explained that this business model cannot work for various fundamental reasons: 1) The restaurant market is very hard, because restaurateurs are old-fashioned people from the 90s, it’s impossible to sell them any new service 2) The market is not big enough for two players 3) Unit economics won’t work — not enough density of orders to make it efficient + restaurants won’t be ready to share their margins 4) No moats — customers will easily switch between different apps Of course, we weren’t sure ourselves, as all of that was relevant, but we felt we had to try. And a lot of things — timing, people, luck (probably the most) played their part. Today Yandex.Eats is a highly profitable market leader with $7B GMV, operating in 12 countries. (part 2/3) 2019 In 2019, when we were already part of Yandex, we launched Y.Lavka — 15-minute groceries delivery from our own dark stores. In 2020 this model became super hyped globally, and dozens of similar companies were launched around the world. A few years later almost all of them went bankrupt. Again, very smart and very experienced people explained the fundamental reasons why this model doesn’t work: 1) Unit economics doesn’t work because of low average check and low gross margins 2) The model works only in a small number of dense locations 3) Customers need a wider choice of SKUs 4) No moats All of that was true. But in our context, with our specific setup, the model started working. (note: it is unfair to compare a launch backed by a corporation vs an independent startup as it makes things so much easier - you get resources, people, brand. But global peers raised billions of dollars, which makes them comparable). Anyway, it was a combination of execution, geo and specific set up on our side that worked out. Today Y.Lavka is $1.5B in GMV with 500+ dark stores, expanding to new geos. (part 3) is in the first comment

153

20 Comments

For those who don't know, one of the biggest bottlenecks in robot deployment is collecting the right data to generate successful AI policies on robot. we are seeing this learning change at breakneck speed. With Generalist and SkildAi coming out with their "in context learning" to now what Machenta is calling "Instant Robot Learning". Time will tell how these new models evaluate against each other but Machenta claiming zero real world data collection and training only in simulation is a bold but interesting claim. if true, it means those companies focussing on building the datasets and the processes for generating successful ai models may need to re-think the entire process. Exciting stuff. Well done guys!

1

Market dominance doesn’t come from competing harder. It comes from redefining the game. Three powerful lessons for dominating the market from Google’s early days, shared by its first product manager, Aydin Senkut, the founder of Felicis Ventures, and championed by the founders, Larry Page and Sergey Brin are focus, doing hard things, and being exceptionally different. He said one of the most brilliant things they did was to launch Gmail. And that project taught him three main lessons. 1. Focus He said Larry was adamant about focusing on very few things. His favorite word, 90% of the time was no. When he said yes, you knew it was something that mattered, and everybody rallied to make that possible. 2. Difficult Goals Sergey always asked for very challenging goals. And you can’t build a truly great company if you strive for a 10% or 20% improvement. You must disrespect the status quo. Sergey told the lead engineer in the ad serving team, who wasn't a specialist in ads serving, to pick 5 engineers and build a functioning ad server in 6 months. And they did. In a large company that world require hundreds of engineers, and probably a couple of years to build. 3. Exceptional Difference They operated the 10X principle. If you want to be really great at whatever you do, you have to be 10 times better, 10 times cheaper, or 10 times faster, or at best, in all three areas than the competition. Transformational companies are built by teams willing to question every assumption about cost, speed, quality, distribution, and even what customers believe is “normal.” They don’t optimize the status quo; they replace them. People with market domination mindsets look for ways to make the old way obsolete, and every alternative irrelevant. That’s how real domination happens.

5