Amazon Accelerate 2025: Fireside Chat with Andy Jassy and Dharmesh Mehta

It's great to be here with all of you. Thank you for having me.

First of all it's great to be here with all of you. Our relationship and our partnership with you is so important for us so it's a thrill to be able to spend time with you. In retrospect, I think the way it's worked out, it seems blindingly obvious that it was great for customers for us to build these partnerships, but at the time that we made the decision to open the store to independent sellers, it was quite an animated debate inside of Amazon. And that's because the entire company had been built the first couple years on the premise of storing items in our own distribution centers and fulfilling them ourselves. And we just weren't sure if we went some other route, what kind of impact would that have on our relationships with publishers and distributors and manufacturers and could people fulfill items in a way that would satisfy customers the way we thought was required and the brand standards we had set. But we started seeing these marketplaces like back then Hafta.com, eBay, Rize and we had this debate whether we should open up the store to other sellers. And after a lot of debate and this took several weeks, we decided that you can't really fight gravity. The reality is that if we had more sellers in the marketplace, there would be better selection for customers and better prices for customers. And the reality is customers want that -- you know, if you look at the cornerstones of a retail experience, there are a lot of things that change. But customers never tire of broader selection, lower prices, and faster delivery. So, we decided that we would do so, and we had some starts and stops along the way. The first thing we tried to do was build an auction site. It was too much of a me too like what eBay had and that didn't work very well for us. And then we tried something that we call Z shops back then, which really was a cul-de-sac on the site where we allowed sellers to sell product, but that didn't work either because nobody went to the cul-de-sac. That's why it was a cul-de-sac. So, it wasn't until we put our independent sellers' products on the same product detail pages as our own first-party selection that customers started interacting more frequently with that selection, and that it became a success for our sellers and for our customers and for us as well. And I think in retrospect, it seemed like an obvious decision. At the time it wasn't. It's really hard to get a culture to change and one that was doing well at that time. And then it's hard to execute it well. And I think it turned out to be a good decision for our customers and for independent sellers and for the company. And today over 60% of our units that we sell are from independent sellers. That's right.

And if you look over the last 25 years, about 2 1/2 trillion dollars of sales has been made by our independent sellers. If you look at the number of employees our independent sellers have employed over that time, it's over 2 million. So, together we have created a lot of jobs. And I still think it's quite early in our adventure together.

Because we built Amazon.com as quickly and as substantially as we did the first several years, we had thought that we had built a bunch of these building block foundational pieces that would let us scale. These are things like search and browse and item catalog and payments and fraud and things like that. And we kind of went along thinking we had built this very strong foundation and then we tried this business called merchant.com. And what it was was allowing third-party merchants like a Target or a Marks and Spencer's to use Amazon's e-commerce technology and fuel their own sites and to deliver those capabilities to the retailers because they wanted all the key building blocks and then they wanted to customize them. And it turned out, and they wanted all those components through application programming interface or APIs, and when we went to deliver the pieces to those merchants, it was way harder than we had ever anticipated because we actually hadn't built those capabilities as foundational building blocks. They were really kind of jumbled together as we were moving so quickly, scaling Amazon for several years. So, one of the things we had to go back and do and it was really painful for us, was we had to go back to all those building blocks and rebuild them as foundational primitives and around the same time we were in the process of defining and building Amazon Web Services which is our cloud computing set of services and I was part of that team that was working on it at that time and we were watching what was happening on the merchant.com side and said we are not going to do that. We are going to actually build these as separable building blocks that we called primitives and a primitive service to us is something that does something particularly well, it's a simple thing that does it very well. It has -- it's not overly paternal, which means we don't make all the decisions for people that are going to use it. And it has a lot of flexibility. So, in the AWS space, the primitives we built back then was an object storage service which became S3 and a compute service that became EC2. And a set of database services, relational database service and a nonrelational database service. And authentication and things like that. So, this concept of building in these low-level primitive building blocks became something that was really at the core, it was a tenet of everything that we went and built. And you can see how we have evolved that here in our retail business and for our partnership with all of you. We realized that we had so many independent sellers who really found it time consuming and expensive to have to store their items, to pick, pack and ship them and to deliver them to customers. So, that's part of why we built fulfillment by Amazon which is a premium service as you all know with very fast delivery that's about 70% less than alternatives for a premium service like that. And then we had a number of customers and independent selling partners who said, I have my own websites, too. And I really would love the opportunity to let your Prime customers be able to buy using the Prime brand and that fast shipping. So, we built Buy with Prime which also has been early on something that a lot of our independent selling partners have enjoyed. And if you look just across the board, every time that we build something to scale Amazon, we are then trying to build it in such a way where we can turn around and offer it to you. So, we have a lot of customers who have to take items through the border, so we built a service that makes it easier for you to do that. Then you need freight to get it to various locations so we built capabilities that do freight. We have a lot of our independent sellers who want their inventory to be upstream of our fulfillment nodes and so we have built those upstream warehouses that you can use. We have something called Multi-Channel Fulfillment, you know, which makes it much easier for you to store in one place and ship through all the marketplaces that you use to ship to customers, that we were able to get items to people 50% more quickly than you can from other retailers. So, every time that we can find foundational building blocks that we find help us scale, we are going to build them so we can scale quicker for you but then we are going to also try to turn them around and make them services that you can enjoy that make your business more customer friendly and more cost effective.

Yeah. I think that Amazon is very much a why culture, and what I mean by that is that it's not top-down driven. It's not my way or the highway. It's not like the senior person on a team says this is what we are doing and everybody says hup2 and does it. It is a place that has ought not just encouraged but asked teammates to do ask why. Do we have to settings for these constraints. Why must it be done the same way it's been done the last 20 or 30 years. And asking those why questions, you know, it can be annoying sometimes for people. You get a lot of people asking why. And the point is not to annoy people, but it's really to get at the root cause of why things have to be the way they are and to really try to work around the constraints that everybody has accepted for a long period of time, but aren't great for customers. And I think somebody -- so, we thought about that as having the why culture. There's something, there's IQ, there's EQ, this is kind of YQ and somebody who has good YQ is not just good at asking questions, but asking questions the right types of questions that get you down to the constraints that help you work around those constraints, and somebody who is not shy about continuing to ask questions when you haven't really gotten to the answer. I mean, there are many times, I'm sure where I feel like I have asked the question four or five different times, I have asked the question, and I'm calculating in my head and I just say, I don't still understand why we can't do something different than this. And then a lot of people will say, okay, this is too annoying to ask another question. But you have to keep asking questions until you really get to what that root issue is that you can't get through. And then somebody with high YQ also knows when you're at the point of diminishing returns there are times where you have asked all the questions you can ask and either you have exhausted what you think you can do to work around those constraints or there just isn't more to ask, where you do have to move on and make decisions. We have this concept inside Amazon of one-way and two-way doors. And the way we think about it is that the overwhelming majority of issues that we all face in our jobs are two-way doors which means if you make the wrong decision, after you have walked through that door, you walk back through it and not much harm. And there are a small number of one-way door decisions that you have to make more slowly, more methodical because if you get it wrong, it's very difficult to walk back. But both of those ideas of two-way and one-way doors assume that the door is unlocked. And the reality is that when you are inventing, you are often trying to work your way around doors that are locked. And if you have a why culture and you have high YQ and you ask enough questions the right way with a smart group of people, you have a much better chance of unlocking that door and inventing.