The Product Model at Amazon
By Marty Cagan and Richard Russell
Marty’s Note:
Recently I co-authored an article on The Product Model at Spotify with one of Spotify’s early product coaches. This proved to be one of our most popular articles, and many people shared that it answered many questions they had about how that company really worked; what was important, and what was just incidental.
In this article, we hope to do the same by describing how the product model manifests at Amazon. My co-author in this case is the European-based transformation coach Richard Russell. Richard had a long career leading product and engineering, first at Google and later at Amazon, and now he helps other companies in their efforts to move to the product model.
Introduction
Many companies have influenced what we have come to refer to as the product operating model, but none more so than Amazon. If you’ve never taken a look at Amazon’s Leadership Principles, hopefully what you see there will look very familiar.
Amazon’s record of consistent innovation, at scale, over the past 30 years, has been nothing short of remarkable. You can debate the policies of any large company, but it’s hard to argue with Amazon’s skills in creating innovative tech-powered products and services.
In this article, we want to try to illustrate how the product model manifests at Amazon through the origin story of Amazon Prime.
We decided on Prime because Amazon Prime did for retail what Google AdWords did for advertising, and what the Apple iPhone did for personal computing. Which is to say, it completely disrupted the industry, and changed the course of the company. And we think most of the industry does not yet appreciate the true impact of Prime.
Many people have already shared the story behind Prime (our favorite telling is from the excellent book Working Backwards: Insights, Stories and Secrets from Inside Amazon, by Colin Bryar and Bill Carr), so we are not trying to repeat what has been shared by others.
Rather, instead of trying to highlight what is unique about Amazon, we hope to illustrate how Amazon’s approach to building tech-powered products is in fact very consistent with what we see in each of the strongest product companies.
As with Spotify, every company uses their own nomenclature to describe how they work, and views everything through the lens of their culture, which can make it hard for people to know if Amazon has some sort of truly unique way of working, or if they are just especially good at working like the best.
For this article to make sense, you’ll want to have at least a high-level understanding of what the product operating model is, as this article will be comparing Amazon’s model to the product model. If you haven’t yet read the series of four articles, you should probably start here.
Background
Amazon was founded in 1994, and ten years later, in 2004, their growth rate was slowing down. It was still quite high, but they knew the market was still nascent, so they felt they needed to maintain a higher growth rate.
The company at this point was still primarily a retailer (AWS would not happen for a few more years). The company’s business strategy was “to relentlessly focus on customer experience by offering our customers low prices, convenience, and a wide selection of merchandise” – that is, compete on price, selection, and convenience.
How They Decide Which Problems To Solve – Product Strategy
Amazon has long understood the power of a compelling product vision, and today the PRFAQ is their go-to vision technique for meeting two important goals: first, ensuring that the product vision is customer-centric; and second, helping to align multiple product teams to all pull together towards the same common goal.
Similarly, the company has long understood the value of data-driven insights powering a strong product strategy. At Amazon, these product strategies are usually captured in what they refer to as “six-pagers” – written narratives that provide the evidence and the reasoning behind proposed product work – and then intensely debated.
In order to boost growth, the company had experimented with advertising, and this did provide a modest uplift, but the leaders believed better and more sustained ROI would result from improving the customer experience.
When they evaluated their market position, they believed they were quite strong on price and selection, but there were several critical areas where convenience was not where it could or should be.
In particular, they had both quantitative and qualitative evidence showing that the convenience was significantly limited by shipping – both the cost, as well as the service itself (the speed and reliability) – and so this became a multi-year focus of their product strategy.
More generally, it’s important to recognize that this focus on shipping meant that there were many other things, well beyond advertising, that the company chose not to pursue. In contrast, most companies would have chosen to spread their resources across a much larger set of initiatives.
How They Solve Problems – Product Discovery
When looking at the problem of shipping, beyond the cost, there were two major areas.
The first was from the point where the customer clicked “Buy,” to when the package was handed off to the shipper. The second was from when the shipper received the package to when it was delivered to the door of the customer. Product teams would need to attack both of these areas.
Amazon already had very good data demonstrating how much customers appreciated faster shipping, so value risk was not the key risk in this case. The major product risks they had to tackle were around technical feasibility (just how quickly could they fulfill?), and business viability (can they find a way to do this that doesn’t cause them to need to raise their prices, or destroy the company’s margins?).
To address these risks, it’s important to understand that another central principle in how Amazon works is that when you have difficult problems to solve, you need to be willing and able to run a series of experiments.
And that is very much what was going on. Here are a few of the many experiments that were run:
- Super Saver Shipping – You can think of this as the first early test of Prime, but for orders over $99, shipping would be free, although this would be a somewhat slower-than-normal shipping. Note that successive experiments gradually reduced the threshold down to $25.
- Faster Super Saver Shipping – like the Super Saver Shipping, but rather than slower-than-normal shipping this provided standard shipping.
- Free 2 Day Shipping (Express Shipping) – similar to Super Saver Shipping, but providing expedited 2-day shipping.
- Annual Fee for Free Standard Shipping – this experiment introduced the idea of a subscription to provide the member standard speed shipping at no extra cost.
The different experiments worked better on certain customers than others. For example, the free shipping worked well for price-sensitive customers. And having a minimum threshold price worked well for Amazon. But for price-sensitive customers, they might choose to go to a physical store, even if that meant reduced selection and slightly higher pricing.
But Amazon’s best customers – the ones spending the most money – were still paying extra for faster delivery when that option was offered for extra cost, and not benefiting from most of these experiments.
While these various business model experiments were going on, one of the engineers, Charlie Ward, realized that the discounting and promotions algorithm was at this point suffering from significant technical debt, and had become distressingly complex, to the point of not being able to meet the evolving needs of the customers and the business, and he believed he needed to both simplify and enable new pricing models. He also saw that the DVD rental team had built a subscription capability, so he put these together and suggested a new subscription for free shipping.
While all these experiments were going on, and new enabling technology was emerging, the leadership, including and especially the founder Jeff Bezos, was considering the implications of this progress, and how this might impact the product strategy.
While initially conceived of as a way to improve convenience through better shipping solutions, Jeff looked at the results of more than two years of ongoing experimentation, and now saw the potential as something larger, a free shipping program as part of a larger customer loyalty program.
The challenge with this new, expanded product was that it would be very difficult to model how the new offering might impact buyer behavior over time. Would people buy more, as they hoped? How much more? Or would they buy the same amount, but now Amazon’s margins would be lower because of the higher expedited shipping costs? And how might shipping costs change as volume rises?
Yet their mission of providing the best place to shop and to take care of their best customers tilted the scale towards betting on the best customer experience they could provide, so the company decided to move forward with the experiments leading to Prime.
While Prime didn’t look particularly strong as a business case due to the potential costs, it was driven forward because of a belief in the product strategy of betting on price, selection and convenience, and to create an ongoing competitive advantage for taking care of their best customers.
How They Build – Product Delivery
From its early days Amazon has been a leader in how to build, test and deploy.
Building out the productized version of Prime included the leverage of several existing technologies, especially around shipping. They were also able to leverage shipping estimate technology, as well as the subscription services built for the DVD rental program.
Remarkably, the first productized version of the Amazon Prime product was built in roughly three months.
There were four main reasons for this:
1) the leadership decided to immediately put a dedicated, empowered product team behind the Prime program;
2) the leadership also made it very clear that this was one of a small number of critical focus areas;
3) due to the many prior experiments related to shipping, much of the supporting infrastructure was already in place, and could be leveraged for this new offering; and
4) the new Prime PRFAQ clarified what was important and what was not, aligning people, and helping them understand the motivation and strategic context, and plan their work accordingly.
The Results
Today Amazon Prime is recognized as one of the most financially successful product innovations of all time.
The Prime direct revenue (from the subscription revenue) is not hard to understand ($35B revenue from 220M subscribers globally).
What is more difficult to estimate, but even more impactful, is how Prime has changed the buying behaviors of the members. What follows is our estimate, based on publicly available numbers:
- In 2023, the 220M Prime subscribers spent on average $1400 each on Amazon (around 3X that of non-prime members), a total of $308B. This represents roughly 45% of Amazon’s total marketplace GMV of $700B in 2023.
- Moreover, Prime members tend to buy from Amazon by default, so they are less price-sensitive, they comparison shop less, and they cost almost nothing to acquire on a per-transaction basis. All this helps significantly with margin.
Product Culture
All of this comes from strong product leaders and strong product teams, obsessing over customers, having a clear product vision and data-driven product strategy, the willingness and skills to test ideas with experiments, then leverage what you have learned into real business outcomes.
It’s not a coincidence that Amazon is such a good example of the principles underlying the product operating model. But a few points are worth highlighting:
– Strong Product Leaders
Amazon’s model depends on strong product leaders (which they often refer to as Single Threaded Leaders). These leaders may come from engineering, or product, or the business side, but what’s critical is that they all bring strong product sense – deep knowledge of customers & data, their business, and the industry.
Furthermore, as explained in the leadership principles: “Leaders develop leaders and take seriously their role in coaching others.” and “Leaders ask themselves: Are my fellow employees growing? Are they empowered? Are they ready for what’s next?”
In our view, this more than anything else has been the key to Amazon’s ongoing success.
– Embrace Experimentation
Nearly every company today gives lip service to experimentation. But for most of them, they mean optimization experiments. For example, they want to see if changing the call to action from green to blue might cause the conversion rate to increase.
But as you saw in this case, the experiments Amazon was running related to Prime were much more meaningful, and involved taking real steps to address the risks, but they did this in responsible ways. Without the skills and willingness to run these types of experiments, Prime simply doesn’t happen.
– Long-Term Thinking
Consider the payback horizon that most companies use to decide what goes on their product roadmaps. Often one or two quarters, and what’s especially remarkable is how few of those short-term efforts generate even modest returns.
Yet Amazon understands that the efforts that generate the kinds of returns they are seeking typically take longer to achieve the vision. Hopefully you can see how this absolutely does not mean that the company moves slowly. Amazon moves faster than most startups. But it does mean that they focus on the longer-term and are willing to put in the sustained effort to achieve real outcomes.
– True Agile
In stark contrast to the many companies where Agile is more of a process than principles, many of the product teams at Amazon don’t talk much about Agile. However, most of them are extremely agile as measured by what’s meaningful: the teams are truly empowered, and the teams build, test and deploy continuously.
Moreover, it’s not just agile in delivery. Notice how the results of the experiments not only informed the successive experiments, but the learnings would quickly make their way to the product leaders and even to the CEO, where they informed the next iterations of the product strategy.
In fact, it was this rapid learning from the many related experiments that gave the leadership the confidence to pursue an even bolder vision than had been originally imagined. This is real agility in action.
– Focus on Outcomes
The motivation for this effort, and the measure of success, was a clear set of business outcomes (improved price, selection and convenience), representing the value of a significantly improved customer experience.
The product vision and the product strategy were all about ways of achieving this outcome, and the many experiments that were run were zeroing in on this outcome, and the ultimate business results delivered this outcome.
Whereas most companies would have a big annual planning process culminating in a big roadmap of a bunch of features that the leaders hoped would generate the necessary results, at Amazon, they start with the desired customer experience that their data shows them is important, and the entire approach to product is about experimenting and discovering a solution that can deliver the necessary outcome.
While Amazon Prime might be their most famous example, and certainly one of the most financially successful, what is most impressive about Amazon is how many such examples – large and small – that they have created through their 30 year history of leveraging technology in order to better serve their customers.
