Contents
This case is part of the Chmltech Business case series. It is written for discussion, in the style of a business school teaching case: it stops at the decision and leaves the outcome to a closed section at the end, so you can decide first. The decision belongs to Jeff Bezos, Amazon's founder and chief executive, in January 2005. It is drawn from public reporting and company statements listed under Sources; it is not based on interviews or private documents.
The decision
A few weeks before Amazon was due to report its results in February 2005, Jeff Bezos asked a small group of employees to meet him on a Saturday at the boathouse behind his home. He wanted a shipping membership: customers would pay one annual fee and get fast shipping on everything, with no minimum order. He wanted it announced with the earnings.
The finance team had modelled it, and the results were, in Bezos's later word, horrifying. The customers most likely to join were the ones who already bought the most, so Amazon would give away shipping it was being paid for today. Two-day shipping was expensive to provide. Nobody could show from data that members would buy enough extra to cover the cost. Bezos had to decide whether to launch anyway, and at what price.
Company background
Amazon had started as an online bookstore in 1995 and by 2004 sold electronics, toys, kitchenware and much more, with net sales of about $6.9 billion. Its central idea was that lower prices and better service would bring more customers, more customers would attract more sellers and selection, and the growing volume would lower costs, allowing lower prices again.
Shipping was the weak point in that loop. Shoppers disliked paying for it, and many abandoned their carts when they saw the charge. In 2002 Amazon had introduced Free Super Saver Shipping, free standard delivery on orders above a minimum, and lowered the threshold over time. It worked, but it trained customers to batch their orders and wait, which is the opposite of the habit Amazon wanted them to form. Fast delivery was available but rare: two-day shipping cost about $9.48 for a typical order, and few customers paid for it.
The situation in early 2005
An engineer, Charlie Ward, had suggested a premium shipping club, borrowing from airline loyalty programs. Bezos took the idea and set a deadline that the team considered impossible. The project, code-named Futurama, pulled people from other teams to build the sign-up flow, the eligibility rules and the changes to the fulfilment centres needed to promise two-day delivery reliably.
Exhibit 1. The proposal on the table
| Element | Proposed design |
|---|---|
| Annual fee | $79 |
| Benefit | Unlimited free two-day shipping on eligible items, no minimum order |
| Extra | Overnight shipping for $3.99 per item |
| Eligible items | More than a million items sold and shipped by Amazon |
| Household sharing | Up to four household members on one membership |
| Comparable cost to the customer | About $9.48 for each two-day shipment |
The economics depended on three things nobody could prove in advance. First, members had to change behaviour and buy more often, from more categories, once shipping cost nothing at the margin. Second, the operations team had to bring down the cost of two-day fulfilment as volume grew. Third, the fee had to be high enough to matter and low enough that customers would try it. According to Brad Stone's account, Bezos settled on $79 because it was large enough to feel like a commitment but small enough that customers would take the risk.
The options
Option A: Launch Prime at $79 for everyone
Announce it with the earnings and accept losses on heavy buyers, betting that changed behaviour and lower fulfilment costs will pay for it over years.
Option B: Run a quiet pilot first
Offer it to a sample of customers, measure the change in spending for six to twelve months, then decide. Lower risk, but slower, and a pilot this visible would be hard to keep quiet from competitors.
Option C: Price it higher
Charge $99 or more to protect the margin. Fewer customers would join, and the habit Amazon wanted to build would spread more slowly.
Option D: Improve Super Saver Shipping instead
Lower the minimum order or speed up free standard shipping. Cheaper and safer, but it does not create a reason for customers to come back to Amazon first.
Discussion questions
- Finance said heavy buyers would join first and cost the most. Is that a reason to stop, or the point of the program?
- Use Exhibit 2. What would you need to believe about member spending for Prime to pay back within three years?
- Which part of the bet was a marketing question and which part was an operations question? Which was harder to win?
- Why might a fee customers pay up front change their behaviour more than free shipping with a minimum order?
- If you ran Prime as a pilot, what would you measure, and how long would you wait?
What happened next
Decide what you would do before you open this section.
Show what the company did and how it turned out
Bezos launched. Amazon Prime went live on February 2, 2005, announced as "all-you-can-eat" express shipping for $79 a year, with unlimited two-day delivery on more than a million items. Tens of thousands of customers signed up in the first month. Press coverage was muted and focused on the cost to Amazon.
The early members were mostly existing heavy buyers, just as finance had predicted, and the program lost money at first. By 2007 internal data showed that customers who joined Prime roughly doubled their spending on the site. Amazon then used Prime to pull in sellers: in September 2006 it launched Fulfillment by Amazon, which let other merchants store goods in Amazon's warehouses and make them eligible for Prime shipping. That made Prime's selection grow faster than Amazon could buy inventory itself.
Exhibit 3. Prime after the launch
| Year | Change |
|---|---|
| 2005 | Launched at $79 a year |
| 2006 | Fulfillment by Amazon opens Prime shipping to other sellers |
| 2011 | Prime Video added |
| 2014 | Fee raised to $99 |
| 2015 | First Prime Day |
| 2018 | Fee raised to $119 |
| 2021 | More than 200 million members worldwide |
| 2022 | Fee raised to $139, citing higher labour and shipping costs |
Prime became the centre of Amazon's retail business: the main reason customers start their shopping there, and the reason sellers accept its fees. The operational bet mattered as much as the marketing one. Each new fulfilment centre and each step toward faster delivery lowered the cost of the promise and raised its value.
Lessons for small and mid-sized businesses
Some decisions cannot be proven with data in advance. When a choice depends on how customers will change, a model can only show the downside. Decide how much you can afford to lose while you find out, and set a date to review it.
A promise is only as good as the operation behind it. Prime worked because fulfilment could deliver it. Before you promise faster delivery or shorter lead times, check your quote-to-cash process and capacity can hold it.
Subscriptions change behaviour. A small business can apply the same idea with a service plan, a maintenance contract or a priority-delivery membership: an up-front fee that removes a per-order cost the customer dislikes.
Price to be tried. Bezos priced Prime to feel like a commitment, not a burden. A first price can be raised later, as Prime's was four times; a price that stops people trying cannot be fixed.
Sources
- Amazon.com press release, results for 2004 and the introduction of Amazon Prime (February 2, 2005)
- Retail Brew, "How Amazon Prime backed the retail industry into a corner" (2025)
- Benzinga, Jeff Bezos on Prime's risky beginnings ("the results were horrifying")
- WRAL / CNN, the boathouse meeting and the $79 price, after Brad Stone's The Everything Store
- Phys.org (AP), "10 years later, Amazon celebrates Prime's growth" (2015)
- Commoncog, Amazon Prime case (launch press release, 2007 spending data)
- Bloomberg oral history via The Lowdown Blog (two-day shipping cost, Charlie Ward, Futurama)
- Wikipedia, Amazon Prime (membership counts and fee history)
More cases
- Costco Case Study: 4,000 Products and a Capped Markup
- Shopify Case Study: Build, Buy or Exit Logistics?
- Apple Case Study: From Months of Inventory to Days
- Starbucks Case Study: Closing Every Store for an Afternoon
- Toyota Case Study: How Lean Is Too Lean?
- Lululemon Case Study: The 2013 Yoga Pants Recall
- Netflix Case Study: Splitting DVD from Streaming
- Tesla Case Study: A Cheaper Car or a Robotaxi?
