Amazon Marketplace Case Study: Deliverable for Submission.

Write Step 4, The Problem and Step 5, The Solution of the Press Release for the proposed Seller Fulfilled Prime (SFP) initiative.

In June 2015, Alex Zimmer, vice-president of new product development for Amazon Marketplace (Marketplace), walked through the Amazon.com Inc. (Amazon) campus in Seattle, Washington, looking over the large open workspaces and admiring some new employee artwork lining the hallway walls outside his office. Every day was exciting and eventful, but Zimmer was especially looking forward to participating in this afternoon's three-year planning (3YP) session (see Exhibit 1 for a glossary of Amazon-specific terminology). At 3:00 p.m., Zimmer would stand before approximately 10 senior executives, with another 50 to 60 operational managers in the audience, to propose an idea that he firmly believed would offer Amazon its next big innovation for Marketplace: the Seller Fulfilled Prime (SFP) initiative. Zimmer knew he would face rigorous questioning at the session about his proposal. He would have to clearly demonstrate the value of SFP for marketplace sellers and end-customers, the potential for SFP to contribute to Amazon's growth, and the link between Amazon's culture and the innovation. He also needed to distinguish between SFP and Amazon's existing Fulfillment by Amazon (FBA) approach and address whether SFP could be seen as competition for FBA. There were practical issues to consider too, such as how to determine whether a seller qualified to participate in the SFP initiative. Another concern for Zimmer was how to demonstrate the value of the new initiative. Decision-making at Amazon was data-driven, but first-time proposals of new innovations could rely more on personal judgment given that solid data often did not exist. Zimmer made a mental note to confirm that the proposal covered the different financial models they had used to evaluate the SFP initiative and its ability to maximize free cash flow (FCF) over time. Zimmer was no stranger to the rigours of the 3YP process, having previously proposed an idea for three consecutive years before finally getting it approved. With his belief in the Amazon motto that Every day is Day 1, Zimmer hoped this new day would lead to the approval of his SFP initiative. AMAZON MARKETPLACE Zimmer had been with the company since the early days of Marketplace, moving between several different leadership roles before landing in his present position. He led a team of approximately 100 Amazonians who were focused on new product and service development for Marketplace. Zimmer's team was responsible for conceptualizing, designing, and implementing new product and service ideas that best realized Amazon's primary vision of being Earth's most customer-centric company. Amazon's leadership principles (see Exhibit 2) created a culture that embodied this vision. Marketplace enabled professional and individual sellers to offer new and used products on Amazon's websites. Sellers had options with respect to how they would fulfill (pick, pack, and ship) customers' product orders. Marketplace sellers could either manage order fulfillment on their own or pay to have Amazon manage that process as FBA. If using FBA, sellers products were stored in Amazon's fulfillment centres, and Amazon picked, packed, and shipped the product and provided related customer service. Regardless of the order fulfillment decision, Marketplace recognized that it had two key customers: Marketplace sellers and the end-customers who bought the sellers' products. Amazon also had Prime customersend-customers who had purchased a Prime membership. The membership offered, among other services, free shipping options on eligible products currently managed through FBA. AMAZON'S 3YP SESSION AND WORKING BACKWARDS The 3YP process began with teams submitting proposals that were evaluated by a board of senior managers. A subset of proposals was selected for presentation at the 3YP event. At the event, proponents discussed their individual proposals, and then a question-and-answer period followed. Zimmer believed his proposal had two attractive features. First, the proposal offered Amazon's customers a greater selection of product choices. Second, the proposal had the ability to scale appropriately within a relatively short period of time. Both of these features were aligned with Amazon's Flywheel concept where growth was fuelled by increasing product selection, which enhanced customer experience, driving more traffic to Amazon. This expanded the seller network, which increased scale, reduced costs, and led to lower prices (see Exhibit 3). In the traditional Amazon approach for developing new initiatives, Zimmer's team worked backwards from the end-customer experience all the way to the seller to develop a sound, innovative proposal following the PR-FAQ process (see Exhibit 4). The working backwards process resulted in the team developing a press release (PR) memo announcing the new initiative. PR memos were written to the new initiative's prospective customers (which could be sellers or buyers or both) and were focused on how the initiative solved customers' problems. PR memos included information about how the current processes (internal or external) were insufficient, and why and how this new initiative would significantly improve upon any existing product or service offerings. Sometimes, solution-based initiatives were developed from specific customer problemsan approach captured by Amazon's "scratched table policy (see Exhibit 1). Other times, teams' initiatives anticipated future problems or opportunities. In addition to the PR memo, the team developed a list of frequently asked questions (FAQ) and associated responses. The FAQ document required the team to anticipate the questions that could be asked within Amazon or by customers regarding the business case, as well as questions about how the initiative would be implemented. Based on his previous 3YP experiences, Zimmer knew that projects could be denied due to insufficient evidence of operational viability. Thus, once they had the concept for SFP, Zimmer and his team of senior managers took a step back by applying the Amazon leadership principle of Insist on the highest standards (see Exhibit 2). They created a list of opportunities and challenges and identified reasonable solutions and suggestions to address questions that might arise in the 3YP. Over the previous few weeks, Zimmer and his team had tightened up their PR memo and expanded their FAQs to address relevant business or function-specific concerns. By working with individual team leads in each of these functional areas, Zimmer was able to identify additional opportunities while considering potential challenges to the SFP initiative. Above all, Zimmer kept reminding himself to be "stubborn on vision, but flexible on details. He would need to remember this concept during the question and answer session later in the day. THE INITIATIVE: DEVELOPING A SELLER-FULFILLED PRIME NETWORK SFP would allow qualified sellers to flag their product listings as eligible for Prime's two-day delivery but send the orders from their own locations and distribution centres instead of from an Amazon facility. At the time, the only seller fulfillment experience with the Prime promise was FBA, which required sellers to ship their merchandise to an Amazon Warehouse and use Amazon Fulfillment to deliver the sellers' products to Amazon customers. Zimmer knew that SFP could unlock the current non-Prime selection of products for customers by leveraging the expertise of trusted sellers. In order to qualify for the SFP program, sellers would have prove their ability to meet the Prime promise for two-day shipping and provide exceptional service to customers. According to Zimmer's proposal, SFP would identify items not already in FBA that could ship directly from sellers to customers. He proposed to initially engage sellers of large items (e.g., appliances and furniture) and sellers of perishable, seasonal, fragile, or customized items. Items that sellers preferred to keep on their premises because of the price point or rarity (e.g., certified and refurbished Rolexes) might also be appropriate for SFP. This expansion in the Prime-eligible selection would delight Amazon customers. Zimmer acknowledged that some sellers might be able to reduce shipping and handling costs by delivering to customers on their own while also having greater inventory control. SFP might also allow Marketplace to attract additional sellers, resulting in a significant increase in the number of items eligible for the two-day shipping service through the Prime program, even beyond the more than 50 million stock-keeping units (SKU) or items now listed, and with even lower prices. Primes free shipping program was an important component to Amazon's sales growth; SFP's potential to increase the number of Prime-eligible items could attract more customers to join the Prime program and increase current members' renewal rates. THE MORNING MEETING To prepare for the afternoon's 3YP session, Zimmer scheduled a review meeting with his advisory team, made up of his top functional managers, to address the anticipated questions regarding the benefits and solutions for the SFP initiative. Since Zimmer had to be prepared to address several functional issues that afternoon, he kicked off the morning team meeting by asking each team leader to reiterate the strengths and potential challenges of SFP from their functional perspectives. Talent Management A project of SFP's magnitude would require a solid team of talented individuals for successful implementation. Given Amazon's leadership principle of Hire and develop the best, Zimmer first asked Jennifer Rodriguez, an experienced human resources manager at Amazon, to review the staffing plan. Zimmer knew that he needed to deploy a critical balance of experienced Amazonians with a few new "blue- badge" external hires (permanent employees), organized into manageable sets of "two-pizza teams (small teams) that were simultaneously interconnected and decentralized (see Exhibit 1). Zimmer and Rodriguez discussed the appealing features of the external hiring process at Marketplace. Typically, the external hiring process involved behaviourally based, one-on-one interviews with several managers and team members, including one bar-raiser"-an employee who had no operational relationship to the hiring team but was a subject matter expert on the Amazon leadership principles (see Exhibits 1 and 2). The role of the bar-raiser was to objectively assess a candidate against the Amazon leadership principles and balance a hiring manager's bias in order to fill positions with the goal of ensuring the long-term hiring needs of Amazon were met. Recruiting internally for the SFP initiative was also a viable option because Amazon encouraged job rotations, an aspect of working for Amazon that was highly valued by employees. Rodriguez reminded Zimmer that there were many potential high-quality internal candidates. Indeed, Amazon leadership strived to make sure all employees progressed in their careers, and the SFP might be the perfect opportunity for some to do just that. Amazon's multi-rater, 360-degree performance evaluation system would be helpful with the internal recruiting process; Amazon looked to hire "all-around athletes who were fungible across roles. Demonstrated successboth functional proficiency and leadership aptitude-on another team had proven to be a strong predictor of likely impact in a new, ambiguous project. The total expected operating expenses for the team, including fully loaded compensation estimates (base salary, restricted stock units, benefit costs), for deploying the SFP project would need to be estimated for budgeting purposes. Zimmer and Rodriguez knew that it would be essential for the SFP implementation team to work efficiently and be flexible in order to meet the projected goals. Finance Amazon traditionally adopted a long-term, patient view regarding financial returns. However, Zimmer realized his SFP initiative still needed to show strong potential to deliver positive returns. For this assessment, he needed financial projections, for which Zimmer next turned to Lynn Buck, his finance expert. Buck had been with Amazon for many years, first as a member of the investor relations team, then as director of finance for Marketplace. Zimmer expected that Buck would provide solid answers to two key questions that would likely be asked during the 3YP session: First, would the SFP initiative maximize FCF over time? Second, would the short-term investment create long-term profit and opportunity for further growth? Buck was fairly confident about their chances at the 3YP session. Twelve years of being in charge of finance and budgeting for Marketplace had taught her to constantly reinvent her approach to financial analysis. The innovation and new product development process certainly required that Marketplace be willing to take significant risk. Over the last few years, feedback from the 3YP sessions indicated that senior executives were increasingly supportive of innovative projects. Double-digit growth during Buck's tenure at Amazon certainly allowed a company of its size to channel internal funds in support of some bold initiatives, which, in Buck's opinion, provided some unique flexibility in supporting innovation with positive investment decisions and actions. Buck ended her brief presentation by stating that, in her opinion, she did not believe that a traditional net present value approach could be used to evaluate the SFP initiative for two main reasons: (1) risk estimates at the cash flow level were impossible to value, and (2) the corresponding cost of capital estimates were equally challenging to define. Since realistic numbers still needed to be provided to appropriately translate SFP's potentials in financial terms, Buck felt that Zimmer needed to be honest by laying out likely outcomes over a three-year investment horizon on FCF and cash management. Three-year projections showing SFP's positive impact on FCF, while supporting the company's current cash management strategy, would certainly play in the project's favour. Estimates needed to be as realistic as possible, but given the uncertainty associated with the proposal's future in- and out-flows, emphasis needed to be placed on scalability and improved customer satisfaction. Do not over-obsess on margins. Focus on profit dollars, Buck advised Zimmer. Buck was accustomed to the company's budgeting approach with respect to innovation: rather than set a precise budget to obsess about, new projects needed to remain financially viable but allow for flexibility within the guardrails or range of acceptable results. Buck prepared a list of selected financials that she believed would provide a strong basis to her financial analysis, which included competitor comparisons (see Exhibit 5). Sales and Marketing Recognizing Amazon's obsession with satisfying customers, Zimmer next turned to Sarah Miller, the team lead in customer service, and Steve Wang, the team lead in marketing. Zimmer looked to Miller and Wang to help address issues related to both sellers and end-customers. Zimmer had two main concerns for Miller to address. First, how would SFP impact the Prime customer experience and the Prime brand? Second, what return policy would be best for SFP? Miller had joined Amazon in 1998, back when all Amazon employees could fit in one building in downtown Seattle. One of the key tools Miller employed when pondering potential innovations was a relentless focus on both sellers and end-customers. She recognized that it was not enough to listen to customers since customers might not always be aware of all possibilities. Her mantra, It's our job to innovate on behalf of customers, was fully exemplified in the SFP initiative. Miller and Zimmer firmly believed that one of the biggest benefits of SFP was its potential to significantly increase the Prime-eligible selection for customers, consistent with Amazon's vision of offering Earth's biggest selection. In spite of this potential, Zimmer expressed some reservations regarding how seller performance could impact the overall customer experience. You have to be able to make the trade-offs, said Miller. Even though it might be challenging in the near term, we know this will help those sellers who can deliver an exceptional customer experience grow their businesses while providing even more great options for end-customers.' Miller recollected the story of Amazon's Jeff Bezos's experience with launching customer reviews as an example. When Bezos first came up with the idea of customer reviews, he received a letter from a publisher saying, You don't understand your job. Your job is to sell books, not to tell the truth about the books. But Amazon's vision was not just about selling books in the near term. Miller concluded, At Marketplace, we want to enable customers to make the best decisions in the hope of building their trust in the long run. In terms of return policy, Miller suggested that a protocol would need to be established as a means of protecting Prime customers from possible seller failures. Miller wondered if Amazon should handle SFP returns and exchanges in a way similar to FBA, including processing returns and determining whether a product was eligible for returns. Zimmer then turned to Wang, who worked closely with the sellers, to address two issues: geographical constraints and seller engagement. Wang had been instrumental in expanding Marketplace's vision and model of creating self-service resources, encouraging sellers to select and adopt features as necessary using Selling Coach (see Exhibit 1). The SFP initiative certainly represented one such product. It offered a self-service resource that would assist sellers with brand promotion and potentially increase sales through access to Prime customers and opportunities such as Prime Day, the annual sales event for Prime customers. While FBA offered a convenient means for companies lacking fulfillment capabilities to get started in Marketplace, SFP would be particularly beneficial for larger sellers that excelled in delivering exceptional fulfillment customer experience and for whom SFP would be an additional, more cost-effective option. For example, highly seasonal products, such as chlorine tablets for swimming pools at 50 pounds per bucket, were difficult and inefficient to handle through a nationwide network that was designed to get product everywhere, rapidly. Sellers could be more efficient fulfilling orders regionally as opposed to using a national FBA service offering. Zimmer and Wang then discussed some potential approaches that could be adopted to increase seller engagement. We usually have a more informal outreach to sellers in the early stages, said Wang. We just check in with them and ask, Does this initiative sound like something that you would be interested in and excited about?' During our three-day Seller Forums, we have our teams interact with sellers and introduce current and proposed initiatives to get their feedback. This is in addition to the annual survey that we send out to sellers to help us understand how well our initiatives are performing and the challenges that sellers face using them." Supply Chain Management Due to the asset and fulfillment implications of the SFP initiative, Zimmer knew that there would be significant supply chain challenges to deploying and successfully implementing SFP. For this reason, Zimmer relied on his supply chain operations team leader, Lisa Ramekin, for recommendations and potential ideas. One immediate concern that Ramekin knew would challenge implementation was figuring out how to qualify sellers to participate the SFP initiative. Evaluating how sellers could be successfully segmented would be necessary to identify a range of acceptable capabilities that could qualify for the SFP initiative. One approach identified by Ramekin's team was to implement a pilot program whereby a few sellers would be hand-selected to participate in an early run of SFP. Once this pilot run was completed, Ramekin planned to survey Prime members to evaluate whether service delight and satisfaction levels had been compromised in any way. Ramekin suggested selecting the pilot sellers from the top Prime performers over the past five years, perhaps identifying those who had shown a consistent shipping record to Amazon FBA facilities according to Prime rules. However, there would be some candidates who were not currently Prime-eligible sellers. Ramekin recalled a recent conversation with one of their potential SFP pilot firms. Castle Clothing was an online apparel retailer that would qualify for SFP based on its current business model. Leo Lyman, Castle's chief executive officer, explained, While we are currently not Prime eligible because we do not participate in FBA, I always ship same-day when it's possible, so customers get the products quickly." Castle operated a 15,000-square-foot warehouse in Ontario, Canada, and relied on Amazon Marketplace for about 70 per cent of Castle's CA$10 million in annual sales. Since Castle would have to absorb the cost of the expedited shipping in Prime, Lyman said he would only list items costing roughly CA$50 or more as Prime-eligible, such as a women's Polo-brand button-down Oxford shirt. Based on Ramekin's conversations at the last three-day Seller Forum, top-ranked sellers were aware that Amazon monitored their shipping performance to Amazon facilities and that it could remove the sellers from the Prime program if they failed to meet specified standards, including on-time shipping rates and order cancellation rates. To ensure that SFP-eligible sellers continued to meet the Prime directive, a qualified seller would have to convey the tracking information to Amazon when an order was fulfilled. As with Prime, sellers could potentially lose their SFP status if they fell outside these metrics. Implementing a new, large-scale fulfillment initiative was not without its challenges. Self-managed fulfillment services were not free for Marketplace sellers, and they also had implications for FBA. We need to educate our sellers on how to identify the best-fit SKU for this kind of initiative, Ramekin explained. The per-item shipping costs for direct-to-consumer sales can be hard to absorb, especially for low-margin items. Even for high volume operations, shipping, order processing, and service costs can be a significant portion of net sales, she concluded. Many sellers could underestimate the cost of last mile delivery. In addition, there were one-time infrastructure, systems, and implementation costs for sellers who had previously only shipped in bulk to retail stores or retail customers. Ramekin's team was also concerned about how sellers could attain a consistent shipping promise given the vast differences in origin- destination transit distances and shipping times. One seller highlighted this exact problem: We can cover the entire East Coast and east of the Mississippi with ground shipping. But once you get further west, the only way to get packages there in two days is to use two-day air. Two-day air costs are back-breaking. "2 Ramekin concluded her session by highlighting an issue of concern from a brand image perspective- sellers' handling of peak demand periods. Periods of significant seasonal demand were of concern in the e- commerce business, and Marketplace was no exception. With the SFP initiative, Marketplace's Prime customers would expect similar levels of service and geographical coverage regardless of whether fulfillment was done by SFP or FBA. If not handled properly, SFP had the potential to jeopardize customer service, especially during peak demand periods. To address this issue, Ramekin reiterated her earlier point regarding how seller education could help identify those best-fit SKUs worthy of the fulfillment investments needed for this kind of initiative. Information Technology Zimmer relied on John Wilson, his technical director, to answer questions about the technical aspects of the SFP proposal. In light of the ongoing technical challenges of scaling up quickly, Zimmer needed Wilson to help address two key issues. First, was it feasible to expect that the technical infrastructure would be able to handle the expected increases in scale that the SFP initiative could produce? Second, how would Amazon Marketplace best use data analytics to benefit its sellers after SFP was implemented? As Zimmer considered SFP's systems challenges, he kept in mind Marketplace's heavy dependence on technical personnel in all its systems efforts. If approved, SFP would immediately need trained technical staff. To encourage shared learning and facilitate implementation speed, technical support personnel always worked side-by-side with project owners. As you know, Zimmer said to Wilson, SFP may require some further technical wizardry as it grows in scale, just to keep things moving smoothly." Yes," replied Wilson. I've been thinking it would be good to do an initial hackathon to brainstorm some possible ways to accommodate SFP's growth system-wise. Zimmer expressed his support for that idea. Wilson continued, But once we generate and vet some starter ideas, we can turn it over to a Scrum project team. With the right mix of technical and business members, we can design the first deliverables and plan each as a four-week sprint to prototype it, run it past sellers, and further refine it." Many technical system challenges existed in SFP, including the scale of operation. So, Zimmer asked Wilson to assure him that the systems requirements could be fully addressed without slowing the main systems down. Wilson replied, I have examined the impact, first, on the main detail product page that customers see. Impact here should be minimal. Implementing SFP will only mean more products will display the Prime option. Zimmer reminded Wilson that impact on back-end systems would also have to be assessed in more detail. Wilson agreed that more detailed assessment would be required. He explained, The back-end systems are complex, with interacting parts, so any change may have an impact. While Agile methodologies speed new application innovation as much as possible, project completion can still take years due to the need for infrastructural changes and the need to coordinate changes that affect other projects or programs. Wilson was more confident on the impact to sellers' experiences. He had completed an early assessment of the impact of SFP on the speed sellers would experience. He assured Zimmer, Here again, the impact should be minimal since similar web screens will be used. Overall, the goal would be to make the entire process easy to use so it can be done in the usual self-service manner. However, Wilson was not certain that his estimates based on this early assessment would pass muster at the 3:00 p.m. session. Alex, we're making decisions based on limited data, so at this stage, I'm not sure we have the data needed to be compelling. With regard to Zimmer's second question about data analytics, Wilson reported, We would build analytic intelligence into the Selling Coach to provide real-time metrics based on SFP standards. The system would inform sellers as they near the standard levels. But that is going to be expensive and time-consuming to build, as we all know from the last project that involved customer metrics. We would need more time to figure out the most cost-effective design. And until SFP gets started, we won't have the full data with which to model what's best. Amazon's data analytics would also help SFP suppliers better understand how much inventory to carry and offer information on potential products that would complement the suppliers' overall product portfolio. Zimmer acknowledged the uncertainties. You are right, said Zimmer. This leaves some key questions. But based on our current experience, let's both present the best plausible answers we can for today's meeting. I'll do my level best, said Wilson. He reiterated his earlier concern regarding the remaining scale-related uncertainties. Due to Marketplace's unparalleled scale, today's solution for a problem may not work as the business scales up. The challenge is often to apply the current change while making it possible to make additional changes at a different scale in one or two years. And I don't have a full answer today for how to do that. It will take further study." THE DECISION Zimmer reflected on the work that had been done over the past few months to put the proposal together and be prepared for that day's meeting. Using the working backwards process, he wondered if the team had captured the value of SFP to marketplace sellers and end-customers in the PR-FAQ and illustrated how SFP could help execute Amazon's Flywheel. Considering Amazon's philosophies and culture, Zimmer hoped he and his team had identified the key questions and the best plausible answers for his presentation, including how SFP related to and competed with other Amazon Marketplace approaches, including FBA. Zimmer felt the team had a good plan for qualifying sellers to participate in SFP and could demonstrate the value of the initiative using different financial models. Now it was time for the 3YP session. Problem Describe the problem your product solves. Solution Describe how your product elegantly solves the problem. In June 2015, Alex Zimmer, vice-president of new product development for Amazon Marketplace (Marketplace), walked through the Amazon.com Inc. (Amazon) campus in Seattle, Washington, looking over the large open workspaces and admiring some new employee artwork lining the hallway walls outside his office. Every day was exciting and eventful, but Zimmer was especially looking forward to participating in this afternoon's three-year planning (3YP) session (see Exhibit 1 for a glossary of Amazon-specific terminology). At 3:00 p.m., Zimmer would stand before approximately 10 senior executives, with another 50 to 60 operational managers in the audience, to propose an idea that he firmly believed would offer Amazon its next big innovation for Marketplace: the Seller Fulfilled Prime (SFP) initiative. Zimmer knew he would face rigorous questioning at the session about his proposal. He would have to clearly demonstrate the value of SFP for marketplace sellers and end-customers, the potential for SFP to contribute to Amazon's growth, and the link between Amazon's culture and the innovation. He also needed to distinguish between SFP and Amazon's existing Fulfillment by Amazon (FBA) approach and address whether SFP could be seen as competition for FBA. There were practical issues to consider too, such as how to determine whether a seller qualified to participate in the SFP initiative. Another concern for Zimmer was how to demonstrate the value of the new initiative. Decision-making at Amazon was data-driven, but first-time proposals of new innovations could rely more on personal judgment given that solid data often did not exist. Zimmer made a mental note to confirm that the proposal covered the different financial models they had used to evaluate the SFP initiative and its ability to maximize free cash flow (FCF) over time. Zimmer was no stranger to the rigours of the 3YP process, having previously proposed an idea for three consecutive years before finally getting it approved. With his belief in the Amazon motto that Every day is Day 1, Zimmer hoped this new day would lead to the approval of his SFP initiative. AMAZON MARKETPLACE Zimmer had been with the company since the early days of Marketplace, moving between several different leadership roles before landing in his present position. He led a team of approximately 100 Amazonians who were focused on new product and service development for Marketplace. Zimmer's team was responsible for conceptualizing, designing, and implementing new product and service ideas that best realized Amazon's primary vision of being Earth's most customer-centric company. Amazon's leadership principles (see Exhibit 2) created a culture that embodied this vision. Marketplace enabled professional and individual sellers to offer new and used products on Amazon's websites. Sellers had options with respect to how they would fulfill (pick, pack, and ship) customers' product orders. Marketplace sellers could either manage order fulfillment on their own or pay to have Amazon manage that process as FBA. If using FBA, sellers products were stored in Amazon's fulfillment centres, and Amazon picked, packed, and shipped the product and provided related customer service. Regardless of the order fulfillment decision, Marketplace recognized that it had two key customers: Marketplace sellers and the end-customers who bought the sellers' products. Amazon also had Prime customersend-customers who had purchased a Prime membership. The membership offered, among other services, free shipping options on eligible products currently managed through FBA. AMAZON'S 3YP SESSION AND WORKING BACKWARDS The 3YP process began with teams submitting proposals that were evaluated by a board of senior managers. A subset of proposals was selected for presentation at the 3YP event. At the event, proponents discussed their individual proposals, and then a question-and-answer period followed. Zimmer believed his proposal had two attractive features. First, the proposal offered Amazon's customers a greater selection of product choices. Second, the proposal had the ability to scale appropriately within a relatively short period of time. Both of these features were aligned with Amazon's Flywheel concept where growth was fuelled by increasing product selection, which enhanced customer experience, driving more traffic to Amazon. This expanded the seller network, which increased scale, reduced costs, and led to lower prices (see Exhibit 3). In the traditional Amazon approach for developing new initiatives, Zimmer's team worked backwards from the end-customer experience all the way to the seller to develop a sound, innovative proposal following the PR-FAQ process (see Exhibit 4). The working backwards process resulted in the team developing a press release (PR) memo announcing the new initiative. PR memos were written to the new initiative's prospective customers (which could be sellers or buyers or both) and were focused on how the initiative solved customers' problems. PR memos included information about how the current processes (internal or external) were insufficient, and why and how this new initiative would significantly improve upon any existing product or service offerings. Sometimes, solution-based initiatives were developed from specific customer problemsan approach captured by Amazon's "scratched table policy (see Exhibit 1). Other times, teams' initiatives anticipated future problems or opportunities. In addition to the PR memo, the team developed a list of frequently asked questions (FAQ) and associated responses. The FAQ document required the team to anticipate the questions that could be asked within Amazon or by customers regarding the business case, as well as questions about how the initiative would be implemented. Based on his previous 3YP experiences, Zimmer knew that projects could be denied due to insufficient evidence of operational viability. Thus, once they had the concept for SFP, Zimmer and his team of senior managers took a step back by applying the Amazon leadership principle of Insist on the highest standards (see Exhibit 2). They created a list of opportunities and challenges and identified reasonable solutions and suggestions to address questions that might arise in the 3YP. Over the previous few weeks, Zimmer and his team had tightened up their PR memo and expanded their FAQs to address relevant business or function-specific concerns. By working with individual team leads in each of these functional areas, Zimmer was able to identify additional opportunities while considering potential challenges to the SFP initiative. Above all, Zimmer kept reminding himself to be "stubborn on vision, but flexible on details. He would need to remember this concept during the question and answer session later in the day. THE INITIATIVE: DEVELOPING A SELLER-FULFILLED PRIME NETWORK SFP would allow qualified sellers to flag their product listings as eligible for Prime's two-day delivery but send the orders from their own locations and distribution centres instead of from an Amazon facility. At the time, the only seller fulfillment experience with the Prime promise was FBA, which required sellers to ship their merchandise to an Amazon Warehouse and use Amazon Fulfillment to deliver the sellers' products to Amazon customers. Zimmer knew that SFP could unlock the current non-Prime selection of products for customers by leveraging the expertise of trusted sellers. In order to qualify for the SFP program, sellers would have prove their ability to meet the Prime promise for two-day shipping and provide exceptional service to customers. According to Zimmer's proposal, SFP would identify items not already in FBA that could ship directly from sellers to customers. He proposed to initially engage sellers of large items (e.g., appliances and furniture) and sellers of perishable, seasonal, fragile, or customized items. Items that sellers preferred to keep on their premises because of the price point or rarity (e.g., certified and refurbished Rolexes) might also be appropriate for SFP. This expansion in the Prime-eligible selection would delight Amazon customers. Zimmer acknowledged that some sellers might be able to reduce shipping and handling costs by delivering to customers on their own while also having greater inventory control. SFP might also allow Marketplace to attract additional sellers, resulting in a significant increase in the number of items eligible for the two-day shipping service through the Prime program, even beyond the more than 50 million stock-keeping units (SKU) or items now listed, and with even lower prices. Primes free shipping program was an important component to Amazon's sales growth; SFP's potential to increase the number of Prime-eligible items could attract more customers to join the Prime program and increase current members' renewal rates. THE MORNING MEETING To prepare for the afternoon's 3YP session, Zimmer scheduled a review meeting with his advisory team, made up of his top functional managers, to address the anticipated questions regarding the benefits and solutions for the SFP initiative. Since Zimmer had to be prepared to address several functional issues that afternoon, he kicked off the morning team meeting by asking each team leader to reiterate the strengths and potential challenges of SFP from their functional perspectives. Talent Management A project of SFP's magnitude would require a solid team of talented individuals for successful implementation. Given Amazon's leadership principle of Hire and develop the best, Zimmer first asked Jennifer Rodriguez, an experienced human resources manager at Amazon, to review the staffing plan. Zimmer knew that he needed to deploy a critical balance of experienced Amazonians with a few new "blue- badge" external hires (permanent employees), organized into manageable sets of "two-pizza teams (small teams) that were simultaneously interconnected and decentralized (see Exhibit 1). Zimmer and Rodriguez discussed the appealing features of the external hiring process at Marketplace. Typically, the external hiring process involved behaviourally based, one-on-one interviews with several managers and team members, including one bar-raiser"-an employee who had no operational relationship to the hiring team but was a subject matter expert on the Amazon leadership principles (see Exhibits 1 and 2). The role of the bar-raiser was to objectively assess a candidate against the Amazon leadership principles and balance a hiring manager's bias in order to fill positions with the goal of ensuring the long-term hiring needs of Amazon were met. Recruiting internally for the SFP initiative was also a viable option because Amazon encouraged job rotations, an aspect of working for Amazon that was highly valued by employees. Rodriguez reminded Zimmer that there were many potential high-quality internal candidates. Indeed, Amazon leadership strived to make sure all employees progressed in their careers, and the SFP might be the perfect opportunity for some to do just that. Amazon's multi-rater, 360-degree performance evaluation system would be helpful with the internal recruiting process; Amazon looked to hire "all-around athletes who were fungible across roles. Demonstrated successboth functional proficiency and leadership aptitude-on another team had proven to be a strong predictor of likely impact in a new, ambiguous project. The total expected operating expenses for the team, including fully loaded compensation estimates (base salary, restricted stock units, benefit costs), for deploying the SFP project would need to be estimated for budgeting purposes. Zimmer and Rodriguez knew that it would be essential for the SFP implementation team to work efficiently and be flexible in order to meet the projected goals. Finance Amazon traditionally adopted a long-term, patient view regarding financial returns. However, Zimmer realized his SFP initiative still needed to show strong potential to deliver positive returns. For this assessment, he needed financial projections, for which Zimmer next turned to Lynn Buck, his finance expert. Buck had been with Amazon for many years, first as a member of the investor relations team, then as director of finance for Marketplace. Zimmer expected that Buck would provide solid answers to two key questions that would likely be asked during the 3YP session: First, would the SFP initiative maximize FCF over time? Second, would the short-term investment create long-term profit and opportunity for further growth? Buck was fairly confident about their chances at the 3YP session. Twelve years of being in charge of finance and budgeting for Marketplace had taught her to constantly reinvent her approach to financial analysis. The innovation and new product development process certainly required that Marketplace be willing to take significant risk. Over the last few years, feedback from the 3YP sessions indicated that senior executives were increasingly supportive of innovative projects. Double-digit growth during Buck's tenure at Amazon certainly allowed a company of its size to channel internal funds in support of some bold initiatives, which, in Buck's opinion, provided some unique flexibility in supporting innovation with positive investment decisions and actions. Buck ended her brief presentation by stating that, in her opinion, she did not believe that a traditional net present value approach could be used to evaluate the SFP initiative for two main reasons: (1) risk estimates at the cash flow level were impossible to value, and (2) the corresponding cost of capital estimates were equally challenging to define. Since realistic numbers still needed to be provided to appropriately translate SFP's potentials in financial terms, Buck felt that Zimmer needed to be honest by laying out likely outcomes over a three-year investment horizon on FCF and cash management. Three-year projections showing SFP's positive impact on FCF, while supporting the company's current cash management strategy, would certainly play in the project's favour. Estimates needed to be as realistic as possible, but given the uncertainty associated with the proposal's future in- and out-flows, emphasis needed to be placed on scalability and improved customer satisfaction. Do not over-obsess on margins. Focus on profit dollars, Buck advised Zimmer. Buck was accustomed to the company's budgeting approach with respect to innovation: rather than set a precise budget to obsess about, new projects needed to remain financially viable but allow for flexibility within the guardrails or range of acceptable results. Buck prepared a list of selected financials that she believed would provide a strong basis to her financial analysis, which included competitor comparisons (see Exhibit 5). Sales and Marketing Recognizing Amazon's obsession with satisfying customers, Zimmer next turned to Sarah Miller, the team lead in customer service, and Steve Wang, the team lead in marketing. Zimmer looked to Miller and Wang to help address issues related to both sellers and end-customers. Zimmer had two main concerns for Miller to address. First, how would SFP impact the Prime customer experience and the Prime brand? Second, what return policy would be best for SFP? Miller had joined Amazon in 1998, back when all Amazon employees could fit in one building in downtown Seattle. One of the key tools Miller employed when pondering potential innovations was a relentless focus on both sellers and end-customers. She recognized that it was not enough to listen to customers since customers might not always be aware of all possibilities. Her mantra, It's our job to innovate on behalf of customers, was fully exemplified in the SFP initiative. Miller and Zimmer firmly believed that one of the biggest benefits of SFP was its potential to significantly increase the Prime-eligible selection for customers, consistent with Amazon's vision of offering Earth's biggest selection. In spite of this potential, Zimmer expressed some reservations regarding how seller performance could impact the overall customer experience. You have to be able to make the trade-offs, said Miller. Even though it might be challenging in the near term, we know this will help those sellers who can deliver an exceptional customer experience grow their businesses while providing even more great options for end-customers.' Miller recollected the story of Amazon's Jeff Bezos's experience with launching customer reviews as an example. When Bezos first came up with the idea of customer reviews, he received a letter from a publisher saying, You don't understand your job. Your job is to sell books, not to tell the truth about the books. But Amazon's vision was not just about selling books in the near term. Miller concluded, At Marketplace, we want to enable customers to make the best decisions in the hope of building their trust in the long run. In terms of return policy, Miller suggested that a protocol would need to be established as a means of protecting Prime customers from possible seller failures. Miller wondered if Amazon should handle SFP returns and exchanges in a way similar to FBA, including processing returns and determining whether a product was eligible for returns. Zimmer then turned to Wang, who worked closely with the sellers, to address two issues: geographical constraints and seller engagement. Wang had been instrumental in expanding Marketplace's vision and model of creating self-service resources, encouraging sellers to select and adopt features as necessary using Selling Coach (see Exhibit 1). The SFP initiative certainly represented one such product. It offered a self-service resource that would assist sellers with brand promotion and potentially increase sales through access to Prime customers and opportunities such as Prime Day, the annual sales event for Prime customers. While FBA offered a convenient means for companies lacking fulfillment capabilities to get started in Marketplace, SFP would be particularly beneficial for larger sellers that excelled in delivering exceptional fulfillment customer experience and for whom SFP would be an additional, more cost-effective option. For example, highly seasonal products, such as chlorine tablets for swimming pools at 50 pounds per bucket, were difficult and inefficient to handle through a nationwide network that was designed to get product everywhere, rapidly. Sellers could be more efficient fulfilling orders regionally as opposed to using a national FBA service offering. Zimmer and Wang then discussed some potential approaches that could be adopted to increase seller engagement. We usually have a more informal outreach to sellers in the early stages, said Wang. We just check in with them and ask, Does this initiative sound like something that you would be interested in and excited about?' During our three-day Seller Forums, we have our teams interact with sellers and introduce current and proposed initiatives to get their feedback. This is in addition to the annual survey that we send out to sellers to help us understand how well our initiatives are performing and the challenges that sellers face using them." Supply Chain Management Due to the asset and fulfillment implications of the SFP initiative, Zimmer knew that there would be significant supply chain challenges to deploying and successfully implementing SFP. For this reason, Zimmer relied on his supply chain operations team leader, Lisa Ramekin, for recommendations and potential ideas. One immediate concern that Ramekin knew would challenge implementation was figuring out how to qualify sellers to participate the SFP initiative. Evaluating how sellers could be successfully segmented would be necessary to identify a range of acceptable capabilities that could qualify for the SFP initiative. One approach identified by Ramekin's team was to implement a pilot program whereby a few sellers would be hand-selected to participate in an early run of SFP. Once this pilot run was completed, Ramekin planned to survey Prime members to evaluate whether service delight and satisfaction levels had been compromised in any way. Ramekin suggested selecting the pilot sellers from the top Prime performers over the past five years, perhaps identifying those who had shown a consistent shipping record to Amazon FBA facilities according to Prime rules. However, there would be some candidates who were not currently Prime-eligible sellers. Ramekin recalled a recent conversation with one of their potential SFP pilot firms. Castle Clothing was an online apparel retailer that would qualify for SFP based on its current business model. Leo Lyman, Castle's chief executive officer, explained, While we are currently not Prime eligible because we do not participate in FBA, I always ship same-day when it's possible, so customers get the products quickly." Castle operated a 15,000-square-foot warehouse in Ontario, Canada, and relied on Amazon Marketplace for about 70 per cent of Castle's CA$10 million in annual sales. Since Castle would have to absorb the cost of the expedited shipping in Prime, Lyman said he would only list items costing roughly CA$50 or more as Prime-eligible, such as a women's Polo-brand button-down Oxford shirt. Based on Ramekin's conversations at the last three-day Seller Forum, top-ranked sellers were aware that Amazon monitored their shipping performance to Amazon facilities and that it could remove the sellers from the Prime program if they failed to meet specified standards, including on-time shipping rates and order cancellation rates. To ensure that SFP-eligible sellers continued to meet the Prime directive, a qualified seller would have to convey the tracking information to Amazon when an order was fulfilled. As with Prime, sellers could potentially lose their SFP status if they fell outside these metrics. Implementing a new, large-scale fulfillment initiative was not without its challenges. Self-managed fulfillment services were not free for Marketplace sellers, and they also had implications for FBA. We need to educate our sellers on how to identify the best-fit SKU for this kind of initiative, Ramekin explained. The per-item shipping costs for direct-to-consumer sales can be hard to absorb, especially for low-margin items. Even for high volume operations, shipping, order processing, and service costs can be a significant portion of net sales, she concluded. Many sellers could underestimate the cost of last mile delivery. In addition, there were one-time infrastructure, systems, and implementation costs for sellers who had previously only shipped in bulk to retail stores or retail customers. Ramekin's team was also concerned about how sellers could attain a consistent shipping promise given the vast differences in origin- destination transit distances and shipping times. One seller highlighted this exact problem: We can cover the entire East Coast and east of the Mississippi with ground shipping. But once you get further west, the only way to get packages there in two days is to use two-day air. Two-day air costs are back-breaking. "2 Ramekin concluded her session by highlighting an issue of concern from a brand image perspective- sellers' handling of peak demand periods. Periods of significant seasonal demand were of concern in the e- commerce business, and Marketplace was no exception. With the SFP initiative, Marketplace's Prime customers would expect similar levels of service and geographical coverage regardless of whether fulfillment was done by SFP or FBA. If not handled properly, SFP had the potential to jeopardize customer service, especially during peak demand periods. To address this issue, Ramekin reiterated her earlier point regarding how seller education could help identify those best-fit SKUs worthy of the fulfillment investments needed for this kind of initiative. Information Technology Zimmer relied on John Wilson, his technical director, to answer questions about the technical aspects of the SFP proposal. In light of the ongoing technical challenges of scaling up quickly, Zimmer needed Wilson to help address two key issues. First, was it feasible to expect that the technical infrastructure would be able to handle the expected increases in scale that the SFP initiative could produce? Second, how would Amazon Marketplace best use data analytics to benefit its sellers after SFP was implemented? As Zimmer considered SFP's systems challenges, he kept in mind Marketplace's heavy dependence on technical personnel in all its systems efforts. If approved, SFP would immediately need trained technical staff. To encourage shared learning and facilitate implementation speed, technical support personnel always worked side-by-side with project owners. As you know, Zimmer said to Wilson, SFP may require some further technical wizardry as it grows in scale, just to keep things moving smoothly." Yes," replied Wilson. I've been thinking it would be good to do an initial hackathon to brainstorm some possible ways to accommodate SFP's growth system-wise. Zimmer expressed his support for that idea. Wilson continued, But once we generate and vet some starter ideas, we can turn it over to a Scrum project team. With the right mix of technical and business members, we can design the first deliverables and plan each as a four-week sprint to prototype it, run it past sellers, and further refine it." Many technical system challenges existed in SFP, including the scale of operation. So, Zimmer asked Wilson to assure him that the systems requirements could be fully addressed without slowing the main systems down. Wilson replied, I have examined the impact, first, on the main detail product page that customers see. Impact here should be minimal. Implementing SFP will only mean more products will display the Prime option. Zimmer reminded Wilson that impact on back-end systems would also have to be assessed in more detail. Wilson agreed that more detailed assessment would be required. He explained, The back-end systems are complex, with interacting parts, so any change may have an impact. While Agile methodologies speed new application innovation as much as possible, project completion can still take years due to the need for infrastructural changes and the need to coordinate changes that affect other projects or programs. Wilson was more confident on the impact to sellers' experiences. He had completed an early assessment of the impact of SFP on the speed sellers would experience. He assured Zimmer, Here again, the impact should be minimal since similar web screens will be used. Overall, the goal would be to make the entire process easy to use so it can be done in the usual self-service manner. However, Wilson was not certain that his estimates based on this early assessment would pass muster at the 3:00 p.m. session. Alex, we're making decisions based on limited data, so at this stage, I'm not sure we have the data needed to be compelling. With regard to Zimmer's second question about data analytics, Wilson reported, We would build analytic intelligence into the Selling Coach to provide real-time metrics based on SFP standards. The system would inform sellers as they near the standard levels. But that is going to be expensive and time-consuming to build, as we all know from the last project that involved customer metrics. We would need more time to figure out the most cost-effective design. And until SFP gets started, we won't have the full data with which to model what's best. Amazon's data analytics would also help SFP suppliers better understand how much inventory to carry and offer information on potential products that would complement the suppliers' overall product portfolio. Zimmer acknowledged the uncertainties. You are right, said Zimmer. This leaves some key questions. But based on our current experience, let's both present the best plausible answers we can for today's meeting. I'll do my level best, said Wilson. He reiterated his earlier concern regarding the remaining scale-related uncertainties. Due to Marketplace's unparalleled scale, today's solution for a problem may not work as the business scales up. The challenge is often to apply the current change while making it possible to make additional changes at a different scale in one or two years. And I don't have a full answer today for how to do that. It will take further study." THE DECISION Zimmer reflected on the work that had been done over the past few months to put the proposal together and be prepared for that day's meeting. Using the working backwards process, he wondered if the team had captured the value of SFP to marketplace sellers and end-customers in the PR-FAQ and illustrated how SFP could help execute Amazon's Flywheel. Considering Amazon's philosophies and culture, Zimmer hoped he and his team had identified the key questions and the best plausible answers for his presentation, including how SFP related to and competed with other Amazon Marketplace approaches, including FBA. Zimmer felt the team had a good plan for qualifying sellers to participate in SFP and could demonstrate the value of the initiative using different financial models. Now it was time for the 3YP session. Problem Describe the problem your product solves. Solution Describe how your product elegantly solves the