Amazon 2027 – The Future of Retail

Amazon only accounts for only about 5 percent of all US retail sales.[1] And online retail as a whole has only 11 percent. So there are plenty of folks who believe that the transition to online retail will happen slowly, and that while Amazon is growing fast it will be long time before it reaches the top of the heap. They see what’s happening as just another twist in the long and winding tale of US retail, the latest installment in a story where department stores rose to replace clusters of small stores, malls and chains exploited the growth of the suburbs, and then Walmart and the big box category killers sliced off large chunks of the US retail market, leaving the big mall anchor stores alive but bleeding. On this view, Amazon is just another shark in a sea of predators, and not a very big or dangerous one at that.

They could not be more wrong. The current “Retail Apocalypse” is just the start. Online retail is entering a period of explosive growth; Amazon is quickly becoming entirely dominant in ecommerce; and  these changes will have a massive impact on retail employment as 3-4 million of retail salespeople lose their jobs over the next ten years.

So this post is about three things:

  • Why Amazon is winning online now and will continue to win right across the retail sector by exploiting the network effects that will make the biggest online player effectively the only online player. Amazon is systematically building strategic competitive advantage to become impregnable. 85 million US households already have Amazon Prime subscriptions.
  • How online shopping is just entering the acceleration phase of technology adoption. Steady growth of about 10 percent annually in recent years will become 20 percent as Amazon’s tools and strategy provides reinforce normal adoption patterns.
  • The impact on work. Workers in retail face a catastrophic future. 2.7-4 million existing jobs at bricks and mortar retailers (B&Ms) will vanish over the next ten years. They will be replaced by far fewer jobs at online retailers, and those new jobs will require different skills and will be located in different places. Amazon also has every intention of automating many of these jobs, as soon as possible, and B&Ms will respond by cutting wages and benefits for those that remain.

Section links
The Rise of Amazon
Amazon and the Double S-Curve: how Amazon is driving us all towards the online future
  Jobs in the age of Amazon
  Conclusions: Amazon and the Double-S Curve – the future of ecommerce
  Appendix A: The three previous shopping revolutions

[1] I use “retail” to describe the entire retail sector excluding gasoline, restaurants, and a few other components that cannot be sold online.

The rise of Amazon

Online shopping is actually the Fourth Shopping Revolution (SR4), as malls and chain stores are replaced for exactly the reasons that allowed them to thrive in the first place: more convenience, better prices.[1]

Ecommerce has well-known advantages, notably lower costs and more choice. But it also faces key challenges: delivery takes time and costs money; buying without touching means more returns; trust is big and local stores have consumer trust and recognition. And it’s still relatively new, so it has the burden of convincing people to change – always a big ask.

Amazon has addressed all these issues by systematically building competitive advantages across the entire supply chain: at the front end (touching the consumer); in its middle layer (touching other sellers); the back end (logistics and delivery); and corporate structure, leadership, and strategy. Each offers formidable competitive advantages. Together, these advantages make Amazon unbeatable, and more unbeatable every day.

Amazon’s Customer-facing advantages.

  • Choice. Amazon offers more than 350 million products in the US – the largest selection of goods available in the US (discounting Alibaba because most US consumers won’t buy direct from China).
  • Extended shelf through Fulfillment By Amazon (FBA). Amazon itself owns only about 12 million of those products – all the rest are third-party products sold through its marketplace. FBA lets Amazon fully exploit the long tail advantages of online retail,[2] offering a cornucopia of choice covering even obscure or technical items. Need a part for a fridge? No problem. Amazon has over just a few years become the go-to source for a hammock, a handsaw, hardwood flooring, or halva (a dessert from Israel – for which an Amazon search generates 264 hits). It’s now a surprise if Amazon doesn’t have something you need.
  • Price. Amazon doesn’t always have the best price, but it does always seem to have a reasonably low price. It discounts all the time (or at least appears to discount), and lower prices from elsewhere are shown on its own web site, so consumers can decide whether to pay Amazon’s sometimes higher prices. UIT seems like bt and large, Amazon prices are low enough.
  • Interface. Amazon has the huge advantage of experience in tuning its site to meet customer needs. Its user interface is intuitive, easy to search and browse, and offers filters that quickly narrow down enormous choice to a reasonable selection. One-click purchases and recurring subscriptions simplify and accelerate buying. A stored address book manages multiple recipients and locations. And Amazon reaps the reward of dominance: even if a better alternative were to be invented, a critical mass of users now expect to see an interface like Amazon’s. If they don’t, they are gone: so there is no opening for competitors to win by using a better interface.
  • User reviews have fundamentally changed shopping, even for obscure and otherwise unreviewable products. Network effects are enormous here, and Amazon’s review engine is the best because it has the largest number of reviews; these attract more buyers who generate even more reviews. In fact, these reviews are so important that companies pay customers to offer unbiased reviews on Amazon. That’s dominance!
  • Customer service. Amazon’s mission is to “be the most customer-centric company on Earth.” Three elements of customer service really matter: first, Amazon takes responsibility for products sold through the site and imposes a no-hassle return policy on third-party vendors who use FBA. Consumers don’t deal with sellers – Amazon does that for you. Second, Amazon reps pick up the phone. Contact is quick and easy, and reps are clearly focused on primarily satisfying the customer. Third, Amazon offers free returns for Prime members. That takes the risk of a mistaken purchase away from the customer, which eliminates a huge barrier to the adoption of online shopping.
  • Amazon Prime. This is the biggest customer-facing advantage of all, partly because free shipping is the single perk most likely to generate additional customers.[3] Prime is key because Prime members are much less likely to go outside Amazon, and Prime offers free 2-day delivery and free returns. Adding in other free Prime features – notably Amazon Music and Amazon Prime Video – creates a set of interlocking incentives to subscribe to Prime. And it’s working: Prime membership is up from 25 million in 2015 to 85 million in 2017 – fully 2/3rds of American households now pay Amazon $99 a year just for membership.[4] Prime membership is up from 10 million in 2012 to 25 million in 2015 to 85 million in 2017 – fully 2/3rds of American households now pay Amazon $99 a year just for membership.
  • This may look like the Costco model, but while at Costco membership is a barrier that has to be overcome before you can to shop there, Prime’s free shipping, Amazon Video and Music, and other benefits “pay” for the membership, which makes is more a tool that locks customers in than a barrier that locks them out. And once you are a member there is little reason to shop elsewhere – for anything.
Where It's Due
Amazon Prime subscribers Sources: Business Intelligence and Consumer Intelligence Research Partners (CIRP). Data are for start of calendar year. * Marks data for June 2017

Amazon has created a virtual mall that solves all the weaknesses of the mall and offers so many additional benefits that it has become what Apple aspires to – a universe from which customers seek no escape.

Seller-facing advantages

Working with Amazon offers overwhelming advantages to other sellers, and they have responded. About 95% of the goods listed, and 40% of the goods (by value) sold through Amazon are owned by other vendors. About 95% of the goods listed, and 40% of the goods (by value) sold through Amazon are owned by other vendors And that share will grow: Nike for example recently removed a long-standing ban to selling direct on Amazon. The advantages of selling through Amazon outweigh even its central strategic interest in managing its own sales channel.[5]

  • Amazon’s market reach. Sellers know that Amazon is uniquely positioned to provide access to US consumers. Beyond the 85 million Prime subscribers, Amazon is by far the most visible and visited ecommerce site, with more than 180 million unique visitors monthly in the US.

  • The full force of FBA. FBA allows vendors to avoid the cost and risk of building their own warehouse, logistics, fulfillment and distribution systems. This is a huge offloading of risk, and at relatively modest cost. Amazon charges far less than the traditional retail markup demands by B&Ms. Amazon gets a cut but more profoundly, Amazon gets to control a marketplace much larger than its own products can reach.
  • Cost of Amazon. Amazon charges a maximum of 18% of the sale price as a “referral fee” (lower for more expensive items). It also charges a low flat rate for FBA services: for example, to a hand pack and ship a book weighing up to one pound costs $4.20.[6] So sellers using Amazon FBA don’t have to invest in warehouse space and services, or pack and ship their own products – a massive offloading of risk, as costs are incurred only when there is a sale.
  • Vendor analytics. Amazon offers analytic tools so vendors can manage both their business and their relationship with Amazon. Third parties have added further capabilities, but Amazon has spent a lot of time and effort optimizing, and its size and dominance mean that it sets the de facto technical standard, so competitors will once again be forced to offer similar tools organized in similar ways.

Basically, Amazon is becoming the Godfather of online retail: using its expanding market reach to make vendors an offer they cannot refuse (and don’t want to refuse).

Logistics and electronic capabilities

Amazon is focused on improving its own logistics services (as Ben Thompson points out, it is its own “first and best customer”[7]). it is also working hard to address last mile and last 50 feet issues: how to deliver faster and more accurately to every customer.

  • Multiple delivery vectors integrated into a single system. Amazon uses its own distribution network (including its own last-mile delivery vehicles). But it also uses UPS, FedEx, and USPS where that makes more sense. UPS currently handles about 30% of Amazon’s 600 million packages annually,[8] USPS a slightly larger share and FedEx a slightly smaller one.[9] That integration minimizes costs and optimizes delivery time, but it’s only worth building such a sophisticated system if you have substantial volume.
  • Amazon uses its own demand for logistics services to build logistics businesses. Amazon Web Services is now the biggest provider of cloud services in the US, but its growth has been driven by Amazon’s role as its own “first and best customer.” Cloud services are almost 10% of Amazon revenues and almost all of Amazon’s overall profits. By owning the infrastructure, it controls strategy and ensures it’s always the top priority customer. And by turning web services into a platform that others can use, Amazon avoids the inevitable complacency of vertically integrated companies.[10] Going forward, Amazon will probably build a fully open freight forwarding competitor to the existing players.[11] According to Bloomberg, internal Amazon documents demonstrate that “Amazon intends to create a revolutionary system that will automate the entire international supply chain.”[12]
  • Amazon has already built a network of warehouses across the US. About 50% of the US population is now within 20 miles of an Amazon warehouse. It directly controls this part of the supply chain linking goods to consumers. And it is squeezing labor fairly hard within this sector.[13]
  • Automation. Amazon is moving fast to automate warehouses. This a major strategic initiative. In 2012, Amazon bought Kiva, a leading warehouse robotics company (now Amazon Robotics), which immediately stopped selling to Amazon’s competitors.[14],[15]
  • Drones. While drone delivery still seems fanciful, Amazon is serious. It recently announced Prime Air, which will use drones to “deliver small packages in 30 minutes or less.”[16] There are plenty of bugs to be worked out, but drones are cheaper and greener if they can be made to work. And they are faster than standard delivery.
  • Automated trucks and the last mile. Similarly, self-driving trucks will drive down costs. Amazon is also experimenting with node delivery – delivery to a single point in the neighborhood rather than the home itself, one part of the new infrastructure needed for automated delivery. Again, Amazon is on the cutting edge here.[17],[18]

Amazon understands that its success rests on the reliability and low cost of its delivery network. So Amazon sees further opportunities to gain competitive ground here, leveraging scale and transforming more cost centers into revenue-generating platforms.

Leadership and strategy

  • The importance of Jeff Bezos. I have a visceral dislike of fawning bios that bolster the singular brilliance of “the leader.” This is usually just lazy nonsense; even huge companies (and governments) work within a tight web of constraining conditions and powerful players. But Bezos is different: his importance at Amazon can’t really be over-estimated. He’s played a key role in developing a culture that gave Amazon the time and space to get it right, to see the long game.[19]
  • The zero-profit advantage. Amazon is famous for not making profits. Revenues are reinvested, and this gives Amazon a strategic advantage over normal companies which have to satisfy their stockholders. Lower profits = lower costs = bigger market share.
  • Independent management and patient shareholders. People who buy Amazon stock are believers in the long game. Bezos says explicitly that Amazon is not interested in short term return to shareholders. Index investors buy and hold because they must, but otherwise if you buy Amazon you accept the Bezos strategy of minimal profits and constant reinvestment. Bezos himself owns around 17% of Amazon, the next largest shareholders have 2.4%, and all the other major shareholders are passive investment funds like Vanguard. Bezos – and hence Amazon’s long-view strategy – has nothing to fear from activist shareholders.
  • Picking off new markets. Amazon has an established track record of successfully invading adjacent markets – remember, it started as an online bookstore. Take groceries. Amazon started selling groceries in 2007. Online grocery delivery requires a different skill set than nonperishable goods like clothes or electronics, notably more detailed quality control, much tighter timelines for delivery, and perishables protection in transit. But if Amazon cracks that nut, the reward is prime access to an almost entirely new market worth close to $1 trillion in annual revenues. As with many Amazon initiatives, the Whole Foods acquisition also works on other levels as well: it extends the reach of the Amazon logistics network into many higher income communities. Amazon is also expanding its efforts to penetrate a wide range of adjacent markets such as tools and home repair, or furniture (see box).

  • Avoiding the complacency of vertical integration. Many large companies become complacent as they integrate vertically and create a soup-to-nuts company where some component divisions become in-house suppliers only. Amazon avoids this by forcing key components – including web services, the customer interface, and the delivery system – to serve other customers. This forces these functions to compete for real customers other than Amazon, and roots out complacency. Amazon is still the top priority customer, but these divisions become better for serving other customers.

Amazon and the double S-curve: how Amazon is driving us all towards the online future

How fast will Amazon grow, and what are the limits of growth for both Amazon and the online shopping more generally?

Patterns of adoption for other technologies and innovations offer a template. The technology adoption S-curve describes how new technologies become widely adopted: the bottom of the curve is a period of relatively slow growth, as new technologies start to enter the market and are picked up by early adopters; the acceleration phase follows as technologies rapidly adopted, and the maturity phase finally reflects market saturation as only the most conservative adopters are left behind(see figure).


The online shopping marketplace will reflect the intersection of two S-curves: the online shopping adoption curve; and the related but independent S-curve that illustrates Amazon’s share of the online market

The online shopping marketplace will reflect the intersection of two S-curves: the online shopping adoption curve; and the related but independent S-curve that illustrates Amazon’s share of the online market. We are now just entering the acceleration phase of the first and are  already well into the acceleration phase on the second.

The first S-curve: ecommerce adoption

The adoption of new technologies is happening faster now (see smartphone, Internet vs telephone, radio). The Internet took about 11 years to go from zero to 70 percent penetration in the US; mobile phones and smartphones have been even faster:[20] landlines are becoming a relic, and cell phones without smartphone capabilities are essentially extinct in the US. Other new technologies are also being adopted rapidly. Look at Uber/Lyft, or Airbnb which is now the largest lodging company in the world by number of rooms available – and was founded in 2008! Compare that to the speed of adoption for telephones, which took 35 years to get to 40% and 65 years to get to 70%.

The first US smartphone was sold in 1994,[21] and the acceleration phase started in 2008, when smartphone penetration hit 11%. It took only 4 years after that for smartphones to reach 50% penetration, and a further 3 years to reach nearly 80%.[22] The story is similar for the Internet: in the US, 9.2% of the population used the internet in 1995. By 2001, that was about 50%, and by 2007 it was 75%. So adoption curves for recent technologies show an acceleration phase lasting 8-10 years, starting when market penetration reaches about 10%, and ending when it reaches 70-80%.

Given the obvious difficulties of predicting detailed outcomes for 2027, it’s useful to begin by defining scenarios that provide boundaries. The lower boundary for online sales can be derived by extending out the pattern of growth from recent years. Under this linear model, online shopping continues to grow at 10% annually, while retail sales overall (excluding autos, gasoline, and restaurants)[23] grow at about 2.3%. Those are the historical averages for the past 6 years. The linear model predicts that by 2027, online shopping will account for about 34% of all retail sales, or around $2.5 trillion.

The upper boundary can be derived from modified S-curve model. Projected growth based on the S-curve model is much faster. While online shopping may not completely match the exponential growth in smartphone penetration, the next 5 years should show year-on-year growth of around 20 percent. After that the curve begins to slow, to perhaps 15 percent for years 6-10. This S-curve model predicts that online shopping will take 55 percent of the retail market by 2027, accounting for $4 trillion in sales. Note that both the Internet and smartphones reached 80 percent penetration, and took 8 years not 10 to do so. So this is a conservative version of the S-curve.

Are these boundaries plausible? It does not on the face of it seem plausible that ecommerce growth will be slower than it has been in recent years. On the other hand, while 55 percent is a big number, other adoption histories end at much higher levels.

Low and high growth scenarios for ecommerce

I use conservative assumptions because retail is a huge industry. It just takes time to turn the Titanic; some people will never shop online; some locations will be very hard to serve; and some kinds of products may be very resistant (e.g. produce?).

So overall, the boundary conditions from these models seem reasonably well grounded.

The second S-curve: Amazon’s dominance of ecommerce

Amazon’s market share of ecommerce is up from 25% in 20912 to 33% in 2015 and 43% in 2016. Some early reports from 2017 have it at 70%. So we are somewhere right in the acceleration phase of Amazon’s own S-curve, and because Amazon is so dominant online, it is affecting the first curve as well: as Amazon grows, it grows the online sector as well.

If we fit Amazon’s market share to the Internet adoption S-curve, we find that Amazon’s market share was in 2015 about half way through the S-curve, and can be expected to reach 70% by about 2021. The extremely rapid recent growth of Amazon Prime membership suggests the acceleration phase may be even shorter, as Prime membership is closely correlated with Amazon-centric online buying patterns, and is a leading indicator of online market share. So using the Internet adoption curve as a model seems conservative.

Once again, conservative and fast-growth scenarios mark the boundaries for Amazon’s market share.  If Amazon just maintains its 2015 share of ecommerce at around 45% of ecommerce revenues – a highly conservative scenario for Amazon – its total revenues in 2027 will be around $1.2 trillion (assuming also that overall online shopping growth is linear, not accelerated).

Amazon’s 2027 revenues will likely be between $1.2 and $2.7 trillion. If, however, Amazon’s share continues to grow through the acceleration phase and then tops out at 70 percent (which is itself conservative given some recent estimates for current market shares), and the online shopping sector also follows S-curve growth, Amazon revenues will be on the order of $2.7 trillion by 2017.  That is 39 percent of the total retail sector.

The double S-curve future of the retail sector

Of course, it may be that unlike previous technologies, ecommerce will somehow not feature an S-curve adoption profile. Maybe it will flatten out at 20 or 25 percent of retail instead of 60-70%.  It may also be that Amazon adoption will follow some other logic – perhaps only 45% of the population is interested in the benefits Amazon offers. But both seem unlikely.

Ecommerce currently grows at about 10 percent annually (according to the Commerce Department). Consumer surveys suggest a faster rate, but using the slower Commerce estimate, compounded growth over ten years means that by 2027, ecommerce will account for 34% of total retail sales.  If, however, we now hit the acceleration phase – which seems highly likely – at least 55% of retail transactions will be made online by 2027.

SR4 is already rolling through the retail sector. Toys “R” Us and Payless are bankrupt. Rue 21 – with it 1,000 stores – may well be next. Sears and Kmart appear to be in a death spiral, with more store closings announced regularly (Fortune recently questioned whether Sears would even make it through 2017).[24] Bloomberg estimates that a record 8,500 retail stores will close this year, far more than even during the previous peak in 2008. This collapse is accelerating: as of April 2017, 2,880 closings had been announced, compared to 1,153 for the same period in 2016. Retailers cut about 30,000 jobs in both March and in April 2017.[25]

As of August 2017

If the second S-curve fully plays out and Amazon’s share of ecommerce eventually rises to 70%, Amazon’s annual revenues would be on the order of $3.4 trillion from the US alone, and from retail sales alone.

Projected Amazon retail revenues 2027

What can the B&Ms do?

The chain stores have only the defense of habit and immediate gratification, and shopping habits can and do change quite quickly. We’ve already seen how this plays out in one leading sector: books. Amazon steamrolled Bor­ders, and it didn’t take very long. Borders operated more than 1,200 stores in 2003, and was still making money in 2006. Between 2006 and 2010, Borders yearly income dropped by more than $1 billion, and the group liquidated entirely before the end of 2011. Amazon ate its lunch.

But eliminating the monster booksellers of the Third Shopping Revolution was just a start.

The retail sector can be divided into predators and dinosaurs. Amazon is the primary predator, with the fastest year-on-year revenue growth, relentless expansion into new retail sectors, and rapidly growing market share.

Dinosaurs are the big retail chains wedded to expensive real estate and focused on instore sales. The Amazon model is already sweeping away the teetering survivors from SR2 and SR3. Sears, Macy’s, Kmart – they are all walking dead, zombie chains that won’t yet admit that their era is over. Walmart is adapting its already somewhat predatory SR3 strategy to SR4, but will its huge real estate footprint be a ball and chain or a competitive advantage? It has a lot of real estate and it’s still almost entirely reliant on in-store sales. But it is trying to meet the threat from Amazon, and has ramped up its own online business and bought other pieces as well (e.g. JET). And obviously, local stores must confer some advantages or Amazon would not itself be considering bricks and mortar.

Why indeed would one go to a store anymore? Only perhaps to window shop before buying on Amazon. So as they thrash around for a strategy, the big chains are in reality creating another free sales channel for Amazon – even while they appear to be Amazon’s fiercest competitors. The local showroom – the greatest apparent strength of Amazon’s competitors – in the end just drives yet more traffic and business directly to Amazon. A recent McKinsey survey of digital shoppers found that almost half the consumers who conduct research on their mobile phones have done so while in stores.[26],[27]

Amazon in 2027

We are a long way from 2027. After all, ten years ago the Great Recession had not even started and Donald Trump wasn’t yet a reality TV star. But Amazon’s emerging system of interlocking advantages seems formidable, and in my view unbeatable.

Nor is Amazon done reinventing itself, its markets, and its ever-deepening ties to consumers. For example, Amazon is introducing radical consumer transparency into its grocery sourcing. It will help customers to track food all the way from the farm. Becoming the “green” grocer has obvious advantages, the has obvious synergies with the Whole Foods acquisition.

More speculatively, why not a Bank of Amazon? Could its huge network of retail consumers soon find some highly attractive offers for financial services? Might Amazon buy PayPal or perhaps something smaller and cheaper? After all, Amazon doesn’t need PayPal’s consumer base. Ten years is a long time in the digital world, and it’s not inconceivable that Amazon could develop its own finance arm – just as Ford and General Motors did. Similarly, Amazon clearly has it’s eye on pharmacies and prescription sales.

Conclusions: Amazon and the Double-S Curve – the future of ecommerce

What will the adoption of online shopping look like in 2027? My money is firmly on the double  S-curve. There are five key reasons why:

  • The power of Prime. Many more people have Prime membership which is a leading indicator of online sales but also a leading facilitator. 85 million Prime members are already paying for access to Amazon’s free shipping and returns. They have already decided that they are going to buy online and from Amazon.
  • Millennials lead the way. Demographics change every day as older consumers are replaced by new digital natives. Just on that basis alone, an additional 2 million or so new adults enter the online retail market every year, and 2 million older adults – who mainly used bricks and mortar – leave.
  • Reimagining online shopping as gratification. We think of online purchases as delayed gratification vis-à-vis store purchases – but that’s wrong: online is instant gratification. Better yet, online purchases offer double gratification – when you buy, and when your package is delivered.
  • Amazon. Amazon is clearly committed to continued improvement – it knows well which challenges and barriers discourage new online shoppers, and it is addressing them. All of them.
  • Evidence from the world’s leading ecommerce China is leading this curve, and its market trajectory looks like an S-curve. Online sales are growing on rocket fuel right now in China.

Amazon is important. It has solved a lot of the barriers and challenges facing ecommerce, and is quite likely to solve others within our timespan. For example, online shopping and delivery will undoubtedly become a much bigger part of the grocery sector. How much bigger? TBD. But Amazon’s presence and drive in that segment can only accelerate adoption.

Amazon’s competitors in contrast have not fully committed to online sales – they have too much bound up in bricks and mortar.[28] Walmart’s online sales are 3 percent of revenues.

We will see if Amazon can really solve some of the barriers to online groceries. We will see whether Walmart can mount an effective challenge to Amazon’s dominance. We will see if Facebook and Google manage to integrate shopping more effectively. But if I was a betting man (and I am), I would bet that within 5 years we will have a new verb: to “Amazon” will mean to buy online, just like Hoover became another word for using a vacuum and Xerox meant to copy something, and Ubering may continue to exist even if Uber does not. Of course, both Hoover and Xerox managed to different degrees to blow their dominant position. But I don’t expect Amazon will follow.

Jobs in the age of Amazon

By 2027 online shopping will have captured 34-55% of the retail sector. How will that affect retail workers?

Previous shopping revolutions shifted retail jobs from one location to another and expanded the number of retail jobs overall.  While new jobs did not directly replace old ones, the number of retail jobs grew steadily during SR2 and SR3 (we don’t have data for SR1) (see Appendix A for a description of these previous shifts).

That won’t be the case for SR4: ecommerce replaces existing bricks and mortar retail with online sales, which fundamentally do not require salespeople, or their supervisors, or cashiers. New jobs will of course emerge – for example in customer service, delivery, and in online development itself, but there will be fewer such jobs, many won’t to be picked up by the workers losing retail jobs, and these new jobs  themselves may be only temporary.

Retail salespeople, supervisors, and cashiers: core working class jobs

The river of retail stuff is huge, and servicing it takes a lot of people. There are about fifteen million employees working in the retail and wholesale sectors, and another six million in transportation and logistics – overall, about 14% of the whole employed workforce supports the flow of goods to consumers.

Three big occupations are in the cross-hairs of SR4: retail salespersons, their supervisors, and cashiers. Together they account for about 10 million jobs[29] – or 38 percent of the jobs in the bottom quintile of the income ladder.

Retail sales offers classic entry-level jobs at low pay. And while these jobs took a hard hit during the Great Recession, they have since recovered and are now close to an all-time high in both employees and share of employment.

Retail salespeople – employment and share of total employment

There are also just under 1.2 million retail supervisors. These jobs are also especially important: they are for many the first step up the employment ladder beyond entry level jobs. But retail supervisors’ share of total employment has fallen about 10 percent since 2011, as retail firms squeeze costs to meet increased competition.

Retail supervisors – employment and share of total employment

Median earnings for both retail salespeople and their supervisors have been falling in real terms since the early 2000’s, and are still down 8-10% from their peak despite a limited recovery in 2014.

Real annual earnings for retail salespeople and their supervisors – BLS

Cashiers too have an archetypical low skill entry-level job, which has been around for more than 100 years. It’s traditionally where high school kids got their first experience at work, women picked up a second income, and older workers returned to the workforce. Most cashiers are women, and many work part-time. Few have any education beyond high school, and many do not have a high school diploma.

There are about 3.5 million cashiers in the US. About 30% work in grocery stores (the largest single industry sector), others in gas stations, other retail stores, and restaurants. The grocery sector divides between a few large employers managing multiple supermarket chains, and thousands of small stores. The four largest companies in the sector account for about one third of total revenues, while the top 20 companies account for 56%. This is not especially concentrated by US standards.

Cahiers – employment and share of total employment

In the early 1990’s, 78% of cashiers were women. But as men with limited skills have been forced out of higher paying jobs some are working as cashiers, and by 2015 women’s share was only 72%. Hispanics account for more than 20%, almost doubling their share over twenty years.

Part-time employment is the norm in retail for younger workers. Almost 80% of teenage workers in retail are part-time, as are half the workers age 20-24, and a third of those 55 or older.

Retail jobs in ten years time[30]

The baseline and fast track scenarios allow us to model the impact on retail workers ten years out.  Critically, online venodrs are much more efficient: as of 2015, Amazon generated about three times as much revenue per worker as Penney’s (which we can use as a proxy for B&MK’s generally).  This is no surprise: Amazon has no huge point of sale infrastructure or labor force, pays for no retail locations, and can scale cheaply and quickly. Penney’s has none of those characteristics.  So as B&M sales shrink, demand for labor in the B&M world will shrink as well.

To develop a simple illustrative model, I made the following assumptionsL

  • Online stores require 1/3rd of the labor required by B&Ms.
  • This shrinkage is pro rata: a 5 percent reduction in B&M sales will lead to a 5 percent reduction in B&M sales floor employment.[31]
  • Similarly, as online sales grow, online vendors will also hire more labor pro rate: a 5 percent increase in sales requires a 5 percent increase in their labor force.
  • Productivity at both online stores and B&Ms will grow but at different rates: online vendors will improve productivity 5 percent annually, while B&Ms do so at 1 percent annually (higher productivity growth for either would lead to higher net losses).
  • Retail sales will continue to grow at their historical pace – about 2.3 percent annually

To determine approximate size of job losses at B&Ms and of gains at online retailers, we have to calculate the impact of the shift in sales from B&Ms to online, adjusted both for changes in productivity and the slow but steady overall growth of retail sales across the period.

Based on these assumptions, bricks and mortar retail will lose between 2.7 and 4 million of its 7 million current jobs for salespeople and first line supervisors (applied to the baseline and accelerated growth scenarios described earlier). The future of online grocery is too difficult to predict at this point, but it seems likely that a significant number of addition cashier jobs will also vanish.

High/low job loss scenarios

So between 30 and 60 percent of the existing retail floor labor force will be eliminated.


Nothing is inevitable. Some new technology might destroy Amazon just as Amazon is destroying the dinosaurs of SR2 and SR3. Maybe Google will come up with a way for consumers to order just by thinking, and delivery will arrive by teleportation. But I’m not betting on it.

New retail modalities sequentially replace old ones. Ecommerce growth is undeniably coming, and so is the accelerating expansion of Amazon. It clearly has a long-term strategy to expand into retail sectors adjacent to core markets, including groceries. It may well strengthen its position further by turning its world-beating logistics chain into profit centers as it already has with AWS, thus producing profits AND ensuring that these services retain their competitive edge.

Amazon has already swept more than 85 million households – about 2/3rds of all US households – into Amazon Prime, which is the bolt that locks the entire edifice firmly into place. But Amazon is not resting on its laurels. For example, it sees home automation as a strategic imperative – unlike competitors for whom it is an interesting playground. It is devoting major resources to its Alexa home automation project – 5,000 people – while Google Home appears to have at best a couple hundred. Why? Because it’s strategically important for Amazon to deny Google, Apple, and others any daylight to insert themselves between Amazon and its customers.

The scenarios described above have profound implications for the retail labor market. Based on our assumptions about the rate at which online shopping will gain market share and the online giants will improve productivity, and standard estimates for the growth of the retail sector overall, we can conclude the following:

  • Substantial job loss among B&Ms. Overall, across ten years, B&Ms will lose between 2.7 and 4 million existing jobs. Those jobs losses will accelerate over time.
  • There will be job growth at online retailers, but not enough to compensate. Based on current staffing trends, and assuming modest improvements in productivity, online retailers will hire between 1.2 and 2 million new jobs by 2027.
  • There is substantial net job loss, which will range between 1.7 and 2 million jobs.
  • These new jobs are not a like-for-like replacement for lost jobs in bricks and mortar chains. They will demand different skills, in different geographic locations.
  • Automation is coming for online jobs too. Many new jobs will be temporary, as online retailers – particularly Amazon – Warehousing and delivery will both be in the front lines of automation. The model does not explicitly address this possibility, which would result in further net job losses.
  • Amazon will continue to improve its productivity, as it commits to an automated future. Automated warehouses are coming very soon. Drone delivery may seem like a fad, but Amazon is serious about it. And Amazon will find other delivery solution for the last mile, including driverless vans. Overall, Amazon will surely grow its current 3:1 productivity advantage over traditional retailers. That means fewer workers will be needed.
  • Online retail is a game of scale that Amazon may already have won. Amazon’s physical and technical infrastructure is enormous and growing very fast. Matching it will soon be impossible. Only Walmart might compete among the bricks and mortar retailers
  • Amazon’s competitors will react by cutting jobs, wages and benefits. This is already happening. Higher wage staff are being let go, and wage differentials are being compressed downward. Floor staffing at B&Ms is already down sharply, and will continue to be cut until the dinosaur dies. Job impacts in retail will also spill over into parallel markets, as relatively unskilled workers let go from retail try to find new jobs.

These effects are already visible in the data. For example, the rate at which Amazon is taking market share from B&Ms is starting to accelerate:[32]

A final key point. The disruption of retail is important partly because of scale – those lost jobs are around 3 percent of the workforce and 10-15 percent of jobs for the lowest income quintile of workers – but also because retail jobs are key for so many low-income families. These jobs can be picked up easily, they often offer flexible hours, they require almost no initial skills, and they can provide a stepping stone to a more responsible supervisory job – an initial step up toward the middle class. Eliminating these jobs means eliminating really important opportunities for low income and unskilled workers. Their decline will leave a huge hole in the US jobs ecology, and it’s not clear at all what could – or should – replace them.





Appendix A. Historical context – the 3 previous shopping revolutions

The first Shopping revolution (SR1): the rise of the department store

The history of retail is the story of new distribution systems replacing old. SR1 began in the first decades of the 19th century, as disposable income for the first time grew to include a new growing mercantile class. While there is no one-to-one correlation between the rise of department stores and the accelerating pace and power of the Industrial Revolution, they were clearly closely aligned. The industrial revolution generated surplus income and a prosperous consuming middle class, and department stores emerged to primarily target the women who had now escaped from the drudgery of domestic service through the growing wealth of their husbands. Harrods for example was founded in 1834, and department stores quickly spread to the US; by 1862 Stewart’s covered a full city block, with 19 departments in 8 floors around a glass-covered atrium.

Figure 11 Four shopping revolutions

Source: McKinsey and Co.

Selfridges (1909) formalized the idea of shopping as a pleasurable activity, and advertised widely. It included restaurants, reading and writing rooms, a Silent Room, and a series of scientific and educational exhibits. William Young stresses the impressive architecture and ornate decoration of department stores at their height.

“Architecturally, these multifloored “palaces of consumption” often featured ornate cast-iron facades with vast, open interiors. At times, they boasted fanciful domes and skylights that flooded the interiors with natural light in the days before electrification. Plate-glass windows on the street level allowed elaborate displays of the treasures within, thus making “window shopping” a new urban leisure activity. Since the store itself was palatial, this focus on display created an atmosphere. It produced the proper environment for purveying goods that were seen as marks of achievement instead of necessities. Everything was ready-made; rather than bolts of cloth, here were racks of dresses. These were items for instant use, for immediate gratification.”[33]

Shopping as an activity and as a means to personal gratification and social status became deeply embedded. The consumer society is an old trope but an accurate one, especially but not only in America. Status symbols continue into the digital age (Apple? Fitbit? Tesla?). but they were most physically tied to the temple of American commerce and luxury, the department store.

SR1 can be summarized thus:

  • The aggregation of many stores into one
  • The invention of shopping as an activity and the tore as destination
  • Technical advances based on refrigeration and more formalized supply chains
  • Validation of the middle-class consumer as a new arbiter of value

But where are they now, those shining beacons of American life for more than 100 years? Sleeping with the fishes, that’s where. And it took less than 50 years for them to sink to the bottom.

The second shopping revolution (SR2): malls and chains

in the 1920’s, chain stores like Sears and JC Penney proliferated across rapidly growing suburbs. By 1930, JC Penny had almost 1,500 stores, while Sears grew from 8 in 1925 to 338 by 1930(although it also had a dominant mail order business). Just as SR1 was rightly linked to the Industrial Revolution, the rise of the chain stores was deeply intertwined with the Automobile Revolution, so it’s not surprising that SR2 started in America. As cars began to dominate the American landscape, they drove the extraordinary rise of the suburbs where working people could acquire all the elements of the American Dream – a house, a car, a family, maybe a boat, even some dogs – while remaining close enough to benefit from the economic drive of the industrial cities. [cite needed on rise of the suburbs].

Suburban families had no interest in going into the city to shop. Too far, too difficult to park, and increasingly seen as too dangerous.  The chain stores saw an opportunity, and these big new stores quickly came to anchor the next iteration of shopping – the mall.

These stores offered better value, and targeted men as well as women. Catering to drivers not pedestrians, the chains had developed the entirely familiar window-less box store by the mid-1930’s, and they dominated retail from the 1930’s until the late 1970’s.

The new chains also led in segmenting the retail market. Upscale chains like Neiman Marcus and Nordstrom differentiated themselves from mid-market stores like Macy’s and Lord and Taylor, which in turn were a step up from Kohl’s, Penney’s, and Sears. In the 1980’s, discounters emerged to focus on price – Target, Walmart, and Kmart. These stores were complemented by off-price retailers like TJ Maxx which carried higher quality goods at discounted prices, often focused on a particular niche (e.g. clothing). Finally, discount clubs and warehouse stores emerged in the 1990’s to challenge the low-end retailers on price, using better logistics and higher volume to create advantage. All these stores created a new shopping ecology around the mall, where a few of the bigger stores drove traffic that could then be exploited by a shifting constellation of specialty stores (e.g. footwear, or accessories) as well as mall services like food courts.

Some of the oldest and most established department stores simply failed to navigate SR2 as it shifted traffic from retail in cities to retail in the suburbs. Icons like Woodward and Lothrop and Garfinkel’s in Washington DC, and Hecht’s and Hutzler’s in Baltimore did not have the scale to make the jump to the suburbs and the malls springing up across America. All went out of business in the 1990’s.

SR3: Walmart, consumer clubs, and category killers

SR2 ended in the 2000’s with many of the lower-cost chains under pressure from the cancerous growth of Walmart with its relentless focus on driving down costs (and prices), its huge variety of goods under one roof, and its ubiquitous presence: it’s said that 90% of Americans live within ten miles of a Walmart. While higher end stores could differentiate themselves, the Sears and Penney’s who had dominated the early part of SR2 now found themselves essentially defenseless.

More widely, small retailers and independent department stores on Main Street were eviscerated by Walmart, and local bookstores got destroyed by Borders and Barnes & Noble. Big box consumer goods stores sprang up, and Home Depot and Lowe’s quickly snatched up large shares of both DIY and construction materials markets. In each case, new logistics became available as computer capabilities advanced, and the bigger chains drove prices down by using their monopsony pricing powers against manufacturers and suppliers. In short, more efficient and cheaper ways to deliver goods quickly replaced local stores with roots a hundred years old.

The relentless pressure of low-cost stores like Walmart with their vast array of items hollowed out existing retail centers, especially in downtown areas. The result has been a Darwinian struggle for survival. Of the 10 largest retailers in 1980, only 4 are still on the list as of 2012. Household names like Kmart and Sears are in sharp decline. And the success of Walmart’s strategy is obvious: in 1980, it was barely larger than K-Mart. Today, it is more than three times the size of its closest competitor.

One can also see SR2 as the industrialization of retail – the application of lessons learned in manufacturing to a different sector. Economies of scale became increasingly important, and the winning chains found many ways to squeeze extra productivity (and costs such as benefits) out of their workforce, indeed and used that to drive prices down and competitors out of business. They found that they could make up in volume what they did not make on each unit, and in the process created a world where only giant competitors could thrive.

So by the early 1990’s, the retail world has settled into the pattern dictated by SR2. Most shopping occurred in the suburbs, either at malls or at stand-alone big box stores (e.g. Best Buy, Frye’s). Different stores catered to different clienteles, and for those chains that had survived first the transition from the city center to the suburbs and then the relentless price pressure imposed by Walmart and the warehouse stores (notably Costco), life if not good was at least stable and tolerable.

But not for long.





“59 Percent of Consumers Say Online Reviews as Trustworthy as a Friend.” Accessed July 11, 2017.

“Amazon Prime Reaches 85 Million US Members.” Consumer Intelligence Research Partners, July 6, 2017.

“Amazon Prime Subscribers Hit 80 Million.” Long Room. Accessed July 12, 2017.

“Amazon Robots Poised to Revamp How Whole Foods Runs Warehouses.”, June 26, 2017.

Anderson, Chris. “The Long Tail.” WIRED, October 1, 2004.

Borthwick, Graham. “Amazon’s Real Threat to Freight Forwarders – the Pieces Are Falling into Place.” LinkedIn Pulse, May 18, 2017.

Bose, Shubhomita. “Amazon, Shopify Q2 Results Speak Volumes About eCommerce Future.” Small Business Trends, August 16, 2016.

Bowman, Jeremy. “Is This the Real Reason Why Wants Its Own Delivery Service?” The Motley Fool, February 15, 2016.

Casselman, Ben. “A Start-Up Slump Is a Drag on the Economy. Big Business May Be to Blame.” The New York Times, September 20, 2017, sec. Economy.

Del Rey, Jason. “Millennials Buy More Clothes on Amazon than Any Other Website.” Recode, March 9, 2017.

Desai, Parag, Ali Potia, and Brian Salsberg. “Retail 4.0: The Future of Retail Grocery in a Digital World.” PWC, n.d.

“How Many Products Does Amazon Really Sell?” Accessed July 16, 2017.

“Infographic: Where Amazon Impacts Most on People Shopping at Stores.” Statista Infographics. Accessed July 18, 2017.

Intelligence, B. I. “Amazon Accounts for 43% of US Online Retail Sales.” Business Insider. Accessed July 8, 2017.

“Iqaluit Residents Live in Fear of Losing Amazon Prime’s Free Shipping.” CBC News. Accessed July 14, 2017.

Kang, Cecilia. “Internet Giants Face New Political Resistance in Washington – The New York Times.” New York Times, September 20, 2017.

Kanter, Zack. “Why Amazon Is Eating the World.” TechCrunch. Accessed July 3, 2017.

Kaufman, Steve. “Will Amazon Logistics Become a Competitor of UPS? And Does It Matter?” Insider Louisville, March 10, 2016.

Kessler, Sarah, and Sarah Kessler. “As Amazon Purchases Whole Foods, Don’t Fear for the Cashier. Fear for the Supermarket.” Quartz, June 18, 2017.

LaVecchia, Olivia, and Stacy Mitchell. “Amazon’s Stranglehold: How the Company’s Tightening Grip Is Stifling Competition, Eroding Jobs, and Threatening Communities.” Institute for Local Self-Reliance, November 2016.

Miller, Claire Cain. “Amazon’s Move Signals End of Line for Many Cashiers.” The New York Times, June 17, 2017, sec. The Upshot.

Pandolph, Stephanie. “THE FREE SHIPPING REPORT: How This Pervasive Perk Is Eating Away at Margins, and the Strategies Retailers Can Adopt to Compete with Amazon.” Business Insider, May 31, 2017.

Pisani, Bob. “Retail Sales Slump Could Be Here to Stay.” CNBC, May 12, 2017.

Robischon, Noah. “Amazon’s Grocery Ambitions Are Far Bigger Than Whole Foods.” Fast Company, June 16, 2017.

Scoper, Sencer, and Alex Sherman. “Amazon Robots Seen as Lined up for Some Whole Foods Tasks.” Arkansas Online, July 3, 2017.

Sen, Conor. “Amazon’s Real Target Isn’t Whole Foods. It’s Everything You Buy.”, June 20, 2017.

Singleton, Micah. “Amazon Prime Wardrobe Will Let You Try on Clothes before You Buy Them.” The Verge, June 20, 2017.

Soper, Spencer. “Amazon Will Sell Nike Shoes Directly Through Brand Registry.”, June 21, 2017.

Staley, Oliver. “Amazon’s Acquisition of Whole Foods Is the Game-Changer.” Quartz. Accessed July 6, 2017.

Stanton, Taylor. “Holiday Update: Amazon’s Market share Is up and so Is Shipping Speed.” Slice Intelligence, December 22, 2016.

Stone, Brad. The Everything Store: Jeff Bezos and the Age of Amazon. Transworld, 2013.

tax, * All products require an annual contract Prices do not include sales. “U.S. Retail Websites by Visitors 2017 | Statistic.” Statista. Accessed July 16, 2017.

“The Future of Grocery E-Commerce, Digital Technology and Changing Shopping Preferences                       around the World.” Nielsen, April 2015.

Thomas, Lauren. “This Chart Shows How Quickly Amazon Is ‘Eating the Retail World.’” CNBC, July 7, 2017.

Thompson, Ben. “Amazon’s Dominant Strategy.” Stratechery by Ben Thompson, August 8, 2013.

———. “Amazon’s New Customer.” Stratechery by Ben Thompson, June 19, 2017.

———. “The Amazon Tax.” Stratechery by Ben Thompson, March 15, 2016.

“US Ecommerce Sales Growth Figures 2016.” Smart Insights. Accessed July 6, 2017.

“U.S. Online Retail Sales to Approach $440 Billion in 2017.” Accessed July 6, 2017.

Wahba, Phil. “In Two Charts, How Amazon Is Killing Its Traditional Competitors.” Fortune, May 11, 2016.

Yglesias, Matthew. “The Real Reason Amazon Buying Whole Foods Terrifies the Competition.” Vox, June 20, 2017.

[1] Previous shopping revolutions are described in Appendix A. Briefly, SR1 created the downtown shopping district, SR2 saw the eclipse of department stores and the rise of malls and suburban auto-driven shopping. Walmart, shopping clubs, and category killers like Best Buy and Borders emerged during SR3.

[2] Chris Anderson, “The Long Tail,” WIRED, October 1, 2004,

[3] Stephanie Pandolph, “THE FREE SHIPPING REPORT: How This Pervasive Perk Is Eating Away at Margins, and the Strategies Retailers Can Adopt to Compete with Amazon,” Business Insider, May 31, 2017,

[4] “Amazon Prime Reaches 85 Million US Members” (Consumer Intelligence Research Partners, July 6, 2017),

[5] Spencer Soper, “Amazon Will Sell Nike Shoes Directly Through Brand Registry,”, June 21, 2017,

[6] Amazon web site. Accessed Aug 24 2017

[7] Ben Thompson, “Amazon’s New Customer,” Stratechery by Ben Thompson, June 19, 2017,

[8] Steve Kaufman, “Will Amazon Logistics Become a Competitor of UPS? And Does It Matter?,” Insider Louisville, March 10, 2016,

[9] Jeremy Bowman, “Is This the Real Reason Why Wants Its Own Delivery Service?,” The Motley Fool, February 15, 2016,

[10] Zack Kantor, “Why Amazon Is Eating the World,” TechCrunch, accessed July 3, 2017,

[11] Graham Borthwick, “Amazon’s Real Threat to Freight Forwarders – the Pieces Are Falling into Place,” LinkedIn Pulse, May 18, 2017,

[12] Kaufman, “Will Amazon Logistics Become a Competitor of UPS?”

[13] Jodi Kantor and David Streitfeld, “Inside Amazon: Wrestling Big Ideas in a Bruising Workplace,” The New York Times, August 15, 2015,

[14] James Vincent, “Amazon’s Latest Robot Champion Uses Deep Learning to Stock Shelves,” The Verge, July 5, 2016,

[15] Louise Masks, “The Future of Robot Labor Is Unfolding in Shipping Warehouses,” Motherboard, September 16, 2016,

[16] Amazon announces Prime Air,

[17] Stephanie Hernandez McBain, “Amazon Builds Team for Autonomous Vehicle Technology,” Automotive News, April 24, 2017,

[18] Borthwick, “Amazon’s Real Threat to Freight Forwarders – the Pieces Are Falling into Place.”

[19] Brad Stone, The Everything Store: Jeff Bezos and the Age of Amazon (Transworld, 2013).

[20] Michael DeGusta, “Are Smart Phones Spreading Faster than Any Technology in Human History?,” MIT Technology Review, accessed July 12, 2017,

[21] Although the BellSouth Simon Personal Communicator was no called a smartphone at the time, it had key smartphone features.

[22] “U.S. Smartphone Penetration Surpassed 80 Percent in 2016” (ComScore), accessed July 12, 2017,

[23] These sectors are usually excluded from ecommerce analysis as they are inherently bricks-and-mortar businesses

[24] Phil Wahba, “Why 2017 Will See Desperate Department Stores and Anxious Apparel Chains,” Fortune, January 3, 2017,

[25] Lindsey Rupp, Lauren Coleman-Lochner, and Nick Turner, “America’s Retailers Are Closing Stores Faster Than Ever,”, April 7, 2017,

[26] Ian MacKenzie, Chris Meyer, and Steve Noble, How retailers can keep up with consumers, McKinsey October 2013

[27] The retail Armageddon may also be bigger and faster in the US, because the US has dramatically more retail store space than comparable countries – about 25 square feet of retail space per capita, compared with about 7.5 square feet in Europe. Space costs money of course.

[28] Phil Wahba, “In Two Charts, How Amazon Is Killing Its Traditional Competitors,” Fortune, May 11, 2016,

[29] Bureau of Labor Statistics. Accessed July 2, 2017

[30] A spreadsheet showing all calculations and assumptions for all models is available at the Great Disruption blog

[31] It is likely that B&M chains will try to cut labor faster to become more competitive, and also that wages and earnings are to some degree fungible – it’s possible that employment losses will be slowed by accelerated earnings losses.

[32] Lauren Thomas, “This Chart Shows How Quickly Amazon Is ‘Eating the Retail World,’” CNBC, July 7, 2017,

[33] William H. Young, “Department Store,” in Encyclopedia of American Studies (Baltimore: Johns Hopkins University Press), accessed June 20, 2017,