eBay vs Amazon: decentralized vs centralized e-commerce

Note: The company I cofounded, Hunch, was acquired by eBay in November 2011. I am now an eBay employee. But all the opinions expressed below are my own, and were developed prior to the Hunch acquisition, through my own research on e-commerce.

Amazon and eBay are the two largest e-commerce companies. As of this writing, Amazon has a market cap of about $87B, trading at a trailing twelve-month P/E of about 139. eBay has a market cap of about $42B, trading at a trailing P/E of about 13. Each company competes with many other companies in many different areas. For example, Amazon competes with Apple on tablets (Kindle vs iPad) and digital media (Amazon’s media store vs iTunes). Ebay’s Paypal unit competes with multiple payment companies, and its marketplaces division competes with other “peer-to-peer” e-commerce sites like Craigslist. But given the potential size of the e-commerce market (not to mention the online-to-offline commerce market), Amazon and eBay’s main competitors are each other. And to understand their large strategic moves (e.g. large acquisitions like GSI and Zappos), it is important to understand their fundamentally opposing strategic outlooks: eBay wants commerce to be more decentralized (around its GSI/Magento partners and eBay marketplaces sellers) and Amazon wants it to be more centralized (around itself).

First, some background. During the dot-com boom, many largest offline brands debated how to best move their businesses online. Some tried to build their own websites from scratch. Others partnered with commerce technology providers. Toys ‘R’ Us took a novel approach and signed a “strategic alliance” to outsource all of their e-commerce operations to Amazon. Over the next few years this relationship soured – apparently Toys ‘R’ Us felt Amazon was competing too directly with them and successfully sued to end the relationship.

The end of the Toys ‘R’ Us – Amazon relationship marked a turning point for a company called GSI Commerce. GSI took an aggressively neutral approach to providing technology and marketing solutions to retailers. Their main appeal over Amazon is that they didn’t compete with their partners (but of course their partners competed with each other). This approach paid off: GSI now powers over 500 large commerce sites, including Toys ‘R’ Us, Adidas, Ralph Lauren, and the commerce sites of all the large sports leagues like the NFL, MLB and NBA.

Last year, eBay paid $2.4B to acquire GSI Commerce. They also acquired a smaller company called Magento that provides e-commerce technologies to smaller retailers. You can think of GSI as the leading commerce platform for the “fat head” of retailers, and Magento as the leader for the long tail.

The key difference between eBay and Amazon isn’t auctions vs. fixed price sales (the majority of eBay sales aren’t auctions anymore). It is that eBay doesn’t take inventory, and prefers to be an intermediary that facilitates peer-to-peer commerce. This strategy wins if e-commerce becomes more decentralized, with the majority of commerce continuing flow through small to medium retailers. In this world, eBay makes money by sending traffic from eBay.com, from fees collected by GSI and Magento, and Paypal transaction fees. In a centralized world, Amazon grows its current 9% e-commerce market share to a much larger percentage, taking advantage of its scale, efficiency, advanced technology, and the convenience of shopping in one place.

One way to view this battle is to think of eBay as a platform a la Windows or Android and Amazon as an end-to-end solution a la Apple computers in the 90s or iOS devices today. Platforms tend to provide greater diversity. In the case of e-commerce, the platform approach could also have a price advantage. As the CEO of TrialPay, Alex Rampell, argues: “Who can beat Amazon on price? The companies whose products are sold on Amazon”. End-to-end solutions like Amazon’s tend to provide greater convenience and a better user experience.

I’m not arguing that one approach is superior to the other. My point is simply that when you understand that the battle is between centralized and decentralized commerce, the strategic moves of the two companies make a lot more sense.


36 thoughts on “eBay vs Amazon: decentralized vs centralized e-commerce

  1. You’re correct, but amazon opened up their marketplace to regular retailers and joe-schmo’s a few years back and now 30% of items sold on amazon are from these 3rd party merchants. Amazon is winning because they have both centralized and decentralized and they use the massive amount of data and content from the open-marketplace to their advantage.

    I know a bit about this industry, for 15 years my company has represented factories that sell to distributors, and 5 years ago, amazon became a new customer and even the amazon buyers are competing with 3rd party merchants on their own site. 

    I was hoping that the GSI acquisition would put ebay in a new market of selling their own goods.

  2. Great additional information.  One item not mentioned here is Amazon Web Services (AWS) and the huge number of e-commerce players using EC2 and S3.  So you have 3rd party merchants within Amazon’s walls and 3rd party merchants on Amazon’s cloud platform (effectively subsidizing Amazon.com’s global distribution costs).

  3. erik112233 says:

    very interesting post! what do you think about startups that try to compete with these big companies? (on my experience it’s more common for ebay alternatives than amazon ones)

  4. Thanks, I actually never thought of it that way, and I have to say that I’m conflicted. It makes me want to route for eBay even though my experiences as a customer of eBay have been negative, while my experiences with Amazon as a customer have been very positive.

  5. rohit_mod says:

    insightful. magento, gumroad, kout are all moving towards smarter distributed commerce. 
    amazon seems headed to being an online mall of malls taking care of the digital equivalents of exits-from-freeways, parking, food courts, and of course all the inventory + fulfilment. 
    distributed commerce seems to be moving towards wherever the users/consumers may be – fertile greenfields.

  6. everyone is trying to create a platform — ebay, amzn, apple, google, all of them. that is the name of the game and everyone knows it. 

    the best platform will create the most opportunities for personalization. some degree of end to end integration is needed as personalization is built atop data profiles, which can best be achieved through tracking that only an end to end solution can deliver. with amazon making computers and delivering web pages via AWS, they are on the verge of becoming in a league of their own on the data collection game, and thus, the personalization game, and thus, the platform game. 

    i do believe niche solutions can beat amazon, but tiny niche platforms is a concept that is currently underexplored, in my opinion. 

  7. My bet is on the Amazon / centralized e-commerce model.

    Amazon’s competitive advantage is the supply chain and that will only gain strength the larger they grow.

    I agree with you that AWS is overblown, but it’s an example of a by-product Amazon is monetizing because they developed economies of scale on their IT infrastructure.

  8. ypanet says:

    Great post. Real eye opener. Having control of magneto makes even more sense because eBay can push for deep integration between eBay and the retailer site. Easy access to eBay helps drive sales, eBay make their cut, everybody’s happy. PayPal was the real brilliancy though. Amazon will never be end to end without a strong payments solution of their own (the lack of PayPal on amazon is close to criminal)

    Now all eBay need to do is fix up the GUI. It’s not a matter of open or close. It’s just plain old (see etsy for a nice example of a decentralized GUI)

  9. There’s room for both approaches as each serves a different need but as you pointed out, end-to-end solutions generally provide a better customer experience – and that is the piece where Amazon excels, both with Amazon.com and its other investments (Zappos, Quidsi etc).

    My money will always be on the business that puts its customers above everything else. Throw in a whip smart, relentlessly focused founder in Bezos and you’ve got a formidable combination.

  10. jtfrazier2000 says:

    This would be a much more interesting argument if it was a pure play. eBay, often, doesn’t care about it’s customers (both merchants and buyers). It doesn’t do a good job making the shopping experience safe, respond quickly to hijacked accounts and or listening to feedback from its merchants.To me, its more like good customer service vs not so good. 

  11. “In the case of e-commerce, the platform approach could also have a price advantage.”
    I disagree. Apple has shown that centralized architectures have economies of scale that are hard to beat, especially if the company innovates like crazy and therefore has a large market share.

    It’s hard for a small retailer to get the price benefits that amazon gets when it orders in volume …… unless a platform helps them get the benefits of scale.

    I guess, what I’m saying is that, ebay should partner with alibaba, yahoo and a capable drop ship company.

    It looks like Mr. Thompson’s move was not so sudden after all.

  12. Chris

    Good luck trying to get ebay to operate like a platform.  There are many of us who spent years trying to push that thought within the marketplaces only to be rebuffed with phantom talk of cannabilization.  Until most folks get over the fact that the transaction doesnt have to happen on eBay.com to satisfy the buyer, you will never achieve true platform status.  Like a few of the comments here, I would argue that eBay is way behind Amazon from an open platform standpoint.  

  13. A nice benefit of the centralized approach is better access to data at scale and the insights you can take from it.  Amazon sees all the queries, all the page metrics (clicks and hover times), all the pricing of the SKUs (their price vs. third parties), etc.  This allows them to build better search (relevance tuning, a/b testing result pages), provide smarter collaborative filtering tools, and generally better optimize their merchandising.  It’s sort of the network effect of data signals.  Same reason why it’s hard to build a better search engine than Google – they see all the queries – so an alternative auto-complete product (for example) will always be lacking.

    The platform approach can still harness all this data, but it’s more difficult to stitch it all together.

  14. Anton Reut says:

    Tough to compete with Amazon if you’re selling commodity products. Those start ups you mentioned have unique products/offerings (most of the time) making it difficult for Amazon to match their inventory and compete on price via Prime and their marketplace reach. Typically they don’t rely on search as their primary cust acquisition channel so they’re less likely to price shopped when they have overlapping product sets.

    Amazon has built their user experience around being the lowest price. Start ups that position themselves around different value props or communities will continue to thrive. I am excited to see the rise of disruptive  “artisan” services (tailored shirts, glasses etc) that aim to “de-commodify”(can’t be a real word) products. I hope we see more of them.

    Would have been interesting to see if Diapers could have competed at scale as a standalone selling commodity consumer packaged goods. Maybe they knew something…

  15. clemnt says:

    Great post (as usual, don’t even need to write it). This difference of approach, I think, comes also from their DNA. eBay was born as a pure marketplace in 1995, it was the garage sales of the internet whereas Amazon was born as a pure ecommerce player.

    The genius of Bezos was to find the right mix of openness / controlled e-commerce model. Bezos always knew he needed 3P merchants to offer universal selection but that he couldn’t let the system as open as eBay if he wanted to guarantee the best buying experience for its customers. The customers can benefit the best of both worlds: the security, the confidence and a uniform buying experience (one shopping basket, same type of product pages wether it’s 3P or Amazon) from the “traditional” commerce model and the large choice from a marketplace.

    I have the impression that eBay was more stuck from his inception. With his image of “garage sales” where you could find anything, they couldn’t changed so easily their model… On one hand users still wanted this “openness” and on the other hand big brands did’nt really want to promote their products next to people who could sell their kidneys… So all in all they couldn’t really impose a model wich involve more “control” (At least it’s my impression)

    eBay suffered from this image in the last years (counterfeits, scams …) whereas Amazon din’t. I have the impression that the resurgence of eBay this year really came with their bet on technology, maybe they will go more and more on the “pure” technology provider side (magento, GSI, mobile, payment) and less and less on the e-commerce marketplace side. 

    Decentralized-centralized difference seems (to me) to strongly come from their DNA.

  16. Simon Lilly says:

    The long-term viability of the Marketplace is questionable though if Amazon continue to take the business away from the merchants and stock the products themselves. There has been ongoing specualtion that once a product or category reaches a certain level Amazon start to look at how they can have a bigger piece of the pie.

  17. willwhutson says:

    This interesting and does stoke some similar comparisons to Apply vs. Android but I have to say:

    Buying a device doesn’t solve what most people go online to purchase goods for.  Device sales have a lot of odd influences (like amazing design/social status etc) that online retailers will have a difficult time packaging and positioning.

    No one says “did you see so-and-so shopping on e-bay last Sunday?” like they’d do with say, Barney’s. Even as a user of the kindle (which I love) aws, and the most cult like thing they offer Prime (which I can’t use out of the US) I do feel your perspective–at least from a marketplace comparison–does have a real leg to stand on.

    This definitely puts their moves/logic/hires/acquisitions into perspective.

  18. dan barker says:

    This is a brilliant post, but I don’t think the reality is nearly as straightforward as ‘centralized vs decentralized’.

    For example, you can host a Magento store (ebay) using EC2 (amazon).

    On the basis of that, you could argue it the opposite way round if you chose to: “Ebay are forcing people to centralize on Magento, whereas Amazon’s EC2 platform supports lots of off-the-shepf ecommerce systems AND homegrown options”

  19. Nice one, Chris.  It seems to me that increasing value and sustainability in e-commerce is closely linked to increasing ownership of the “purchase process”.  By the purchase process I mean product search, discovery, education, payment, fulfillment and shipping.  In many cases Amazon owns and has competitive advantages in each step.  They’ve “flattened” the purchase process resulting in that enormous valuation.

    Conversely, eBay participates in a more vertical purchase purchase process where in many transactions they play a very small role — as you say, they’re simply “facilitating peer-to-peer commerce.”  

    There are hundreds of startups trying to compete and steal market share along each step of the purchase process (there are countless startups focused on payments alone).  

    By owning less of the purchase process eBay risks getting kicked out of it and becoming a glorified search engine.

    GSI on the other hand doesn’t directly participate in the purchase process, they power it from the back-end.  So GSI doesn’t risk getting kicked out of the process, but because they don’t “own it”, there is risk that what they sell becomes a commodity (if it hasn’t already).

    It won’t be easy but it seems Amazon is better positioned than eBay is to win this battle.  Given the P/Es listed above, the market seems to agree.

  20. sahilz79 says:

    The centralized vs. de-centralized framing reminds of Life Inc., by Douglas Rushkoff and Jaron Lanier’s commentary on the global-local flip (http://edge.org/conversation/the-local-global-flip). I am more inclined to agree with Chris Dixon on this, I don’t think its one model or the other, I think both have their pro’s and con’s. Having almost been a victim to more than one CraigsList scams I can attest to the limitations of the peer to peer model.

    However this is a very important topic in terms of how the world economy reshapes itself for the coming entrepreneurial era. We are seeing the emergence of a three tiered economy:

    Level 1:Comprising local merchants, artisans, free agents and even a vast majority of scientists; when 3D printing becomes mainstream we will see another boost to this sector of the economy; also combining AWS with increasingly sophisticated big data capabilities opens the room for data scientists doing some very sophisticated and interesting things at this level

    Level 2 (platforms):
    Cost minimization platforms: SohoOS, AWS etc. that enable level 1 players to achieve economies of scale in the past only possible to the largest players in the economy
    Monetization platforms: combined with revenue enablement through Mobility apps that enable the local discovery, as well as platforms like AirBnb, eBay and Amazon that enable peer-to-peer or B2C commerce

    Level 3:- where global capital markets, fiscal and monetary policy are organized- where the largest social/professional media platforms start serving as authentication, reputation and personalization engines
    - where the largest R&D and scientific projects – LHC, Space Exploration etc. operate 

    Every industry sector, including energy, as well as governments themselves as we have seen over the past 12 months, are susceptible to this re-organization. The hardest part of this change though is individuals realizing that they are largely responsible for charting their own professional lives, and the days of working somewhere for 3+ decades and retiring with a gold watch and a plaque ended when Ronald Reagan took office.

  21. NicolasVDB says:

    If you mean the influence (“synergy”) of AWS to Amazon’s  e-commerce business, I agree: they’re two unrelated businesses.
    But if you mean contribution, I think the opposite: AWS will become a larger and larger part of Amazon’s business, to the point where you’d call it its core business. AWS is simply changing the $1Trillion IT industry as we know it!

  22. yoavsch says:

    Based on Clay Christensen’s theory, when value chains are in their early stages the winners are vertically integrated companies. As those value chains mature, they become more modular/decentralized and profits move to specialized companies. This happened for instance as the PC industry’s profits moved from IBM (fully integrated at the time) to Microsoft and Intel. 
    Since eCommerce is still in its early stages this approach suggests that Amazon can succeed with its current model for a while, but at some point the eBay model will win.

  23. Nokia My says:

    A platform like Facebook will be a third way for e-commerce. In this sense, both
    Amazon and Ebay are old model and social/crowd sourcing e-commerce is the new

  24. Pingback: Trackback
  25. Pingback: Trackback

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s