Raghav Toshniwal

Home Projects Blog Contact

Why Amazon Wouldn't Survive in the Long Run

The 10x better lens

In this 1995 clip, David Letterman is interviewing Bill Gates about this new thing called the “internet”. It is telling how Letterman completely misses the point and how useful the internet would turn out to be. He ridicules the internet, stating how live broadcasting is akin to radio and the on-demand part is equivalent to taping something.

The internet did turn out to be very useful, but before that it had to get a lot better. Most disruptive technologies start out looking like a toy. Therefore a good measure of wether a technology is useful is to figure out how impactful it would be if it was 10x better.

If the underlying technologies that power e-commerce became 10x better, I think Amazon1 would find it hard to survive.

E-commerce primitives


Selection - They need to have a range of products that the customers might consider

Social Proof/Trust - There needs to be some level of trust/social proofing for the customer to know what they’re buying2

Infrastructure - Everything from payments and fulfilment to physically deliver the goods to the customer

All of these things are individually complex to solve and Amazon has delivered on all three.

The Everything Store

In 1993, when Jeff Bezos was tossing around the idea of internet based commerce while he was at D E Shaw, he called it - The Everything Store. He wanted a marketplace to connect manufacturers with sellers directly. Manufactures would drop-ship the products and the customers would enjoy lower prices The idea was starkly different from what ended up being Amazon.

It was a much purer, ideal internet business. He got it wrong on several counts though, there wasn’t enough selection on the internet. Trust was very hard to solve where you would not own neither the buyer side/seller side of the experience. Above all, the logistics for internet enabled commerce was simply non-existent.

Amazon initially won by solving for selection, they had the world’s “largest book store”, they solved for the long tail of customer needs. People weren’t using Amazon because it was easy to use, or for their shipping times, they used it because they would not find the books they were looking for elsewhere.

It wasn’t until Barnes&Noble tried to launch their own Amazon competitor that the ever prescient Bezos invested heavily in solving for infrastructure.

Over time, Amazon added so many categories (selection), launched Prime (trust), launched 2 day shipping (infrastructure), strengthening their moat in all these areas.

Everything else is getting 10x better

The internet has changed quite a bit since Amazon launched. People have taken pieces of Amazon and built is an enabling layer for other e-commerce operations. Commerce on the internet sans Amazon has improved by tremendously in the last decade.

Selection - Given enough time, this is perhaps the easiest moat to lose. The few categories, where selection is a bigger problem (apparels), already have specialty portals giving a much wider range. We’re still in the early days of Amazon category unbundling. After we have unbundled more long tail categories, there would be attempts to consolidate various categories.

Trust / Social Proofing - Amazon has tons of legitimacy, and a wealth of reviews/existing customers. Prime continues to give better value to its customers and the lock-in seems real. Would take a lot for customers to switch to a different offering. However, given a better offering/experience, people should start switching because there are no network effects for Amazon. Nobody uses Amazon because their friends also use it.

Infrastructure - Amazon has 300+ Fulfilment Centers, 73 planes, an Ocean Liner and millions of employees to manage the mammoth logistics network. This is what enables fast delivery and efficiency. On the margins though, cracks are beginning to show. There are logistics providers that take up different parts of the problem and can seamlessly integrate with each other. First and Last Mile provider, Middle mile providers, warehousing partners, picking & packing companies, payment infrastructure providers are all coming together to rival the Amazon experience.

Bezos’ original dream of The Everything Store seems more and more plausible, it was just a few decades premature. It seems unreal to have so many players can coordinate with each other and provide the unified/seamless experience that Amazon can provide. However, if the success of D2C is are any indicator, people are figuring out the

Counters and Counters to Counters

A list of reasons why I might be wrong but also why I don’t think I am.

  • Economies Of Scale - Amazon has amassed so much volume that everything is cheaper for them. For ex - They buy packing tape by the truck-loads at pennies on the dollar.

    However, in the real world, there is a physical cap to economies of scale. Things do get cheaper but beyond a point there is a diminishing effect to cost.

  • Headstart - Things do get better but Amazon would always stay ahead of the pack, they would witness the challenges first-hand, therefore everyone will continue playing catchup.

    I think this holds true till it doesn’t. As more number of companies that try their hand at unbundling Amazon, a subset of them would get close to matching or bettering Amazon on metrics central to their existence.

  • The Flywheel Effect - Amazon has the most sellers because it has the most buyers and therefore it can offer the cheapest prices and attract more sellers and more buyers and on and on.

    Amazon holds inordinate power over sellers, this is because they have customer loyalty and cost efficiency. They have absolutely no lock-in with sellers and once Amazon’s unbundled alternatives reach cost parity, sellers and customers would jump ship. Among the big-tech companies, Amazon has the worst network effects. Nobody is using Amazon because their friends are using it too.

How long till this happens?

I don’t know.3

In his famous 1999 letter to shareholders Bezos claimed that upto ‘15% of retail commerce may ultimately move online.’ We crossed that threshold in 2022.

Cracks are starting to show, there are now logistics companies, payment infrastructure companies, various other utilities that make internet commerce better and Shopify to tie them all.

Before someone comes and aggregates various e-commerce plays under one roof, various parts of Amazon would have to be unbundled. Today, D2C brands can provide a commerce experience that rivals Amazon in customer delight. The unbundling of Amazon is happening right now.

  1. I am referring to Amazon’s e-commerce business only here, they have AWS and bunch of ancillary bets that might pan out. 

  2. Flipkart’s big innovation with Cash-on-Delivery was solving for the non-existent trust that Indian Customer had on internet commerce 

  3. On a long enough timeline, the market capitalisation of every company tends to zero. It is intellectually dishonest to not commit to any timelines when making these claims.