What’s Amazon’s market share? 35% or 5%? — Benedict Evans

These assumptions give us $81.5bn US ecommerce, but you could very easily flex it $5bn up or down.

Now let’s come back to that ‘third party’ thing. Amazon lets other companies list products on its website and ship them through its warehouses as the ‘Marketplace’ business. It charges them a fee for this, and it reports the fee as revenue. That’s the $42.745bn. But, Amazon doesn’t treat the value of the actual purchases as its own revenue, which is in line with US accounting rules, since technically Amazon is only acting as an agent. So, if you buy a $1,000 TV on Amazon from a third party supplier, Amazon will charge the supplier (say) $150 in fees for shipping and handling and commission, and report $150 as third party service revenue, but won’t treat the $1,000 as Amazon revenue at all. eBay has the same accounting policy, but it also reports the value of the actual purchases as a separate number – ‘gross marketplace value’, or GMV.

In the past Amazon didn’t disclose its GMV, but in the 2018 shareholder letter it finally gave a bunch of rounded numbers: global first party sales were $117bn (note this is slightly different from the $123bn above, given in the formal accounts in the same report), and global third party sales were $160bn, giving global GMV of $277bn and meaning that 58% of global Amazon ecommerce is actually made by third parties using its platform, not by Amazon itself. To repeat: Amazon doesn’t treat that $160bn as its own revenue and so it’s not in the revenue tables above – it only reports the $43bn of fees for handling it. But clearly, this belongs in any discussion of Amazon’s market share.

(As an aside, one should note that the revenue from the third party platform services business is almost twice that of AWS, and certainly pretty profitable; Amazon discloses revenue and profitability for AWS but only revenue for this. Remember that the next time you see someone claiming that AWS subsidises Amazon.)

So, we need to consider both Amazon’s own sales and the value of third party sales (together, GMV) in market share, and again we know the global number but not the US number. However, if we assume the same 66% split, and switch to using the $117bn ecommerce number instead of $123bn, that means that in the USA in 2018:

• Amazon sold $77.5bn of products itself,

• And also sold another $106bn for third parties,

• Giving a total US GMV in round numbers of roughly $184bn. 

Second question: $184bn sounds like a big number, but how does that compare to the competition? What market share would that give Amazon? 

The simple answer is that the US government gives a number for total ecommerce sales as an economic statistic: in 2018 the number was $522bn. Hence:

• Amazon’s first party business had about 15% market share of US ecommerce

• The third party business had about 20%

• And the total GMV had a 35% share. 

• (As a backstop, if the 66% US/international split is too low and Amazon has more of its ecommerce than that in the USA, treating global GMV of $277bn as though it were all in the USA would give Amazon a total US ecommerce share of 53% – in other words, it must be significantly lower than that)

Splitting out GMV is important because Amazon isn’t setting the price or choosing the selection for the third party Marketplace. This is especially relevant for any conversation about predatory pricing: Amazon is setting the price directly for 15% of US ecommerce, not 35%. On the other hand, some of those third party products will be competing with products that are sold and priced by Amazon, and setting their own prices accordingly. Life is complicated.

However, the problem with comparing Amazon’s online sales with all online sales is that this presumes that Amazon only competes with other online retailers. It presumes Amazon doesn’t compete with malls or department stores or Walmart or with any other shop. It presumes no-one ever asked ‘should I buy this from Amazon or from Barnes & Noble’? That seems like a problem.

So, what if we look at Amazon as part of US ‘retail’ instead of part of ‘online’? Now we have more numbers to choose from. Amazon US GMV plus the US physical stores (which we need to include now since we’re comparing with physical retail) was roughly $200bn. As for US retail: 

• Total US retail in 2018 was $6tr, but that includes bars and restaurants, and though some of the restaurant business is clearly moving to online ordering and delivery it’s not there yet, so we should probably exclude this.

• Gasoline stations had $511bn of revenue (co-incidentally almost exactly the same size as ecommerce) and it seems unlikely that Amazon will ever sell gasoline. It also doesn’t sell cars or repair them, and car dealers and parts were $1.2tr. This market will change radically over the next few decades as we go first to electric and then to autonomy, but for the moment we should exclude this as well. 

• That leaves ‘addressable retail’ (i.e. excluding cars, car parts, gasoline stations, restaurants and bars) of $3.6tr in 2018.

• Hence, Amazon US retail revenue of $200bn was about 6% of US addressable retail. 

• (Incidentally, this means that $522bn total US ecommerce is about 15% of US addressable retail.)

Is 6% market share a lot? Well, how big is Walmart? Walmart’s US revenue in 2018 was $318.5bn. So, in the USA in 2018, Amazon was a little less than two thirds of the size of Walmart.