A recent report from the Institute for Local Self-Reliance made the improbable claim that Amazon accounted for 46% of all online retail sales in 2016.
A quick look through Amazon's quarterly reports and US Census Bureau estimates will quickly debunk the claim, but it turns out there is a grain of truth here.
Released in late November, the report (PDF) reads like it was penned by David Streitfeld or published by Salon.com. The report is a bundle of misleading statements, misrepresented stats, and unsubstantiated claims, all wrapped up in a pretty bow.
For example, the report mentions last year's The Authors Guild survey on member income, and misrepresents it as showing the income of all author (including the 98% which don't belong to TAG). Later in the report the guesstimate that half of American households belong to Amazon Prime is stated as fact, and not an analyst's best guess.
I could go on, but I would prefer to focus on the most clickbaity factoid in the report rather than point out the spin the ILSR applied to its report.
Five years ago, Americans spent $170 billion shopping online, and Amazon accounted for one in four of those dollars. By last year, online spending had ballooned to $343 billion and Amazon’s share of the market had grown to 40 percent. This year, we estimate, it has soared to 46 percent.
According to Amazon's Q3 financial report, the retailer generated $18.9 billion in sales in North America. That comes to about 20% of the $93.672 billion spent online during that quarter (US Census Bureau estimates - PDF).
So obviously this claim is false, and easily proven to be wrong, yes?
While at first glance this report looks about as wrong as the BBC's claim that Amazon controlled half of UK online retail (a claim which was withdrawn once I caught them in the error), when the authors of the report were asked to show their work they presented a plausible explanation.
For overall online sales, we used the Commerce Department’s figures.
For the share that is Amazon, we added together their direct sales and the third-party sales on their platform (in the U.S. only).
Their direct sales comes from their annual report. We stripped out the other revenue, like AWS, so we just end up with the retail revenue. We also subtracted the part of their revenue that is commission on the third-party sales (using an average commission rate of 12 percent).
For third-party sales, we used quarterly data from Channel Advisor. Their figures are global, so we estimated the share of 3P sales that is U.S. versus foreign based on the breakdown of Amazon’s direct sales.
Okay, they still haven't shown their work, but the above explanation is not an unreasonable way to calculate how much is spent on Amazon.com (if you're willing to accept the estimates made by a group hostile to Amazon, that is).
That calculation doesn't tell us "Amazon’s share of the market", which is what the report said, but instead reflects the combined size of Amazon's marketplace and its own sales.
That is not the same thing, but it is also close enough that the dispute comes down to semantics rather than factual accuracy.
image by Phil_Parker