Everyone already knows how Amazon.com (AMZN) reported a loss, guided lower, part of the loss was due to a charge from LivingSocial and things like that. I am not going to cover that again. What I am going to do in this series of articles is I am going to dive deeper into Amazon.com's latest 10-Q and bring out a few details which have gotten little or no press.
In this first article, I will start with gross margins. I have two theses regarding Amazon.com's gross margin:
1) Amazon.com's gross margin cannot be calculated the way it has been calculated both by analysts and Amazon.com itself;
2) Given the public data, Amazon.com's gross margin can indeed be calculated, but only for products where Amazon.com is the seller of record.
Gross margins are not what they seem
In the last two quarters, Amazon.com's gross margins have been presented as a reason to go wildly positive in the stock in spite of plunging earnings and plunging earnings estimates. This is what analysts in general saw - a large jump in gross margins due to a changing mix brought about by faster growth on AWS and third party sales (3P):
It is my opinion that analysts are going gangbusters over something misleading. The reason is simple - it makes no sense to calculate the gross margin that's being calculated and said to be increasing!
This is so because Amazon.com does not divulge the cost of sales for services sales. Indeed, Amazon.com thus accounts for AWS and 3P sales at 100% gross margins, so with these two components growing faster than overall sales and increasing their weight in the mix, the reported gross margins are naturally expanding.
However, this is wrong, because neither AWS nor 3P sales really have 100% gross margins. What happens is that Amazon.com includes the cost of sales of these activities lower in its P&L, below cost of sales. Amazon.com includes AWS's cost of sales in the "technology" line and several 3P costs in "fulfillment" and "marketing." Another easy way to see this is by looking at Equinix (EQIX), a business comparable to AWS. Does it report 100% gross margins? No - it reports 50% gross margins. Certainly AWS is no different. In short, gross margin including services cannot be calculated because the analysts - and indeed anyone outside Amazon.com - do not have the cost of sales for services.
Another problem is that Amazon.com has taken to the practice of selling some relevant products, such as the Kindle eReader or the Kindle Fire, at 0% gross margins. Obviously, when sales of these products increase, gross margins as a whole suffer, and when they decrease (as they did in early 2012, especially Q2 2012), gross margins increase. It makes no sense to applaud Kindle Fire and such when it seems to be doing well, and then applaud gross margins when Kindle Fire seems to be doing poorly.
Finally, even Amazon.com recognizes that gross margins are a bad indicator of its own health due to the way costs are spread over its P&L. This is from Amazon.com's 10-Q:
We believe that income from operations is a more meaningful measure than gross profit and gross margin due to the diversity of our product categories and services.
The true gross margins
But the reason I am bringing up gross margins is not just to explain them away like that. No, we have something new and rather more interesting to see here. You see, recently Amazon.com has started breaking down its revenues between "product sales" and "services sales," and this is what goes into each category as per the 10-Q (bold is mine):
Net sales include product and services sales. Product sales represent revenue from the sale of products and related shipping fees and digital content where we are the seller of record. Services sales represent third-party seller fees earned (including commissions) and related shipping fees, digital content subscriptions, and non-retail activities.
Also we know that the line "cost of sales" on Amazon.com's 10-Q includes only the cost of the product sales (hence the way it gets inflated by the changing mix, as I wrote earlier). Here's the description from the 10-Q:
Cost of sales consists of the purchase price of consumer products and digital content where we are the seller of record, including Prime Instant Video, inbound and outbound shipping charges, and packaging supplies. Shipping charges to receive products from our suppliers are included in our inventory, and recognized as cost of sales upon sale of products to our customers.
Knowing this, we can do an interesting exercise; we can actually calculate how the gross margin on Amazon.com's product sales has been evolving over time, for the time period where Amazon.com gave us this information. This is what we get (remember Q4 always present seasonally lower gross margins):
Or, if we calculate TTM for the quarters where we have the data (Q4 2011 - Q3 2012) to eliminate the seasonality, plus the 2 years (2010, 2009) reported in 2011's 10-K (the large drop from 2010 to Q4 2011 TTM also happens because of the absence of Q1 2011 TTM to Q3 2011 TTM due to lack of data, the drop was smoother in reality):
And there you have it. Not only is there no gross margin increase in sight, but the product gross margins are actually decreasing steeply and consistently over the timeframe where we have information on them. And certainly gross margins are not decreasing "because Amazon is investing in the future"!
Indeed, we might even go further. What would have been the impact on Amazon.com's EPS if there had been no deterioration in gross margins since 2009 or 2010? The table below calculates that:
The impact was dramatic! Which leads us to a very simple conclusion - contrary to the popular explanation put forward by Amazon.com and the analysts, Amazon.com has seen earnings deterioration not because it is "investing in the future," but because its product gross margins have suffered severely over the last few years, as the implosion in product gross margins more than accounts for the entire drop in earnings!
By gathering whatever little detail Amazon.com puts out about its own operations, I was able to reconstruct the true gross margins for part of Amazon.com's business, namely the part regarding product sales (physical and digital) where Amazon.com is the seller of record. What this has shown beyond any shadow of a doubt is that this part of Amazon.com's business is seeing decreasing gross margins over time. This ugly trend cannot be blamed on "investing on the future." In part this trend can be due to the introduction of low margin products such as the Kindle Fire, but even that fails to explain the already evident deterioration that was taking place well before Q4 2011.
This gross margin deterioration also accounts for ALL of Amazon.com's earnings drop since 2009/2010, making it unnecessary to believe in the "we are investing in the future" theory. A drop in product gross margins cannot be blamed on investment. It can only be blamed on an attempt to grow sales faster.
On the other hand, it is impossible to estimate the gross margin for AWS and 3P sales. The cost of sales for both those businesses is spread across Amazon.com's operating costs. It is of great intellectual dishonesty to just account for those revenues at 100% gross margin in an effort to paint Amazon.com in a more favorable light, as most analysts have been doing.
Amazon.com remains a deteriorating business on many accounts, where the heavy investments can only be partly responsible for the deep plunge in earnings pictured below (I do believe heavy investment in AWS might be responsible for some deterioration).
There will be a couple more articles on this series as I have uncovered other interesting effects having a misleading influence on Amazon.com's financial reporting, namely affecting (inflating) its operating cash flow, EBITDA and capex (which means part of Amazon.com's investment isn't what it seems).
Disclosure: I am short AMZN.