Amazon's revenues came in slightly below consensus ($15.7 billion vs. $15.73 billion lowered consensus). More relevant was Amazon's guidance for Q2 2013, which was as follows:
Third Quarter 2013 Guidance
Net sales are expected to be between $15.45 billion and $17.15 billion, or to grow between 12% and 24% compared with third quarter 2012.
Operating loss is expected to be between $440 million and $65 million, compared to $28 million in third quarter 2012.
This guidance includes approximately $340 million for stock-based compensation and amortization of intangible assets, and it assumes, among other things, that no additional business acquisitions, investments, restructurings or legal settlements are concluded and that there are no further revisions to stock-based compensation estimates.
As I predicted before the earnings were released, this constitutes another guide-down for Amazon's revenues. The midpoint of the guidance falls at $16.3 billion whereas present consensus sits at $16.98 billion. I'd expect consensus to be revised lower to around $16.6-$16.7 billion or so.
More amazing, though, is yet another guide-down on operating margins, with Amazon.com now estimating negative operating earnings even at the high end of its guidance. Obviously there's a measure of sandbagging involved, but it's still surprising that Amazon.com would guide so low. As well as that it would have no impact on the share price.
Comparison to my model
The model predictions compared to what Amazon actually reported as follows:
In what regards my own modeling, where I use my model for both short term and long term predictions, the major differences were:
- Product margins came in at 13.0% versus my 11.2% assumption. I believe there were differences in accounting treatment having an effect on this particular variable, but since the 10-Q does not highlight those differences (though we're aware that they exist), it's impossible to quantify the effect;
- GMV margins came in at 14.2% versus my 13.8% assumption. My long-term assumption is 13.0%, which I will revise towards 14.2%. It's noticeable that the increase in GMV margins was much lower than the increase in product margins - I believe that the GMV margins better represent the move in margins which existed, and the product margins were further helped by an accounting change affecting at least part of AMZN's ebook sales;
- The product/GMV margins were the main reason why gross margins came in 5.4% ahead of projections. Some of it was due to volume in "other" revenue, some of it was accounting changes. Impossible to tell how much is each;
- Technology costs, which came in 12.8% above my estimate. However, this implied a ratio of Technology/Other revenue of 177.8% versus my Q2 2013 assumption of 178%, so really close to my estimate. This probably means that the 12.8% difference came from volume inflating costs. My long term assumptions will remain the same;
- Marketing costs, which came in 7.7% above my estimate. This implied a ratio of Marketing/GMV of 2.14% versus my assumption of 2.00%. My long term assumptions will remain the same;
- G&A costs, which came in 7.4% above my estimate. This implied a ratio of G&A/GMV of 0.91% versus my assumption of 0.85%. Since my 2013 assumption is 0.80% I will increase the long term assumptions to 0.85%;
- Although Fulfillment costs came in just marginally (2%) above my expectations, I will increase long-term assumptions by 0.05%, since there seems to be a small bias towards those costs slowly creeping up, which is consistent with a wider SKU selection and more numerous fulfillment center network.
My own long-term model already implies that technology will get better (less costly) over the long-term. G&A has some volatility so it won't imply much of a change.
All in all the cost relationships held quite well, with most of the deltas coming from unexpected volume and expected - but unknown - accounting changes. The major difference in product margins is certainly down to accounting changes, and the overall revenue number also appears to have been inflated by those. However, since rules and accounting treatment didn't change -- what changed were the very nature of the transactions -- Amazon.com did not explain these impacts. This makes it hard to estimate exactly what the impacts were as Amazon.com is not transparent enough.
Revised long-term model
My long-term model was based on product margins. However, since accounting changes are affecting these more than GMV margins (since accounting changes shift margin from 3P towards 1P while GMV margins remain more stable), I am shifting the long-term model towards using the GMV margins model.
Taking into account the differences explained, my revised long-term model now predicts the following:
Basically by shifting to GMV margins the model now predicts losses almost as far as the eye can see, except for 2013, where it sees a below-the-Street $0.76 EPS.
Due to the shift from 3P to 1P in some ebook sales, Amazon.com's Q2 and following quarters will be significantly affected. And since Amazon.com does not detail the accounting effects, although we know these effects are there we cannot know for sure how they're manifesting themselves except for the obvious inflation of revenues as revenues that used to be accounted at 30% or so commission now get accounted at the full 100% sales price.
Another effect is that this accounting change shifts some margin from services towards products. This explains at least part of the jump in product margins during Q2 2013. But we don't really know how much of it is explained by the accounting change, since again, Amazon.com does not tell us.
Amazon's earnings report brought nothing new. It was yet another miss and guide-down, of which Amazon.com has done around 11 in a row. The Street continues buying not Amazon.com as it today, but some mythical, hugely profitable, Amazon.com that's supposed to exist down the road.
With GMV (Gross Merchandise Value) already above $100 billion, it's very unlikely that Amazon.com's margins will expand tremendously. As I've shown, when Wal-Mart (NYSE:WMT) was investing as much as Amazon.com and growing even more, its margins never suffered. Also because of the very large GMV it's likely that the cost relationships to GMV (Fulfillment, Marketing, SG&A) will remain stable, thus preventing a large increase in margins down the road. Indeed, as it stands my own long-term earnings models for Amazon.com turned negative.
There was nothing in the report to change my opinion that Amazon is a clear short which will never produce enough profit to justify the levels it trades at. I expect this report to lead to another round of downward estimate revisions in terms of revenues and EPS. These revisions are systematic because the long-term models the Street uses do not respect the stable cost relationships that I have identified in my articles.