Where the money is the value investing in digital age 重點摘要 (二)

 

I doubted it. For one thing, since its IPO Amazon has consistently said that the operating margins of its online retail business should at maturity be 10% to 13%.II For another, Walmart has more than 10,000 stores to maintain; Amazon has only 500 Whole Foods stores, a handful of Amazon branded stores, and fewer than 1,000 distribution centers. How could Walmart be three times more profitable than a company operating with less than 10% of a physical presence?

Either there was something structurally broken with Amazon’s online model, or the company was underreporting its true earnings power. It was clear to me, both from instinct and by making comparisons to comparable companies, that the latter was true. Other, less ambitious e-tailers such as eBay report 25% operating margins

At this point, I could have just slapped Amazon’s 10% to 13% long-term profitability goal on the entire e-commerce segment, or I could have imputed eBay’s 25% margin to Amazon. Instead, I decided to dig another layer deeper. The company had given me discrete revenue—but not profit—disclosure for five of its commerce divisions. If I worked through the numbers of these subsidiaries and tried to tease out a profitability profile for each one, perhaps I could generate a more accurate estimate of Amazon’s earnings power. So that’s what I did.

Common sense told me that this segment’s margins should be at least equal to Walmart’s 6%. Several calculations seemed to back that up. Walmart’s depreciation expense, the proxy for how much it must spend to maintain its physical plant, amounts to 2% of its annual sales. Amazon has no physical customer traffic in its virtual store, so it was logical to assume that Amazon’s depreciation expense is de minimis by comparison. Moreover, because it operates online, Amazon does not have to worry about “shrink,” the euphemism retailers use for shoplifting. While Walmart works hard to minimize theft—those “greeters” aren’t at the door just to welcome you—last year Walmart lost roughly $5 billion to sticky fingers. This amounts to one percentage point of margin lost to shrink.

If we assume that Amazon’s margins begin at Walmart’s 6%, and then we add two points of depreciation that Amazon doesn’t have to incur and one percentage point for shrink, Amazon’s online retail earnings power becomes 9% of sales. That’s half again as profitable as Walmart and roughly in line with Amazon’s low double-digit goal.

In the end, I assigned a 10% operating margin to Amazon’s legacy online store. Ten percent is just above the 9% margin I had calculated and at the low end of Amazon’s long-term 10% to 13% aspiration.

Putting a 10% margin on $140 billion of core e-commerce sales generates $14 billion in earnings power. That is an interesting figure: it nearly equals Amazon’s entire 2019 reported operating profit.

( 把推算出來的營業利益率10%, 導回去計算網路e-commerce 部門140B , 得出來營業淨利14B, 很接近2019財報顯示的數字) 

我特別上去查証當初2019年報, Operating income 欄位顯示14,541( 百萬) , 換算14.5Billion

2. and 3. Physical stores and subscriptions. When I looked at two of Amazon’s smaller segments, physical stores and subscriptions, it was clear there would be no such dramatic changes. “Physical stores” refers mainly to Whole Foods Markets, which Amazon bought in 2017. Before its acquisition, Whole Foods was reporting operating margins in the 5% range, higher than the average grocer and consistent with Whole Foods’ premium brand. Because this segment was less than 10% of Amazon’s total corporate sales, however, it didn’t matter whether I assigned physical stores a 2% margin, a 5% margin, or a 15% margin. The difference wouldn’t have been big enough to affect Amazon’s earnings power in a major way.

Like other major tech platforms, Amazon views its various subscriptions businesses not as profit centers but as ways to bind customers to its platform. To use the old grocery term, they run subscriptions as loss leaders, low-margin products that they know will draw people into their stores—and keep them coming back. Paying $139 a year for Amazon Prime delivery strikes many people as a good deal, but getting a free video service thrown in makes it even better. Amazon spends billions each year to populate its streaming service with compelling new shows. Because the company charges nothing for it, some might consider this spending wasteful. Amazon, however, views it as a good investment. Offering subscriptions is an attempt to build a “switching cost” moat. Having Prime Video makes customers less likely to drop their annual subscription for e-commerce delivery, where the company does make money.

Given all these puts and takes, it was hard for me to know how profitable Amazon’s subscription segment was. On the one hand, Prime delivery brings in billions of annual revenues. On the other hand, free shipping costs Amazon a lot, and much of Prime’s revenue goes straight to buying video content. Alibaba, one of the few tech platforms that reports operating profits in its subscriptions segment, runs this business at a loss. On the other hand, it doesn’t have customers paying it $139 a year for delivery. Given this, I thought it was fair to assume that Amazon’s subscriptions business ran at roughly breakeven, so I assigned this segment margins of zero.

4.

Third-party seller services. When Amazon began online, it bought and sold merchandise the old-fashioned way: it purchased a book from a publisher, or a CD player from a manufacturer, then marked the product up and sold it for a profit. Gradually, however, Amazon opened its website to other merchants who lacked their own online presence, and this service proved enormously popular. In 2000, such third-party seller services represented only 3% of Amazon’s gross merchandise sales, but by 2015 the figure surpassed 50%. Today, roughly two-thirds of all sales on the company’s platform come from outside merchants.

These third-party sellers pay for the goods they sell on Amazon, and this arrangement has material implications for Amazon’s profitability. When Amazon sells protein bars or a selfie stick on its website, but a third-party merchant pays for the goods, Amazon avoids the single biggest expense of a retailer: buying the merchandise itself. Moreover, Amazon charges these merchants a fee for the right to transact business on its platform. Because Amazon dominates e-commerce, merchants are happy to—or are forced to—pay for access to Amazon’s eyeballs.

By turning its storefront into a place where the goods are paid for mainly by other merchants, Amazon has become a platform company. Like the fees Apple charges app developers, Amazon’s third-party merchant fees have almost no expenses associated with them.

As a result, like Apple, Amazon has transformed itself from a relatively low-margin “hardware” business into one that can generate levels of profitability usually associated with software companies like Facebook and Alphabet. To use Buffett’s metaphor, Amazon has become a toll bridge, and in 2019 it collected $54 billion in tolls from third-party merchants

What are the margins associated with this revenue stream? Amazon discloses nothing about it, but I figured the margins had to be higher than traditional e-commerce. Amazon doesn’t pay for the goods themselves, and the fees it charges merchants are pure gravy. However, Amazon stores, packs, and ships most third-party orders, and that costs a lot of money.

eBay is a pure third-party reseller and, as I mentioned earlier, enjoys operating margins of 25%. Was 25% the right profitability figure to assign this segment? On the one hand, eBay has almost no logistics operations. It leaves the shipping to its third-party merchants, and this fact suggests that margins for Amazon’s third-party segment might be lower than 25%. On the other hand, eBay’s single biggest expense is sales and marketing. Without Amazon’s market power and instant brand recognition, eBay must spend 25 cents of every revenue dollar it earns to promote itself. Amazon has no such need.

In the end, I decided to give Amazon’s third-party business segment the same 25% operating margins as eBay. eBay’s higher marketing costs, I estimated, were roughly equivalent to Amazon’s higher distribution costs. This estimate was nothing more than a guess, but it struck me as a sound one. If Amazon’s legacy e-commerce margins were 10%, its third-party business was likely two to three times more profitable.

Assigning a 25% operating margin to Amazon’s third-party revenues of $54 billion generated roughly $14 billion in operating profit. This is another interesting number: by assigning reasonable margins to both Amazon’s traditional online segment and its third-party selling segment, I found that Amazon’s earnings power was roughly double the company’s entire reported operating profit. And I hadn’t yet analyzed the company’s most profitable segment: advertising.

 “Other”—i.e., advertising sales. Amazon has turned its ubiquitous web presence into another powerful toll bridge. With the company averaging 90 million online visits per day, its website is becoming an increasingly popular place for companies to advertise. Amazon reports ads in its “other” segment, and in 2019 this segment generated $14 billion of revenue. According to a footnote on page 68 of the company’s 2019 SEC filing, most of this revenue comes from selling ads.

This business is surely even more profitable than Amazon’s third-party selling service. When selling goods for others, Amazon must still pay for the extra labor and the incremental distribution space associated with the merchandise. Advertising, on the other hand, takes place in the virtual world. Beyond perhaps paying a team of engineers to configure the ads, it doesn’t cost Amazon anything at all.

What is the margin on this business? It likely approaches 100%. If, for argument’s sake, we say that Amazon incurred $1 billion in annual engineering costs to run the ad segment, that would equate to a 93% profit margin. On the other hand, perhaps some of this “other revenue” was very low-margin. I doubted it—but to be conservative, I assigned this segment a 50% profit margin. This estimate added another $7 billion to Amazon’s earnings power.

What were the net results of my earnings power exercise? As you can see from the table that follows, Amazon’s reported e-commerce profits were $5.3 billion—but its earnings power was $35 billion, or nearly seven times more. 運用R&D 計算出來的

When combined with projecting Amazon’s revenues three years into the future, this adjustment had dramatic implications for the price the market was asking me to pay. The multiple on reported 2019 earnings was eighty-seven times, but the multiple on 2022’s estimated earnings power was fifteen times. That moved the earnings yield from 1% to 7%.

If, having worked through this exercise, you find the uncertainty in both financial statements and valuation unsettling, then I’ve done my job. As an investor, you must get used to a certain degree of haziness. The world is uncertain, the future is uncertain, and what counts as earnings is uncertain.

This lack of exactitude drives engineers like my son crazy. Engineers live by precision, and rightly so. A single line of bad code corrupts an entire program, and a millimeter off in the placement of a jet engine’s fuel nozzle means the plane will malfunction. As anyone familiar with them knows, however, income statements and balance sheets are filled with all sorts of estimates.

Treating financial statements as gospel is especially dangerous in the Digital Age, because GAAP’s old-economy bias distorts them so badly. The truth, however, is that beneath their neat, well-ruled surface, a company’s financial statements have always been somewhat squishy. Bad debt reserves, warranty expense, depreciation expense—all of these are not precise calculations. They are approximations that companies must make, and this leaves considerable room for fudging. “By making excessive or insufficient allowances for these items,” Ben Graham wrote in 1937 about the depreciation reserve, “the net earnings may readily be under- or over-stated.”