Amazon - An Outstanding Company at an Outrageous Price

Amazon claims to be Earth's most customer-centric company. And they have evidence to back that claim: American Customer Satisfaction Index consistently ranks them amongst them top when it compiles its list.

A laser-like focus on customers has enabled them to grow like a weed. Here are some measures of growth over the last 10 years:

Revenue/Share: 26% CAGR
Operating Income/Share: 11% CAGR
Book Value/Share: 27% CAGR

I believe that by having this customer-centric approach, Amazon has created a lot of value in the world. They have connected many entrepreneurs with customers through Marketplace, developers with datacenters through AWS and aspiring authors with readers through Kindle Direct Publishing.

Most of the created value is redeployed back into the business, at "high" rates of return, as claimed by CEO Jeff Bezos. So the company doesn't have much of an accounting profit (it has massive depreciation expenses from prior years' CapEx). I don't believe that the company will have an accounting profit for the next 15 years or so. Whenever it makes any efficiency gain, it passes those savings to customers. Without an accounting profit, it is difficult to place a value on Amazon.

How does one value Amazon? We can get some hints from the shareholder letters. I believe that Miss Market is valuing Amazon on the basis of cash flows far into the future (15+ years probably):

"Why focus on cash flows? Because a share of stock is a share of a company’s future cash flows, and, as a result, cash flows more than any other single variable seem to do the best job of explaining a company’s stock price over the long term." -2001 Letter to Shareholders

Amazon had operating cash flows of $6.8B, $11.9B and $16.4B in 2014, 2015 and 2016. I believe these do not take into account stock-based compensation, which is a real expense. Also, it is difficult to project these far into the future (given that they are consolidated figures). So let's try a sum-of-parts valuation instead.

Let's break down the business into: Retail (North America and International) and AWS.

North America had $80B of sales in 2016 and $2.3B of operating income. Growing at 25%
International had $43B of sales in 2016 and $-1.2B of operating income. Growing at 24%
AWS had $12B of sales in 2016 and $3B of operating income. Growing at 55%.

Optimistic Scenario after 15 years

Let us assume the current growth rates are sustained over the next 15 years.

Total Retail will grow to $3.5T.

Note that the total world retail spend is around $23T. Assuming it grows at 6% CAGR, it will be $55T in 15 years. So Amazon will be 6.3% of total global retail spend.

Also, Amazon US retail will be $2.2T and if US retail spend is $7T in 15 years, Amazon will be 31% of all US retail spend.

Assuming an operating margin of 5%, operating income from retail will be $175B.

Current rate of AWS doesn't seem sustainable to me (44%), so let's assume it will also grow at 25% CAGR for 15 years. Sales of AWS will be $341B in 15 years. Assuming an operating margin of 20%, operating income will be $68B.

Assuming multiples of 10x operating income for retail and 15x operating income for AWS, the total valuation comes out to be $2.7T.

The current market cap is $700B, so the rate of return is around 10% CAGR.

"Realistic" Scenario after 15 years

I suspect growth rates of 25% are not sustainable for 15 years. To me, something like a 15% sales growth for retail and a 20% sales growth for AWS seems more realistic (though it is no small feat).

Assuming the same multiples as described in the "optimistic" section above, the market cap in 15 years will be: $1T. Given the current $700B market cap the return will be a meager 3% CAGR.

Valuation Matters

15% sales growth for retail and 20% sales growth for AWS for 15 years is excellent. Any other company would be very happy to hit 15% sales growth sustained over 15 years. The reason for the meager returns in the "realistic scenario" above are because of valuation contraction.

Given the current market cap of $700B, it seems to me that Miss Market is putting a multiple of 100x on AWS operating earnings and 70x on hypothetical retail "operating earnings" at a 5% operating margin.

So the negative CAGR for "valuation contraction" that Amazon as a stock has to fight is -12%. So while I am bullish on Amazon as a business, I am not sure whether there is enough margin of safety for Amazon as a stock. Anything positive CAGR that Amazon as a business has to produce will have to overcome the -12% CAGR valuation contraction.

What else is Miss Market thinking?

Why is the multiple so high for Amazon? I suspect Miss Market is thinking about Amazon entering new businesses and applying the same "customer-first" mantra to seize market share. If she is not thinking about Amazon creating new businesses, she must be assigning a very high growth rate for retail and AWS.

From Jeff Bezos' commentary, it seems they only want to enter businesses that are highly scalable and have high returns on capital. Online retailing seems to pass that bar while physical retailing does not:

"I often get asked, “When are you going to open physical stores?” That’s an expansion opportunity we’ve resisted. It fails all but one of the tests outlined above. The potential size of a network of physical stores is exciting. However: we don’t know how to do it with low capital and high returns" - 2006 Letter to Shareholders

So apparently back in 2006 physical retailing did not pass the "returns bar" for Amazon. If we take Walmart as an example of a retailer with operational excellence, we can try to estimate the height of the "returns bar". Walmart's ROIC is around 12% or so. So I am guessing Amazon will only enter an industry where the ROIC is greater than that. My hunch is that their bar is perhaps 15%-20%.

A study by McKinsey ranks industries by median ROIC. We can look at industries whose median ROIC is > 15%: pharma, consumer products, software services and media. Amazon has already entered software services and media. They could enter consumer products or pharma next. To me consumer products seems more likely as pharma tends to have a wider range of returns on capital than consumer products. They already have the AmazonBasics brand that they can build up upon.

Facebook: megacaps can grow too

Facebook as a business owns the facebook.com website along with some other social media properties like Instagram and WhatsApp.

The main revenue driver is the facebook.com website. Within facebook.com, the most important revenue source is the news feed.

The news feed is a list of articles that is customized to each user that shows the activities of their friends. Sprinkled within that news feed are ads that are also tuned to the user's likes.

The numbers

2016 Revenue: 28B

2016 Net Income: 10B

Net Margin: 35%

2017 Estimated Revenue: 36B

2017 Estimated Net Income: 17B

Net Margin: 47%

This is an extremely capital-light business. 2016 property and equipment was only 9B. They have 0 long-term debt.

The fundamentals behind the numbers

Net Income is a function of the following variables:

MAU = monthly active users

ARPU = average revenue per (monthly active) user

Net Income = MAU x ARPU x net margin

MAU is about 2B.

Worldwide ARPU is about $18 currently. In the US ARPU is close to $75.

Net Margin is about 47%, which is outstanding.

In the base case, in 5 years:

MAU goes up to 3B, i.e. ~8% CAGR

ARPU goes up by 5% CAGR to $23 (global GDP goes up 2%; 3% is secular shift from traditional to digital ad spend)

Net Margin expands due to operating leverage by 1% CAGR.

Valuation comes down from 25x to 20x, i.e. -4% CAGR

Share dilution of about 3% CAGR

In this scenario, the net income will increase by 14% CAGR. Due to valuation decrease and share dilution, shareholder return may only be 7%.

In a realistic scenario, I think ARPU will go up more than 5% CAGR. To me, 10% or so CAGR seems realistic so the final ARPU would be $28 in 2023. This may be because of higher ad load, or more tuned ads. Also, the valuation may not shrink (and if it shrinks, I suspect it will be because the benchmark valuation also shrinks). In this "realistic scenario" the earnings CAGR would be 19% and the shareholder return will be 12%.

In optimistic scenarios, worldwide ARPU can go up to $50 or more. While this can happen, it is unlikely due to more users joining from geographical locations where ARPUs are much lower than worldwide ARPU. Here are the 2017 approximate ARPU figures by geography:

US&Canada: $75

Europe: $24

Asia-Pacific: $8

Rest of World: $6

Side Note: There is quite a disparity between US and Europe, which cannot be explained by difference in GDP per capita alone. Europe / US GDP per capita = 65% while Europe / US FB ARPU = 32%. Perhaps this difference is cultural (Europeans may have fewer friends, or are less interested in friends' activities). It would be interesting if Facebook provided some insights into these trends in their annual reports.

The next billion users to be added will surely come from Asia-Pacific and Rest of the World. US MAU growth rate is ~5.7% while Asia-Pacific is ~26% and Rest of World is ~15%.

The drivers behind the fundamentals

We can dream up any ARPU or MAU figures, but ultimately these will be as a result of happy users engaging with the facebook.com platform. The platform must provide meaning and value to users' lives in order for users to keep spending time on it.

Ultimately, users want platforms to stay in touch with friends and family. So that will likely not change in the future. The relevant question to me is what platform they will use to stay in touch with friends and family. While there are many choices (Snapchat, Facebook, Instagram, Google+, etc.), I think facebook.com will remain the platform of choice because of network effects. Users will want to stay there because their friends, photos, comments and other digital memories are there (likes, stickers, etc.). Starting a new social network is not easy (think of Google+) because your friends aren't already there.

How does one estimate future ARPU? I think the key factors involved are:

1. average time spent per user per year
2. average "engagement" (think of this as how much the user enjoys looking at the posts on their feed): are users skimming through the feed, or are they looking at it with concentration
3. number of ads shown in the news feed (though this knob has quickly diminishing returns)
4. quality/relevance of ads shown in the news feed (ads should be tuned to the user's tastes)
5. how much the advertiser is willing to pay for an ad

(1) and (2) depend on the content of the news feed which is user-created. Facebook only has indirect control over those and can provide tools to improve on this (like auto tagging, filters, etc.). Facebook has direct control over (3) and (4) and especially (4) is likely to increase a lot in the future due to advancements in machine learning. Plastering too many ads by increasing (3) can actually cause (1) and (2) to go down, so that knob may be saturated. Growth in (5) can be attributed to two components: GDP growth and advertisers moving their budget from TV/newspapers to facebook.

At the minimum ARPU will grow inline with GDP + secular trend of ad budgets moving from offline to online.

Other factors to consider

Other positive or negative factors to consider are:

1. Buybacks may increase in the future. Current buyback is rather insignificant at $6B which is ~1% of market cap.
2. Company is run by the CEO who is an owner-operator. I like that he sacrifices short-term results for long-term results.
3. WhatsApp is totally unmonetized right now. This may take a long time to monetize though.
4. Management may get distracted with money-losing acquisitions or ventures. I thought paying $20B for WhatsApp was a very high price but if they can have an ARPU of $2 there (with 50% net margin), they paid 20x earnings which seems reasonable.
5. Longevity in the technology business is rare.

Bottom Line

FB is currently a great company and the price of the stock also seems reasonable at 28x forward earnings (where the S&P500 is trading at 19x).

My estimate of forward S&P500 returns is around 6% CAGR and FB should be able to beat this by 5% CAGR as long as user engagement doesn't drop.

ROST and TJX: Recession-proof retailers

Two of the strongest retailers I know in the US are ROST and TJX. The business model is simple: buy branded goods by the truckload at dirt-cheap prices, and sell them at a discount to the department store prices. All but the most premium brands have a tendency to over-produce. ROST and TJX buy merchandise when department stores need to "close out" items. They then hold or distribute them to their outlets and sell them at large discounts.

Because the inventory is very eclectic (a full selection of sizes and designs is not available), it creates a "treasure-hunt" like environment in the store. From personal experience, females in my household love shopping at Marshall's (owned by TJX) and Ross Stores (ROST) "you never know what you may find." (I guess we are a family of "value" investors).

These two publicly traded companies are killing it and the numbers do not lie:

ROST

EPS 2007 - 2017: 0.42 - 2.72 (21% CAGR)
Net Income 2007 - 2017: 242 - 1081 (16% CAGR)
Shares Outstanding 2007 - 2017: 568 - 398 (-3% CAGR)
Store Count 2007 - 2017: 890 - 1446 (5% CAGR)
Payout %age 2007 - 2017: 15% - 20%
Return on Equity 2007 - 2017: 27% - 42%
Net Margin 2007 - 2017: 4% - 8%

TJX (2007 - 2017):

EPS: 0.78 - 3.41 (16% CAGR)
Net Income: 738 - 2287 (12% CAGR)
Shares Outstanding: 960 - 669 (-4% CAGR)
Store Count: 1623 - 2223 (3% CAGR)
Payout %age: 17% - 27%
Return on Equity: 35% - 52%
Return on Assets: 13% - 18%
Net Margin: 4% - 7%

These numbers are excellent. I am really surprised that both these retailers have thrived in our competitive, capitalist world without attracting a lot of competition. Side note: I have noticed another slightly higher-end off-price retailer in my local area called Fox's. I will be keeping an eye on when it goes public.

What's even more impressive is that both these retailers increased Net Income and EPS during the 2008-9 recession. This business model really is superior to "regular" retailing branded clothing/accessories.

How/Why did the growth come about over the past 10 years?

Net Income growth can be attributed to store count growth, same store sales growth and margin expansion. For ROST, these 3 factors contributed roughly 5% CAGR annually.

The "why" is the most important question but I do not have any good insights into why competition hasn't moved in to eat the lunch that TJX and ROST are enjoying. I suspect a combination of these comprise of the competitive advantage:

1. Scale advantage. About 10 years back, the net margins of TJX and ROST were 4% and now they are at 8%. An incumbent would have lower net margins than 4% at the beginning. I suspect this keeps competition out.
2. Established relationships with branded department stores. An incumbent will need to build these relationships again from scratch.

The Future (next 10 years)

Where will things be 10 years from now?

I think ROST store count could be will be around 2200 (4% CAGR). Net margin will be 10% due to higher operating leverage (2% CAGR). Same store growth will be 2% CAGR (totally made up). Therefore Net Income should compound at around 8%.

To compute total return, we need to add dividend yield (1%) and buyback yield (3%) for a total of 11% total CAGR over 10 years (assuming valuation doesn't change).

Management

This is a noticeable dark halo around these two companies. Management compensation is, IMHO, excessive. ROST pays its executives 35M out of 1B total profit, which is 3% "management fee." TJX pays 60M from 2.3B total profit which is also around 3%.

I wish management paid itself less than 3% (something like 1% seems reasonable to me), especially because forward returns don't seem very high compared to the past. But then again, most companies are not like CHKP (where the CEO makes minimum wage).

Valuation

Currently selling for 24x trailing and 21x forward P/E. I suspect valuation will decrease over the next 10 years which will dampen total return.

From the universe of retailers, these would be good candidates to hold in a portfolio.

Subjective Thoughts on the Business Model

I like businesses that create value in the world. Most successful, sustainable businesses in the world create a lot of value for their customers. These two businesses certainly create the world a better place. Without ROST and TJX, department stores would be forced to trash their excessive inventory. Consumers who love to wear branded goods may also be forced to pay a high price for them. ROST and TJX enable "value" consumers to wear branded clothing, etc. at a fair price. It's a win-win for everyone.

Novo Nordisk: Making money while saving lives

Novo Nordisk is a pharmaceutical company based in Denmark. It sells drugs to manage diabetes, obesity, hemophilia and growth hormone disorder. Let's look at the numbers:

2006 - 2015 Revenue Growth: 15% CAGR
2006 - 2015 EBITDA Growth: 19.2% CAGR
2006 - 2015 EPS Growth: 22.9% CAGR
2006 - 2015 Debt: Negligible
2006 - 2015 Net Margins: 16% - 32%

The numbers are salivating. What's even more impressive is that the company has been paying about about 20% - 45% of net income as dividends. So they definitely have way more cash than they can put to use.

The Why

I don't have a very complete picture into why they have been so profitable in the past 10 years, but here is my hypothesis:

1. Oligopoly: The insulin industry has only a few players and NVO is the biggest one by market share. Along with NVO, there is Ely Lilly and Sanofi. NVO supplies about 50% of the world's insulin by volume. So this is an oligopoly with rational pricing.
2. Scale Advantage: being the biggest player has its advantages. Net income has increased more than revenue because of margin expansion because the fixed costs in this business are limited.
3. Barriers to Entry: This is because of a few reasons.
1. It is unattractive for a smaller competitor to come in because prices for insulin are already quite low (it is a high-volume business).
   
   “It’s important to know that the insulin market is a high-volume, low-price business compared with many other pharmaceuticals,” says Jesper Brandgaard. - 2012 AR
   
   In the insulin market, biosimilar competition is not a new phenomenon for Novo Nordisk. For decades there have been and still are local insulin producers in some countries, but their attempts to grow internationally have so far failed. The insulin market is very different from those for other biologic medicines, Jesper Brandgaard reiterates: “Insulin is a high-volume, low-price business, so it doesn’t have the same appeal to biosimilar producers as other biologic medicines. Even when some of the modern insulins lose their patents it will be very difficult for biosimilar producers to achieve the economies of scale that established insulin producers have. Furthermore, to be successful a new producer with global ambitions must also be able to deliver the products in sophisticated pen systems, which further adds to the up-front investments needed and increases the total manufacturing costs.” - 2012 AR
2. Some of the newer insulins have patents like Tresiba.
3. It takes a lot of R&D and government approvals, etc. to bring a new drug to market. It's just not worth it for other companies to enter this market.
   It is interesting that their most popular drugs (like NovoLog/NovoRapid) are actually off-patent. And yet there is little generic competition. Why?
   The reason for this based on chemistry: insulin is a long and complex protein. Unlike small molecules that have a fixed chemical formula that can easily be identified and copied, insulin does not. In fact, two batches from the same manufacturer can have molecules that are different. So, in order to prove safety to the FDA, generic manufacturers have to go through the approval process all over again (they cannot re-use the proven safety of a small molecule drug). I believe this is the reason we have not seen generic insulins even though it was discovered in 1922.

As volumes have increased, their margins have expanded from 16% to 32%.

Industry Trends

The industry trends are very favorable looking to the future. 387 million people have diabetes today vs. ~700 million estimated in 2035. According to the "Rule of Halves" mentioned in the annual report, about half of the diabetics in the world are diagnosed, of which about half receive care, of which about half achieve treatment targets.

So there is a long runway ahead of this company.

Why I think profitability will persist in the future

We invest for the future, so it is important to assess future profits. I think NVO will continue to be highly profitable because:

1. Barriers to entry as mentioned above. Specifically, since insulin is not a small-molecule drug, we will probably not see a generic anytime soon. Since its discovery in 1922, there has not been a generic insulin on the market in the US.
2. Newer drugs like Tresiba have patent protection until at least 2022, and they have a pipeline of future drugs.
3. As volume increases, they should be able to increase margins even more.

Capital Allocation

This is a business that is a cash gusher. So what do they do with the cash? About 40% of Net Income is given out as dividends, and some of it is used to repurchase shares. Share count has gone down 21% in the past 10 years. R&D is already taken care of in Net Income and CapEx is quite low (10% or so of Net Income).

Management Quality

I like management with discipline, and I see some signs of that in their annual reports:

"In May 2010, the Greek government announced temporary price cuts of up to 27%. As a consequence, we made a decision to temporarily withdraw some products from the Greek market" - 2010 AR

Also, management is paid about 1% of Net Income, which is not too bad.

Valuation

Because this is such a great business, it has always been valued at a premium to S&P500. However, due to a recent sell-off, it is trading at around the same multiple as the S&P500 (20x earnings). I think it is a great time to be a shareholder of this wonderful business.

CHKP - Selling virtual security

I'll start with the 10-year figures:

2006-2015 Revenue: 575M - 1630M = 11%

2006-2015 Net Income: 278M - 686M = 9%

2006-2015 EPS: 1.17 - 3.74 = 12%

2006-2015 Net Debt: 0 - 0

These numbers look quite impressive. Even though net margins have decreased over the years, there is still solid revenue and net income growth.

Note that this company does not capitalize software development cost and I like that. I could not find glaring accounting shenanigans in my brief look. Contrast this to Cerner, where the real net income drops a lot if you do not capitalize software development cost.

What they do

Checkpoint sells security-related software to businesses: firewalls, public and private cloud security software, etc.

Capital Allocation

This company is engaged in my type of acquisition plan: acquiring partial ownership of a business they understand very well. Aka, buying back their own shares. This is much better than acquiring terrible speculative software companies that do not produce profits. Note that they have made acquisitions of other companies, but at a small scale.

Because the company is very capital-light (capex is 3% of net income), it can use almost all the net income (and then some) to repurchase its own shares.

CEO

Gil Shwed is the founder/CEO of this company. He seems passionate about this field and holds several patents himself. I really like his compensation:

"Mr. Gil Shwed, Chief Executive Officer and Director. Cash compensation expenses recorded in 2015 consisted of $14.0 in salary expenses, and $12.7 in benefit costs. Mr. Shwed requested to forego his salary and bonus for 2015, as he has done for the past several years." - 2015 Annual Report

He also owns 17% of the company.

The future

This is a company in a fast-changing industry (technology). Moats in this industry are fleeting and companies need to be nimble to protect them.

I believe this company can stand out from the crowd because of the following reasons:

1. CEO is brilliant and thinks like an owner.
2. They have patents.
3. They have relationships with customers.

However, a complete change in landscape (like Cloud offerings including security bundles) may take away the pie this company has been consuming.

Valuation & Final Thoughts

Currently going for about 20x earnings. I think I may bite and take up a small position that I can increase in case the price moves downwards.

ODFL - An oasis in the trucking desert

Old Dominion Freight Line (ODFL) is a trucking company. They focus on transporting less-than-truckload (LTL) loads for their customers. Their motto is "premium service at a fair price."

First, a look at some numbers:

2005 - 2015 EPS CAGR: 20%
2005 - 2015 Revenue CAGR: 10%
2005 - 2015 Operating Income CAGR: 18%

2005 Debt: 129M
2005 Equity: 345M
2005 Net Income: 53M
2005 Net Margin: 5.7%

2015 Debt: 134M
2015 Equity: 1685M
2015 Net Income: 305M
2015 Net Margin: 10.1%


Management has not given out any dividend or done any significant repurchases or acquisitions in the past 10 years. All growth has been organic. Management are outstanding operators.

The trucking industry is generally very lousy: it is capital-intensive, cyclical and undifferentiated. Profit margins and return on equity is very low.

ODFL is different because it provides a slightly premium product and excellent customer service. It's return on equity has been > 15% for many years.

Capital Allocation Policy

From their 2015 annual report:

We have three priorities for our capital allocation strategy. Our first priority has been, and will continue to be, investing in the organic growth of our business through our ongoing capital expenditures for capacity and technology.
Our second priority will be to continue to evaluate strategic acquisition opportunities in an industry experiencing increasing consolidation pressure. Finally, we remain focused on continuing to return capital to our shareholders through our stock repurchase program in order to increase total returns. 

ODFL 2015 Annual Report

They have not done any acquisitions recently which shows patience and discipline.

Risks

I think this company has key-person(s) risk. Earl Congdon is 84 years old (Chairman) and David Congdon (CEO) is 58 years old. The Congdons own ~15% of the company and are owner-operators. Without an owner-operator at the helm, I fear return on equity will decline significantly.

Another key risk is that of a recession. This company did not fare well in the 2008 recession.

A note on capital intensity

I have noted in the past that I like businesses that are not capital intensive. So why pick this business to study?

It is because this business has a long ramp. Capital-light business are my preferred type of businesses to study and "acquire" in my "conglomerate", but typically those businesses cannot reinvest their income back at high returns. So, although they return most of their cash back to owners, the net income typically does not increase by that much.

If the economy does not hit any bumps, ODFL can reinvest income at 15% returns for a long time (10+ years), given management runs the business well.

Leaving the question of price aside, the best business to own is one that over an extended period can employ large amounts of incremental capital at very high rates of return. 

Warren Buffett, 1992 Shareholder Letter

Visa - A toll road on the world's consumption

Visa is a technology company that acts as a middleman between banks when transactions happen via their cards.

Business Model

When you swipe a visa card, the following happens:

1. The seller's bank takes the buyer's card number, and gives it to Visa.
2. Visa contacts the buyer's bank, checks the account balance and authorizes the payment.

So, by sitting between the seller's and buyer's banks, Visa enables the transaction to take place. In exchange for this, it gets a small fees. And boy, do those small fees add up. In 2015, Visa recorded 7B of income. In 2010, the income was 3B. Per-share income rise was even higher (22% from 2010 to 2015).

Business Stats

2010 - 2015 EPS CAGR: 22%

2010 - 2015 Net Income CAGR: 16%

Net margin in 2015: 50%

CapEx / Net Income in 2015: 7%

This is an extremely capital-light business gushing out lots of cash flow. Bulk of the cash flow goes to the owners, to repurchase shares.

Outlook

Electronic payments account for only 15% of the world's total transaction volume. Cash and check are still prevalent in many parts of the world.

Visa and Mastercard have a lot of room to grow. My suspicion is that both Visa and Mastercard will grow at 10% or so CAGR for at least 10 years. The EPS may grow faster because of share repurchases.

Eventually, it may be a winner-takes-all situation, or an oligopoly. Or perhaps both will be replaced by some other form of electronic payment. I am not smart enough to figure that out.

Summary

Both Visa and Mastercard are amongst the finest businesses in the world. They are like a toll road on the consumption of the world with very low maintenance cost. I would love to own either.