AMZN - Amazon.com Inc.

[frith2012]

It's interesting that people often cite Amazon's warehouses vs. Walmart's retail stores as a justification for a higher operating margin.  Walmart typically acquires their sites at a subsidized price, sometimes at less than developer cost, because of their ability to anchor the development.  Their presence enables the developer to secure a construction loan as well as lease off of their anticipated traffic.  As a result, they typically pay (either through rent or by purchasing a pad site) a rate well below market rents vs. other retailers.  As a result, I'm not as certain as many Amazon bulls their is a cost advantage in Amazon's real estate vs. Walmart.

[JAllen]

http://www.businessinsider.com/amazon-robotics-kiva-systems-2014-5

 

 

BTW, this video about Kiva is awesome.  Roboticized [sic] FCs are one of the reasons I think AMZN will have lower costs than pretty much all others that currently exist - because they have a head-start on automation and scaling vertically in various ways (software, bandwidth, delivery now).  This automation stuff is non-trivial to implement, to say the least.  It takes years to hone this stuff and get it right.  This particular video shows additional detail above the one that Kiva produced prior to the AMZN acquisition, but I'm not sure how much of that are AMZN's changes/improvements or they're just choosing to release more detail now.

 

 

A big question I've had is whether or not Kiva is still selling the pods to other retailers like they were prior to the AMZN acquisition, or perhaps AMZN is selling an older version to appease other retailers and prevent them from complaining.

[JAllen]

It's interesting that people often cite Amazon's warehouses vs. Walmart's retail stores as a justification for a higher operating margin.  Walmart typically acquires their sites at a subsidized price, sometimes at less than developer cost, because of their ability to anchor the development.  Their presence enables the developer to secure a construction loan as well as lease off of their anticipated traffic.  As a result, they typically pay (either through rent or by purchasing a pad site) a rate well below market rents vs. other retailers.  As a result, I'm not as certain as many Amazon bulls their is a cost advantage in Amazon's real estate vs. Walmart.

 

 

The increasingly-automated FCs designed to fulfill orders in a purely e-commerce mode are a small part of what I believe will be AMZN's OEM retail cost advantage.

 

 

I attribute the vast majority of the advantage I believe AMZN will have compared with WMT (and had prior to the accelerated expansion in 2010; 14-15% OEMs compared to WMT's 18-19%) to the lack of any retail real estate, not just automation advantages.  WMT has 1.1B square feet of retail square footage (still growing 3+%) to maintain.  Regardless of retail land costs, no one sells cement, copper, and other building materials to WMT for below the market price.  Retail land will always cost more than industrial or rural land where most of AMZN's FCs go (even in NJ their Fresh warehouse by definition has to be in an lower-cost industrial area [often blighted looking for redev], not a high-value retail/residential area).  Contractors may bid lower construction gross margins, but that's about it. 

 

AMZN has essentially zero retail real estate (save the lockers).  WMT doesn't have as much flexibility to move as AMZN does once a FC lease expires.  AMZN can say "we'll just move FCs to a different part of the suburbs" where there are many more options to choose from (for many, but not all of their FC locations) to the RE developers.  Where AMZN built their two recent Bay Area FCs, you can look around and as far as the eye can see there's empty land except for a few other warehouses - it's farm land out there.  The small rural towns are literally competing over having AMZN build a new FC there for jobs.  How many towns are dying to have a WMT go there?  I'm sure there are a couple but I've never heard of this.

 

So yeah, the lack of retail real estate is where most of the cost advantage will come from, I believe.

[Spekulatius]

AMZN has essentially zero retail real estate (save the lockers).  WMT doesn't have as much flexibility to move as AMZN does once a FC lease expires.  AMZN can say "we'll just move FCs to a different part of the suburbs" where there are many more options to choose from (for many, but not all of their FC locations) to the RE developers

 

AMZN does not own much real estate but they lease a lot. if I remember correctly, they lease about 94M sqft of real estate. I think that WMT has ~10x AMZN real estate (much of it is fully owned, at least in the US) but they also have 5x AMZN sales. AMZN business is not really that capital light, if you include all the leased hard assets.

 

In fact, I think WMT owning their real estate at a low average cost basis is a strong advantage, at least relative to leasing it.

[Palantir]

I think JAllen has made a pretty persuasive case for AMZN. If he is right, then it is a steal at this price IMO.

 

 

Long AMZN.

[dwy000]

What's the basis for the assumption that opex levels can (and will) return to the levels of 5 years ago?  As a % of revenues opex has been rising for 5 years now. 

 

Also, shouldn't you subtract out stock comp ($1.2bn last year).  It's not cash but it ultimately comes out of the shareholder's pocket. And it's been growing rapidly.

[JAllen]

What's the basis for the assumption that opex levels can (and will) return to the levels of 5 years ago?  As a % of revenues opex has been rising for 5 years now. 

 

 

Growth spending went from less than sales growth rate to three times the sales growth rate.  So from 2002-2009 FC space increased at a CAGR of 20%.  Since then it's increased at almost 60%.  This is a proxy for overall spending.  This is why I believe that OEM is inflated right now, not to mention the fact that AMZN should have a lower OEM than WMT long-term.  WMT's OEM is 18-19%.  I just don't see how AMZN's will not be lower without all of the expensive retail real estate.  They do exactly the same thing: buy tons of stuff from suppliers and deliver it to customers, except one company uses only much cheap FC space, that is increasingly automated.  The other uses FC space, but then keeps some of it in expensive buildings in dense cities and has a ton of employees running around managing that inventory and moving it around.  This is my basis for believing that AMZN will have sub-WMT OEMs.

[JAllen]

Also, shouldn't you subtract out stock comp ($1.2bn last year).  It's not cash but it ultimately comes out of the shareholder's pocket. And it's been growing rapidly.

 

Please read my previous comments about accelerated expensing of stock-comp.  It's a discretionary accelerated expense method that AMZN chooses to take to maximize expenses (clearly stated in 10-Ks).  Because stock-comp expense is accelerated and there are also forfeitures (not all employees stay four years for full vesting; AMZN's vesting schedule is 5, 15, 40, 40% FYI), stock-comp in any year is overstated, but by how much I'm not certain of.

 

One can easily account for stock-comp accurately by just diluting your FCF projections per share by some X% growth rate of shares outstanding per year.  I'd call it 1% or so.

 

AMZN has quite a proclivity of buying back shares at opportune times, so I'm hoping that gets to happen in a material way about every five years which could offset some of the dilution.

[dwy000]

Hi JAllen,

 

The accelerated spending on the FC's and other growth initiatives would be capitalized (as is much of their software spending) so there shouldn't be much one-off growth spending in the opex number.  As implied, this should just reflect ongoing operating costs. Are you aware of them putting one off growth costs in here?  If so, can you point that out for me?

 

On the share expense, yes, they are straightlining the expense and back end weighting the issuances.  Maybe I'm looking at this wrong but that's a bad thing in a rising share price environment. For example, if they award $100 of shares that vest 10-10-40-40 but expense them 25-25-25-25, the expense is higher in the first two years but then that reverses in later years. But more importantly the expense number is based off the share price in the first year (say $200) but with a rising share price the majority of those shares are issued at the higher price (say $300) so effectively they are under expensing vs the cost to shareholders of those new shares coming on the market.  Either way, over a 4 year cycle the expense differential will even out - and I think you have a longer term view than 4 years.

[dwy000]

Sorry, forgot to add:  my point on excluding share based comp was not about how it is expensed but the fact that it is dilutive unless you use a similar amount of cash (or more in a rising share environment) to repurchase the stock being issued.

[JAllen]

Hi JAllen,

 

The accelerated spending on the FC's and other growth initiatives would be capitalized (as is much of their software spending) so there shouldn't be much one-off growth spending in the opex number.  As implied, this should just reflect ongoing operating costs. Are you aware of them putting one off growth costs in here?  If so, can you point that out for me?

 

On the share expense, yes, they are straightlining the expense and back end weighting the issuances.  Maybe I'm looking at this wrong but that's a bad thing in a rising share price environment. For example, if they award $100 of shares that vest 10-10-40-40 but expense them 25-25-25-25, the expense is higher in the first two years but then that reverses in later years. But more importantly the expense number is based off the share price in the first year (say $200) but with a rising share price the majority of those shares are issued at the higher price (say $300) so effectively they are under expensing vs the cost to shareholders of those new shares coming on the market.  Either way, over a 4 year cycle the expense differential will even out - and I think you have a longer term view than 4 years.

 

 

 

I wrote about how AMZN also has growth expenses, not every investment is capitalized.  All of the non-shipping Prime benefits are expenses - digital video licensing, giving books away to Prime members, etc...

 

 

"Another thing that AMZN does well is to invest using expenses, not always capex, which as you know are immediately tax-deductible and help maximize long-term cash flows by minimizing taxes (remarkably similar to John Malone's modus operandi actually), the only thing Bezos cares about.  So the market thinks AMZN's underlying operations aren't profitable at all when in reality the profit is being obscured by the company's massive growth expenses, much of which are growth expenses.  A great example is the significant sum AMZN is now spending on HBO shows."

[dwy000]

I don't think that's correct.  According to the 10K, not only are content costs put on the balance sheet and expensed (that would include HBO payments) but the expense is consider a cost of goods sold, not an operating expense.

 

 

[txlaw]

This is an interesting read by the founder of Bonobos.  The post is essentially about Amazon and how one can or can't compete with it.  Nice to read an industry-insider's perspective and not an investor's or journalist's.

 

 

https://medium.com/what-i-learned-building/d233f02d52a5

 

This is an interesting read. 

 

However, there is a company I follow (and am invested in) called OSTK that does a pretty damn good job of making real cash profits, despite being an e-commerce company. 

[Spekulatius]

I wrote about how AMZN also has growth expenses, not every investment is capitalized.  All of the non-shipping Prime benefits are expenses - digital video licensing, giving books away to Prime members, etc

 

I don't get is. Why is the real cash that they spent on prime members or on HBO considered an investment by some? It's money that goes out the door to deliver a certain service. There is nothing tangible left after that spending, or do you believe that spending let's say 100M$ in cash for HBO creates higher profits in the future?

 

I agree it will entice members to sign up for the service as long as they offer it, but if they drop the service, the same members that like it for their very reason, would be likely to drop amazon prime.

 

If anything, offering something like HBO to Amazon prime members creates a liability to offer it in the future. I think the beneficiary of deal is HBO which likely get's a 100M$/year annuity.

[Palantir]

JAllen, what do you think of their growth slowing down? Do you expect the rate of decline to moderate, if so what are you estimating as your long term rate?

 

 

Their international operations seem comparatively slow, but I guess that's understandable as it does seem a very US-centric firm.

[JAllen]

I don't think that's correct.  According to the 10K, not only are content costs put on the balance sheet and expensed (that would include HBO payments) but the expense is consider a cost of goods sold, not an operating expense.

 

 

They do state that "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", so you're right about Prime Instant Video being in COS.  However, this doesn't really alter the overall thesis which is simply that AMZN should have a lower OEM than WMT long-termand because AMZN's gross margin has increased so much even with these extra Prime Instant Video expenses supports my GM - OEM = ~7-10% operating margin thesis seeing as AMZN's gross margin continues to increase y/y and q/q (28.8% 1Q 2014) and WMT's OEM is 18-19%.  So if AMZN doesn't achieve lower, just an equal to WMT OEM, we're talking 10% operating/EBIT/probably FCF too margins over time.

 

Another thing to add is that after re-reading everything about video and content in the 2013 10-K, it is not perfectly clear where all content related items are accounted for to me. There's the self-produced video which should absolutely be capitalized (STRZA capitalized and amortized its over three years if I recall correctly), and licensed content is probably also capitalized. In 'Operating Expenses'on page 26 however, they also list 'Technology and content' so I'm not sure if there's any video in there or whether that's music or books etc. 

 

Video isn't the only content AMZN is giving away for little if any revenue, there are book expenses too.  I should note that Cost of Sales does not mention Prime book lending, just Prime Instant Video.  There's also 5 years of Fresh Development, drone development, smartphone development, Kiva development, and hopefully lots of other things they're working on and incubating that we don't know about.

 

It's also hard to come up with a total estimate of how much operating expenses are inflated by - I just know AMZN generally prefers to throw money at customers than to pay 39% of operating income in taxes.  I doubt it's an accident they run at essentially break-even year after year.

[JAllen]

I think JAllen has made a pretty persuasive case for AMZN. If he is right, then it is a steal at this price IMO.

 

 

Long AMZN.

 

 

Glad we've managed to convince at least two people thus far. 

 

 

If you have an indefinite perspective and analyze the company with that lens, it's super attractive.  If you look back at the last handful of years, not the prior ten, and extrapolate the last four years into the future, well I can see how it's hard to imagine what the company would look like financially if they weren't spending so much.  But if you look back 12 years and then ask yourself which company is AMZN most similar to that is operating at a much more stable growth profile and what are that company's financials like long-term, then you've got quite an interesting opportunity to ponder.

 

 

I think the 'isn't earning money headlines' and not-for-profit stuff turn many people off in general from even glancing at AMZN's financial statements.  I know they did that to us for a number of years!

 

 

Newspapers, book publishers and smaller stores are already talking about the competitive advantages AMZN has - they will be obvious in hindsight - in ten years or so.  But they aren't now to many.

 

 

 

 

Why don't people think about the big picture here? What is the future of worldwide B2C product distribution?  How will it work?  Will it be store or Internet-based?  Which is more convenient and has lower costs?  Storing 20 million different items in a warehouse and ordering as many as you want in a few seconds with zero hassle or offering 150,000 in a WMT supercenter that requires driving for at least a few minutes, walking around a 4 acre store for the various products you need, waiting in line for a few minutes, then sitting in traffic for another 10-20 minutes.  Not to mention the gas expenditure for this (even if it's only a couple bucks, shipping is less if you're a Prime member and ordering doesn't take any time).

 

 

WMT is WAY behind in U.S. e-commerce; who wants to shop there online?  WMT has tried online grocery in the Bay Area.  They reported that 'people don't want it'.  Of COURSE people don't want to buy groceries online from WMT - WMT is known for cheap stuff and many people don't want cheap food.  The people that DO want cheap food aren't shopping online yet. 

 

 

If WMT is so far behind, who's going to catch AMZN?  I know Alibaba will try, but aside from ShopRunner, which is nearly the same price as AMZN with a fraction of the selection, and none of the content benefits, I don't see how AMZN will not dominate at least the U.S. e-commerce industry (hopefully cloud-computing too) over the next 1-3 decades.  AMZN is to WMT as WMT was to Sears and K-mart.

 

 

 

The simple fact that the company has had consistent positive operating cash flows should tip investors off to the fact that maybe, just maybe, AMZN could, or is, actually 'making money'.  Think of the companies that have little or negative FCF year after year but report GAAP income.  The headlines should actually state "AMZN reports positive FCF for the year and has every year for more than a decade", but they never do.

 

 

Here's to the very long life of the 'ablest CEO in America'.

[JAllen]

JAllen, what do you think of their growth slowing down? Do you expect the rate of decline to moderate, if so what are you estimating as your long term rate?

 

 

Their international operations seem comparatively slow, but I guess that's understandable as it does seem a very US-centric firm.

 

 

Gross profit is growing 33%.  Doesn't seem slow to me.  If GP was only growing 23% like sales, AMZN wouldn't be nearly as attractive.  I would like to make a long-term quarterly histogram of sales and gross profit growth.  Sales have slowed, but I'm quite sure this is because they are allowing third-parties to sell stuff, but they still earn high-margin commissions (try to sell a used book on AMZN to learn the commission rate).

 

 

But again, gross profit is what I care about (operating expenses too but they can be managed and inflated in periods of rapid growth).  Gross profit is a great single metric that shows your financial relationship with customers.  If you're a restaurant chain you could easily inflate your OEM by employing people to open new stores, opening them and having that 1-2 years of sales ramp up, but gross profit and its growth are a direct indicator of changes in your sales mix (e.g. increasing third-party FBA high-margin commission revenue and).

 

 

There is greater than sales growth revenue growth in third-party units, digital sales, and AWS (which if AWS is at all comparable to RackSpace has way higher gross margins than AMZN as a whole).  These are all causing gross profit to grow faster than sales.

 

 

Yes, I hope they can accelerate international.  Only 18% sales growth there, but the digital transition is just beginning there so hopefully things will improve and we will have a nice foothold in one of the massive countries (India or China).

 

 

I'll see what I can do about that sales growth histogram.

[Spekulatius]

Why don't people think about the big picture here? What is the future of worldwide B2C product distribution?  How will it work?  Will it be store or Internet-based?  Which is more convenient and has lower costs?  Storing 20 million different items in a warehouse and ordering as many as you want in a few seconds with zero hassle or offering 150,000 in a WMT supercenter that requires driving for at least a few minutes, walking around a 4 acre store for the various products you need, waiting in line for a few minutes, then sitting in traffic for another 10-20 minutes.  Not to mention the gas expenditure for this (even if it's only a couple bucks, shipping is less if you're a Prime member and ordering doesn't take any time).

 

This is true, but is AMZN the right vehicle to play this trend. There are already companies that benefit from this trend and are very very profitable. GOOG to find stuff online, MA and V to pay for stuff online and FDX and UPS to ship it. All these stocks have GAAP earnings, good growth and a decent valuation even by traditional metrics.

[moustachio]

JAllen, what do you think of their growth slowing down? Do you expect the rate of decline to moderate, if so what are you estimating as your long term rate?

 

 

Their international operations seem comparatively slow, but I guess that's understandable as it does seem a very US-centric firm.

 

 

Gross profit is growing 33%.  Doesn't seem slow to me.  If GP was only growing 23% like sales, AMZN wouldn't be nearly as attractive.  I would like to make a long-term quarterly histogram of sales and gross profit growth.  Sales have slowed, but I'm quite sure this is because they are allowing third-parties to sell stuff, but they still earn high-margin commissions (try to sell a used book on AMZN to learn the commission rate).

 

 

But again, gross profit is what I care about (operating expenses too but they can be managed and inflated in periods of rapid growth).  Gross profit is a great single metric that shows your financial relationship with customers.  If you're a restaurant chain you could easily inflate your OEM by employing people to open new stores, opening them and having that 1-2 years of sales ramp up, but gross profit and its growth are a direct indicator of changes in your sales mix (e.g. increasing third-party FBA high-margin commission revenue and).

 

 

There is greater than sales growth revenue growth in third-party units, digital sales, and AWS (which if AWS is at all comparable to RackSpace has way higher gross margins than AMZN as a whole).  These are all causing gross profit to grow faster than sales.

 

 

Yes, I hope they can accelerate international.  Only 18% sales growth there, but the digital transition is just beginning there so hopefully things will improve and we will have a nice foothold in one of the massive countries (India or China).

 

 

I'll see what I can do about that sales growth histogram.

 

I've read your posts in the past several pages and I think you make somewhat of a compelling case for AMZN. I haven't thoroughly looked into AMZN, but AWS is one thing that gives me pause and makes me question the thesis. Those data centers cost a lot, and much of the hardware seemingly would have to be replaced at a pretty quick rate. When I looked at Rackspace somewhat briefly in the past, it simply seemed too capital intensive to be attractive.

From AMZN 10K:

"Depreciation is recorded on a straight-line basis over the estimated useful lives of the assets (generally the lesser of 40 years or the remaining life of the underlying building, two years for assets such as internal-use software, three years for our servers, five years for networking equipment, five years for furniture and fixtures, and ten years for heavy equipment). Depreciation expense is classified within the corresponding operating expense categories on our consolidated statements of operations."

 

Unlike depreciation for a lot of types of capex, three years might actually be pretty close to the useful life for servers, might it not? I wonder what percentage of their capex is being spent on datacenter hardware, and how much of it will be recurring. In short, because of AWS I think some of their growth capex might become recurring maintenance capex, which might make your thesis a little less compelling.

[Palantir]

Can you believe AMZN was $37 per share in November 08?  :o

[PatientCheetah]

http://www.businessinsider.com/the-future-of-retail-2014-slide-deck-sai-2014-3

 

a lot of interesting data points

[JAllen]

Can you believe AMZN was $37 per share in November 08?  :o

 

 

We had sold some naked puts in early 2008 but covered them as AMZN declined instead of letting ourselves be put the stock.  Oh well!  At least we wised up after only five more years... 

 

 

Here's to hopefully at least five more great years, though they won't be an 8 bagger.

[Fat Pitch]

This is turning into a great thread. JAllen got me to think about Amazon in a different light. I went ahead and looked back on 10yrs+ of the company financials and I do see strong evidence the company is stuffing their growth capex into the expense items: Fulfillment, Marketing, and Tech & Content. This will in turn reduce Operating Cash Flow and give the perception the company is grossly overpriced.

 

% of Revenue          2013                 04-06 Avg

Gross Margins –    27.23%                    23.3%

Fulfillment –            11.53%                    8.75%

Marketing -              4.21%                    2.46%

Tech & Content –      8.82%                      5.4%

G&A -                        1.52%                    1.86%

Net Op                      1.15%                    4.83%

 

If you strongly believe these costs will revert back to pre-ramp up spending someday then it is easy to see Operating Margins of 9% or even higher. For those concerned about FCF generation... it's there, but it is hidden in Bezos clever accounting.

 

My only concern is that the increase in these cost centers are permanent just to keep their edge on the competition in perpetuity so that’s the risk.

 

[dwy000]

Fat pitch - I agree that IF they can get back to opex levels from 10 years ago they will be very cash flow positive.  But what I don't get is why the belief that the expense levels of 10 years ago (when growth was much faster than today) should be considered the "normal" level.  What is the evidence that today's expense levels are not the "new normal"?

 

There seems to be a suggestion that they are stuffing growth expenses into the operating expense line.  What's the basis for this?  I can't find anything in the 10Ks or Q's.  And top line growth is somewhat slower than it was 5-10 yrs ago.  Also, this is a bit of a different beast than in 06.  They now manufacture hardware and their fastest growing business (AWS) is a high opex operation which would suggest the higher expense levels are less than temporary.

 

Again, IF they can revert to historic levels this can be very profitable but trends and evidence suggest it's headed in the other direction.  Curious as to why the belief this can and will reverse sharply?