Friday, October 2, 2009

Green Mountain Coffee Roasters (Nasdaq: GMCR): A Dangerous Short

<<<< Spiderwebs without and with caffeine, from Wikipedia. GMCR continues to be one of the most shorted stocks in the market but I think it is a dangerous game to play. I have never worked in the coffee business but I have a thesis. I can't say "go long"- at least not right now- but I think there are some dangers to shorting them. Here's why: At first, let's discuss why GMCR is so heavily shorted (something like 32% of the float as of 9/10!). One, according to some views, it is overpricing the future earnings growth of the company. Second, the stock has returned 165% for the last 52 weeks, and 1,310% for the last five years which is nothing short of phenomenal for a "consumer staple" stock. And, yet, it keeps going up, aided by short squeezes here and there. Third, there are the more qualitative views on competitive entries from allegedly better competitors (Kraft, Sara Lee). I am sure there are more reasons (i.e. algorithmic trading shorting runner-ups and the like) but I am a complete dilettante in those areas. Now, let's take a look at the coffee business overall. (1) Caffeine is an addictive substance. It is often mixed with sugar, high fructose corn syrup and various flavors to induce additional intake by the consumers. The FDA just banned flavored/"sweetened" cigarettes (but menthol is somehow ok, NAACP, where is the outrage?) but, fortunately for the caffeine purveyors, caffeine is still widely available in a wide range of flavors from energy drinks, colas, teas to coffee. There are certain benefits to dealing in addictive substances, as Altria (NYSE: MO), Philip Morris International (NYSE: PM), Reynolds American (NYSE: RAI), Lorillard (NYSE: LO) and others can tell you. The three big ones are (1) repeat business, (2) relatively inelastic demand and (3) juicy margins. As a matter of fact, the CPG company with the highest margins until recently was UST (US Tobacco), makers of various brands of moist snuff (aka dip), such as Copenhagen. They were at something like 70% gross and 50% EBITDA! Compare these margins to, say, Microsoft's, a near monopolist. And UST, unlike MSFT, was all about one thing: returning cash to shareholders via dividends and buybacks. I just wish MSFT would stop throwing good money after bad with their failures, and separate the Windows and MS Office business in a nice, slow-growth, stale toll-collecting company that pays big dividends and does not waste money on bricks like Zune or antiperspirants for Steve Ballmer. But I digress. Unfortunately, for us, the cold-blooded profiteers, UST was swallowed up by Altria as a part of their "total tobacco" platform.

While caffeine is not as addictive as nicotine, there are similarities, especially when mixed with sugar and established in a ritual ("having my morning coffee", anyone?) just like smokers are looking forward to taking a "smoke break" as a part of their ingrained daily behaviors. There are similar mechanisms of addiction related to food but this is a whole different topic: it is not only the biological/neuro-chemical response is my point.

(2) So let's expand on the "juicy margins" point from above.

Coffee is a commodity. It costs about $1.30 per pound these days. "Better varieties", whatever that is, cost a bit more.

Roasted coffee, Green Mountain brand, was $6.50 for 12 oz in the local supermarket yesterday. This is $8.67 per pound (vs. $1.30ish for the green commodity)

GMCR are the main maker of the Keurig single-cup "k-cups" but I was not able to find online the weight of the portion packs used in the machine. Why do you think they hide it? I called customer service to find out the product weight, and they quoted me a weight in grams (8.5-9.5 grams for most varieties). Call it 9 grams, which is ~.32 oz. A k-cup retails for about 50 cents a piece. There are about 50 9-gram k-cups per pound, or in other words, a pound of coffee is now selling at $25 per pound. Yes, there is more packaging and all that but still.

So let's review the numbers: $1.30/lbs commodity to $8.67/lbs roasted on the shelf to $25/lbs in the form of a k-cup. This is called "value added." Putting more colloquially, someone is getting really fat here, and it sure isn't the farmers and it sure isn't the coffee drinkers, either. No wonder when you order online, you never see the net weight, just cup count per box!

(3) Let's also expand on the "repeat business" and "inelastic demand" from above.

One of the simplest business models out there to ensure repeat purchases is the so-called "installed base" model. The most famous historical case is Gillette: the story is, if I remember correctly, that he gave away the safety razor but he sold the replacement disposable (!) blades. The model still works well for them, 100 years later. You can see the "installed base" concept everywhere: printers/ink cartridges, iPods/iTunes, soap/cleaning dispensers ("free" from the likes of 3M and EcoLab), custom toilet paper and napkin dispensers ("free" from Kimberly Clark), fast food ketchup pumps (fitting only the Heinz bag, of course), Coke/Pepsi dispensers, free, that hook up only to Coke/Pepsi concentrates, satellite dishes that work only with one network, hardware that works only with an ongoing software "subscription", and so on, and so on. I remember even Reebok tried selling shoes with inflatable insoles, with CO2 cartridges sold by...Reebok, of course.

Great business models but not so great models from a user perspective, as it makes it very easy to raise pricing due to the high switching costs created by these modern-day Trojan horses. It is one of the main reasons I am the last person without an iPod (or any other similar device): I view them as the leaches and the ticks in the world of personal finance.
Some, like the whole Apple suite, can even be upgraded to herpes, gifts that really keep on giving, especially to the world's most important liver transplant recipient. But I digress again.

What does it have to do with coffee?

A lot, in my view. Today's GMCR is pretty different from the GMCR of 3-4 years ago. GMCR has become a qualitatively different company: they have largely moved from the $8.67/lb coffee business to the $25/lb coffee business. This fact, of course, has not gone unnoticed by Mr. Market, as you can see by the returns I point out above.

They have had an ongoing relationship with Keurig, a marketer of the Keurig single-serve coffee makers and the k-cups for those, for several years, culminating the outright purchase of Keurig in 2006. Keurig is a major player in the single-serve coffee machines, along with Kraft (Tassimo) and Sara Lee (Senseo). I should also include Mars (Flavia): makers of the most hated coffee in offices nationwide.

They all work on the same principles: marketed as better-tasting "gourmet" coffee with little wait (under 1 min), they all use their own disposable cups or pods, and they are marketed as better value than Starbucks. Somehow, they are able to make consumers spend money on buying a machine that virtually locks them in with a provider. That much for rationality.
Then you have the cup cost of about $0.50 each. On top, this one cup makes 8 oz of coffee. For better or worse, who drinks 8 oz of coffee? Realistically, a 16-oz cup is now closer to $1 in cost.

So, GMCR has been working very diligently at expanding the built-in base of proprietary coffee makers thus ensuring a good stream of near-automatic, not-too-price-sensitive reorders from its caffeine-addicted customers. They have also worked hard at expanding the heavy-usage office machine business, as well as the hotel/hospitality business. While most other businesses are reporting declining sales, GMCR machine sales are actually accelerating! So are their coffee sales! See the spreadsheet. I should also note that GMCR tries to cover its cost on the machines, and is not making money on them.

GMCR now owns the whole chain from the roasting and the machines, to the cup shipments. GMCR does not make all k-cup brands that are available. About 40% of the volume comes from k-cup licensees. The notorious, gravity defying widow-maker Diedrich Coffee (Nasdaq: DDRX) is a major k-cup licensee. The profitability of the k-cup business has not been missed by Mr. Market, of course, and DDRX has been topping the charts this year as the top gainer. I used the adjectives to describe the stock and its effect on the shorts, not the company itself. I have no impressions of them either way.

So GMCR has a nice revenue stream from the licensees, who pay Big Brother by the k-cup. I dug around to find out how much they charge (DDRX actually filed the licensing agreement with the SEC) but all the numbers were blacked out. GMCR mentions somewhere in one of the K's that they raised the fee by $0.01 per cup, so I estimate it at $0.02-$0.04, or 5-10% of the retail price. Don't ask me why, just a gut feel of what a "typical" franchise fee would be as a % of topline sales: there is a fine balance between confiscation and yet, making sure the franchisees have something to eat. Ask McDonald's. This estimated fee per cup will come in handy soon.

Then GMCR has the "legacy" GMCR business, the food service and retail coffee. I remember them from 10 years ago when they were taking over numerous gas station coffee businesses around New England and getting increased shelf-space vs. the mass-market coffees in supermarkets in the region. Obviously, this business is not as profitable as the k-cup business, and it accounts for about 30% of the sales of the company.

So, to summarize for my ADD readers, GMCR sells an addictive substance at jacked-up prices (20x the wholesale) to addicts who also give them money to have the proprietary IV drip on their kitchen countertops. In addition, GMCR collects protection money from the little people that make 40% of the k-cup volume in the form of license fees.

Now on to the fun part.

GMCR publishes both machines shipped and k-cup volume by quarter, which is very helpful.
I went back to 2006, and compiled the data, along with growth rates.

A very, very relevant statistic for me here is the average annual k-cup usage per machine.

Since 2006, GMCR has shipped 3.3 million machines and systemwide 3.3 billion k-cups (landfills be damned!). I am ignoring the pre-2006 machines as those numbers are likely very small in comparison to the recent volume (i.e. Q1 '06: 71k machines, Q1 '09: 711k machines). To estimate the usage, we need a relevant metric. This metric is what I would call "machine-quarters", a measure of how many quarters has each machine been in use. It is very similar to the standard unit of labor, man-hours. So, those Q1 '06 machines have been in service roughly 15 quarters, while the 711k machines have been in service for 3 quarters. I say "roughly" because one should really use the midpoint, but whatever.

So, GMCR's machines have clocked in a grand total of 16,739k machine-quarters.

System-wide k-cup shipments for the same period have come in at 3.3 billion units, so, doing some simple arithmetic, we get that the actual k-cup usage per machine-quarter is 196 cups. Makes sense, using my Garage Logic University skills: a quarter has 90 days x 2 cups per day is 180 cup usage. Sounds like a regular home coffee drinker to me (this is their biggest market; institutional is smaller). No red flags here.

Q4 (Q/E 9/30/09) is not out yet, so I am ballparking some of the figures. In Q4 last year, they sold 315k machines, and YoY growth in the three quarters since then has been 120-190%. This means that in Q4 it would be reasonable to expect about 700k in machines shipped. This would put the total installed base at over 4,000,000 machines as of end of Q4.

GMCR has sold nearly 2 million machines LTM, with growth rates each quarter b/n 76% and 190%, with growth rates accelerating YoY. So, for next FY, I think 3 million new machines would be a conservative estimate (deceleration to about 50% from LTM of 100%+). Since those are sold throughout the year, from a usage perspective we have a net add of 1.5 mm machines for the year, if we take the midpoint.

So, for FY2010, the year that just started, I estimate that GMCR will start with an installed base of 4 mm units, growing to 5,500,000 million mid-year and to 7 million by year-end. Again, this builds in a fairly substantial growth slow-down in units.

So now we have to look at the spreadsheet.

GMCR has 4 streams of income: GMCR produced and marketed cups, licensing fees from the licensees, machine sales and GMCR "legacy" business.

The easiest is the machine business: they say they cover costs, or NI from there is $0.

The next easiest is the legacy business. I estimate $250 mm in sales for next year for that line, with a 4% net income margin. GMCR has ~7-8% NI combined margin on the cups, the licensing and the legacy, so legacy should be lower than the combined.

Third is the licensing business. Since it is shrouded in secrecy, I estimate that they charge $0.03 per cup sold, with the licensees being responsible for 40% of the volume for the year. I estimate the margin on the licensing business to be 95%, not dissimilar to IP holder businesses. This simplifies it a bit, as there are taxes involved (i.e. 35%ish) but they also have some related R&D credits, if I remember correctly.

Fourth, and trickiest, is the main NI source, the GMCR owned cup business. I estimate, based on historicals, that GMCR will have 60% of the volume, and the system-wide volume will be based on 5.5 million installed machines, with 160 cups/machine/quarter. I lowered it vs. the historical calculation of 196 because even 160 cups/quarter implies a 100%+ growth rates of the cups business, which is a bit higher than historical norm. It seems that there is a bit of a lag b/n machine sales growth and cup sales growth, so with the machines really taking off LTM, cup revenues should not be far behind, and should move towards 100% from the 60%+ rates. Very similar to the move from 40% to the 60% that has happened since '07. I also estimate the NI margin of the GMCR own cup business to be 10%, again based on the aggregate margin of 7-8%.

So, in all, my back of the (FedEx) envelope calculation implies that GMCR can hit $120mm in NI for 2010, making them valued at 20-22x. Not something I would call "overvalued".

As a caveat, their interest rate expense will be changing because of the Tully's acquisition this last summer, there might be some sloppy merger-related charges, coffee prices at wholesale and packaging can go out of whack, the divergence b/n the units shipped and cup growth rates might mean more than meets the eye, my work is not warranted, etc., etc. caveats. Too many to list. You should do your own analysis: there must be a reason why they are so shorted.

But in aggregate, I think that shorting GMCR going into Q4 results (announcement date is Nov. 11, 2009) has the potential of burning your fingers, especially if we get a massive squeeze like we did a few months ago.

Here's a one-pager in Scribd, select view full-screen, and then zoom.

UPDATE MARCH 29, 2010:   I feel I should do a quick update here as I still get a lot of hits from people searching for information on GMCR (often "GMCR short" and similar). The article above was current as of late October last year. Several material changes have happened since then that you should be aware of. One, GMCR has reported two more quarters of results, and the shares have advanced accordingly, currently trading at around $95 vs. $65 when this article was written. A second material event is the pending acquisition of Diedrich Coffee (DDRX) which, when combined with the smaller acquisitions over the last two years or so, brings the k-cup production back in house. As the company moves from collecting licensing fees to own production, the economics change as, in my view, GMCR will be growing both sales and profits at a quicker pace than before. You can do your pro-forma on the DDRX acquisition: my estimate is that GMCR probably has another 20-25% upside from here, provided that it keeps the high forward P/E multiple. Entry now at $95 does not offer a compelling risk/reward in my view, and, hence, I have no position.

(PLUG: the author of Barbarian Capital blog is available for the right consumer- or inflation-focused analyst opportunity within the US) 


Ian said...

Great post. I suppose the shorts are betting that in a recession customers will realize the shitty economics from their perspective. But still cheaper than starbucks...and Americans are lazy.

Ankit Gupta said...

Excellent analysis - very thorough and well written.

Did you look into patent concerns surrounding the K-cups? They expire in 2012 and because these are where they actually have a profit margin, the loss of patent protection could shrink the income they do have.

Barbarian Capital said...

I have seen this as an argument. I do not think it is as big of a threat as it is in the pharma or other industries. One has to look beyond the patent, and into the whole coffee brand portfolio. What are the major brands that are not a part of the k-cup group, or that are not a part of the competing systems groups? Once you answer that, I think you will see that a new k-cup entrant will have hard time gaining meaningful share.