|Shares Out. (in M):||352||P/E||-29||24|
|Market Cap (in $M):||8,526||P/FCF||-19||-30|
|Net Debt (in $M):||-2,139||EBIT||-202||-292|
|Borrow Cost:||Available 0-15% cost|
“It’s expensive, long-lasting hardware, stupid!” Peloton is the next Fitbit (whose stock price is still -93% off its peak five years later) or GoPro (-96% five years later), plus an OK but small subscription business, minus much longer product-replacement cycles, current operating losses, and a lot more recession risk. Those hardware sales are going to crash some day. We have already seen a lot of skeptical investment commentary on Peloton, but we haven’t seen anyone hammer home this simple and powerful fact as hard as it should be hammered, even by those who have explicitly compared Peloton to Fitbit and GoPro. (We also haven’t seen enough attention paid to the new competition that just launched or is about to launch from SoulCycle and NordicTrack.)
As the bloom has come off the rose of money-losing high-growth technology unicorns, plenty of people have been quick to note that Peloton superficially fits the profile of its more famous recent-IPO peers like Uber and Lyft. Like the others, the business currently loses money, and its stock is trading below the IPO price. Our initial hypothesis was that sentiment was likely too negative out of the gate to merit shorting it now. But Peloton doesn’t fit the general profile; it’s worse. Most of the others probably have enough demand headroom that they could keep growing revenues rapidly for many years, as long as they keep selling dollars for 75 cents. The issue for them is whether, how soon, and to what extent they can reach profitability. Peloton sells luxury durable goods and is currently filling the unmet demand for a product and service category that previously didn’t exist. As with Fitbit and GoPro, once that initial demand is filled and those who want Peloton already have Peloton, Peloton’s revenue growth will fall from positive triple digits to negative. The stock price will fall materially then if it is anywhere near its current level. We can't predict the decline’s timing, but given the valuation, the likely small size of Peloton’s effective TAM, and the current market environment, we have taken an initial short position now to wait it out.
THE RELEVANT PELOTON BUSINESS BASICS
Almost every professional stock analyst already knows the Peloton basics, so this overview will stock to what matters most for the stock. Peloton gets ~80% of its revenue from selling $2,245 exercise bikes and $4,295 treadmills. Peloton launched its bikes in 2014 and its treadmills in 2018. Of course, management has refused to disclose the number of units sold in any given period, let alone a split between bikes and treadmills. The likely reason for not splitting out treadmills is obvious: There is no second equipment category to which spinning classes translates well, not even treadmills and certainly not anything else Peloton might try to expand to in the future. Spinning classes have been popular for decades, but do you even know anyone who has tried a treadmill class? We assume that bikes are the vast majority of recent sales and almost all of the installed base.
The other ~20% of revenues come from subscriptions to online-streamed classes, overwhelmingly from $39/month “connected fitness” subscriptions that use the bikes and treadmills. (1% of revenues comes from “other,” which is mostly clothing sales.) Peloton offers regular live workouts with a real-time instructor and a library of on-demand recorded workouts; it produces “over 950” new programs per month. Peloton also offers $19.49/month “digital” subscriptions to its program library. As of June 30, Peloton had 5x as many “connected fitness” subscribers as “digital” subscribers, which means it had 10x as much connected subscription revenue as digital subscription revenue. Connected-fitness subs grew 108% in FY19, while digital subs grew 122%.
In both FY 2019 (June end) and in the June quarter, equipment revenues grew about 105% year over year, subscription revenues 125%, and total revenues 110%.
Gross margins are running 43% for both hardware sales and subscriptions. The subscription business is very un-SaaS-like, with substantial variable costs that will prevent high margin expansion as revenues grow. Peloton streams music during every workout and must pay public-performance royalties for the songs. (It also faces a $150 million lawsuit by music publishers for Peloton’s past streaming that was allegedly unlicensed.) To create all those workouts, it must also pay instructors and provide studio space for each session.
It is also worth mentioning that Peloton has a hefty 66 million shares worth of options and warrants outstanding, equal to 23% of basic shares, with a blended strike price of only $7. They are hidden from Peloton’s GAAP numbers because of the ongoing net losses, but if and when Peloton’s net income turns positive, they will inflate the diluted share count and market cap by hefty amounts. The opening valuation table includes the if-exercised option shares, exercise cash proceeds, and IPO cash proceeds.
EXPENSIVE, LONG-LASTING HARDWARE
Peloton’s hardware is expensive, relative to alternatives and in absolute terms. A $2,250 Peloton bike costs 6x a good non-connected exercise bike, and a $4,295 Peloton treadmill costs 3-4x a good non-connected treadmill. Amazon segregates its exercise bicycles into “below $200,” “$200 to $300,” and “above $300.” The “Amazon’s choice” product in the above-$300 category costs $399, and its best-selling bike costs $155. It splits treadmills into under-$500, $500-1200, and above-$1200. The Amazon’s-choice product for above-$1200 costs $1499, and Amazon’s best-selling treadmill costs $899. Apart from the high cost relative to other bikes, $2,245 or $4,295 is a big-ticket purchase that few U.S. households pay for much of anything other than their homes and autos.
Everyone buying Peloton’s equipment is also signing up for the $39/month subscription. That, too, is expensive for what you get, relative to some alternatives. It is the same price as the average full-service gym membership, which gives you access to many different types of cardio and weight machines, other equipment, and often a pool. It is 3x the cost of a Planet Fitness membership and 2x the cost of Planet Fitness’s Black Card membership with gym reciprocity and other extras. It is infinitely more expensive than biking outdoors for free. It is, however, cheaper than in-person classes. $39 is enough to get you two or three in-person fitness studio classes per month in most cities. (Maybe only one SoulCycle class in Manhattan, but come on.)
For the high-income exercisers who have chosen Peloton, the high cost is worth it to get a fitness-studio class, in their home with no travel time, at any time of day. Indeed, Peloton customers are unusually enthusiastic and, so far, loyal. The subscription service is averaging 0.65% monthly churn, so far, and the average customer is doing three workouts per week. They love it. Just like people still love their Fitbits, GoPros, and Teslas.
Peloton’s larger long-term business problem is that gym equipment lasts a long time. Even fitness chains, whose bikes and treadmills get used far more often per day than do in-home machines, only replace their equipment every 5-7 years. A high-quality bike or treadmill like Peloton’s usually lasts around 10 years at home before breaking down. It won’t become technologically obsolete, either. Thus, after Peloton makes its initial sale to its customers, its replacement-equipment revenues will be small.
THE HISTORY OF FITBIT, GOPRO, AND TESLA
To see how Peloton’s equipment revenue growth might progress, it is instructive to look at how Fitbit’s and GoPro’s revenue growth progressed:
In both cases, when the growth rates were in the triple digits, analysts were forecasting that the rates would taper off relatively slowly and stay positive. Instead they both crashed to -24% and -27% in relatively short order. The absolute sales levels stayed relatively flat after the initial big drop, because demand transitioned from first-time purchases to more-sustainable replacement purchases. To repeat, these 24-27% revenue drops caused 93-96% stock price drops.
Peloton’s eventual revenue drop should be steeper than Fitbit’s and GoPro’s were, because its equipment replacement cycle is closer to 10 years while Fitbit’s and GoPro’s are closer to 3 years. To see what can happen to sales for a durable, expensive category-creator that fills its initial demand, one can look at Tesla’s deliveries for the Model 3. Tesla has been able to report revenue growth lately only because it launched the Model 3 in 2018, which draws from a new source of unmet demand (for $45,000 EVs instead of $90,000 EVs), then launched the Model 3 serially in new countries quarter by quarter to tap yet more new pools of demand (first North America, then continental Europe and China, then the U.K. and other right-drive countries), and also launched cheaper and cheaper Model 3 trim levels with lower and lower sales prices. Tesla does not disclose deliveries split by model and region, but good outsider estimates exist based on reliable external data sources. The next chart shows North American quarterly delivery estimates with quarter-over-quarter growth rates. The numbers are skewed for analogy purposes by the cheaper trim levels in later periods and by the 50% drop in U.S. EV subsidies for Teslas at year-end 2018, which boosted 3Q and 4Q deliveries and dampened 1Q deliveries. Adjusting for the subsidy effect, North American deliveries peaked only three or four quarters after introduction and have declined steadily since then, despite the launch of cheaper trim levels.
When recast on a full-year basis, Tesla’s Model 3 sales are going to show infinite percentage growth in 2018 after zero in 2017, low-teens growth in 2019, and declines already in 2020 despite the introduction of cheaper trim levels.
PELOTON’S LIMITED DEMAND HEADROOM
Obviously Peloton’s pop and drop won’t be anywhere near as brief as Tesla’s; Peloton launched its bikes five years ago and is still growing sales by triple digits. We don’t think it’s possible to estimate when the peak will come, but one can get a sense by examining the several hurdles that severely limit Peloton’s effective TAM. Peloton must sell to:
with an exercise enthusiast
who highly values the in-home experience instead of going to a gym or studio or, you know, bicycling or running outdoors
who doesn’t value the social connections in in-person classes
who likes cardio equipment and, among the cardio equipment options, likes exercise bikes or treadmills
who doesn’t mind the monotony of a single exercise mode (or will pay multiples of Peloton’s cost per month for multiple exercise modes)
and who doesn’t choose a competitor’s offering.
High income households -- Peloton’s own materials suggest incomes above $50,000 are high enough to make Peloton attractive, but $100,000 seems like a more realistic cutoff for market sizing. Peloton says that one of its fastest-growing segments is households with incomes under $75,000, but of course they don’t say what size that segment is. It could be that under-$75k subscribers recently grew from 1% to 3% of Peloton’s base. The U.S. has 37 million $100k+ households and 65 million $50k+ households.
Exercise enthusiast -- The CDC says that only 21% of American adults exercise the “recommended” amounts. In 2018 roughly 25% of adult Americans had a fitness club membership, 62 million out of ~250 million. Of that 62 million, 14 million are Planet Fitness members at $10 or $20 per month (most of whom are light users), perhaps another 7 million at other low-cost gyms. Peloton’s own survey of 1,000 people suggested that 12 million = 18% of the $50k+ group would “express an interest” in Peloton products.
Values in-home / likes exercise bikes or treadmills / accepts monotony -- Of the 25% of Americans with fitness club memberships, 40% report paying for a boutique studio (so 10% of adults). I don’t know how to provide other numbers on these other preference factors, but they clearly slash the true TAM down by large amounts. I can use myself as a useful anecdotal data point, even though rule #1 in marketing and TAM-related investing decisions is “you are not the market”: I am high-income. I exercise 2-3 times per week. Bicycling (outdoors on real bikes) is one of my primary exercise activities. I highly value time savings from in-home activities and occasionally use YouTube videos for some class-like workouts rather than going to classes. I use a desk treadmill every work day. Those facts suggest I should be dead-center in Peloton’s TAM -- but I would never consider buying a Peloton. I dislike exercising on stationary bikes or treadmills. I own a bicycle resistance trainer for indoor use in winter, and I haven’t touched it in years. I almost never use cardio equipment and, if that’s the only option, I want an elliptical trainer. I also want more variety in my exercise and couldn’t stomach paying $2250 + $39/month for one video workout per week while I do other activities for the other one or two workouts per week.
Moving from single anecdote to decades worth of many anecdotes, here are thoughts from a friend in the fitness industry who has tried, and has watched thousands of other people try, every exercise type there is: “People either LOVE IT [spinning class] or don't do it. There are some ‘tweeners’ but very few. I fwiw have done spin maybe a couple dozen times in the past 25 years and never once liked it. I have friends otherwise psychographically identical to me who do it 3 times a week year in year out. It's honestly a cult. And it has persisted - it was popular 30 years ago, 20, 10. So if there is a ‘niche in boutique fitness that HAPPENS to require an expensive machine to boot,’ it is spin. It's a huge winner in that regard. Soul Cycle and Peleton hit the jackpot. Alas, that means when you turn to ‘what next’ for Peloton - there is no ‘next’ anywhere near it.”
Competition -- So far Peloton has largely had its niche to itself. That is now changing quickly. Although the new competitors might not make a dent, they probably will, because (1) some of them are exercise-industry heavyweights with large operational resources at their disposal -- no one is better equipped to offer “SoulCycle at home” than SoulCycle -- and (2) most of them are offering lower price points.
Equinox Fitness, owner of the SoulCycle, Equinox, PURE Yoga, and Blink Fitness chains, and the Precision Run concept in its Equinox clubs (SoulCycle for treadmills), has announced it will launch Peloton-like equipment and streaming services this winter. It will offer content from its existing studios, including SoulCycle and Precision Run.
NordicTrack, the world’s #1 seller of treadmills and a large seller of exercise bikes, recently began offering a Peloton clone at a cheaper price. NordicTrack’s “Commercial Studio Cycles” come with a one-year subscription to its iFit service, “valued at $396” (which implies iFit costs $33/month), as well as two hand weights. They are selling for $1,599 for a 14-inch screen and $1,999 for a 22-inch screen. Ignoring the hand weights but subtracting the one-year membership, the effective bike prices are $1,203 and $1,603. NordicTrack has to be spending real money on the digital marketing for its service; once I started researching Peloton online, NordicTrack started feeding me multiple video ads every day.
Flywheel Sports is mostly a SoulCycle clone (with 42 in-person studios vs. SoulCycle’s 98) that is now offering a Peloton clone, again with content created in their legacy in-person classes. The subscription is $39/month. Their bike + screen is selling on Amazon for $1,599 including in-home setup and 2 months of subscription, so $1,521 ex-subscription. Or you can buy the same bike without the screen and bring your own device (tablet or phone) for $1,399 including set-up and 2 months subscription.
Echolon Fit Connect offers streaming classes attached to a wide range of equipment, all at lower prices than Peloton’s and some at much lower prices. The lowest-priced equipment is bring your own device.
Zwift is a start-up with a different type of offering that nonetheless competes directly in the in-home-biking/treadmill space. Users bring their own in-home equipment but access Zwift’s streaming platform that is essentially a massively-multiplayer gaming environment, with “virtual worlds” to exercise in together and “1,000+ structured workouts made by top coaches.” Subscriptions start at $15/month.
YouTube, etc. Peloton’s $19/month equipment-free subscription competes with tens or hundreds of thousands of free workout videos available on YouTube. Barriers to entry for equipment-free workouts are lower than for the equipment-connected workouts. If this piece of the business takes off, competition will proliferate.
Instead of thinking about a theoretical TAM number in a vacuum, here is my SWAG at how large Peloton might actually grow its installed base of machines:
U.S. $100k+ households = 37m
25% with exercise enthusiasts = 9.3m
25% want in-home bike or treadmill workouts as their primary exercise = 2.3m
Peloton takes half the market share in its niche, competitors take the other half = 1.2m
Add some international growth. Currently 2% of Peloton’s equipment sales have come outside the U.S (Canada, UK, and Germany). They will try for much more, but this type of offering seems a middling candidate for international expansion -- not terrible, but not remotely a winner-take-all global offering. Classes are language-specific and to some extent culture-specific. If Peloton can get international sales up to 50% of U.S. sales, that would bring the total to 1.8m.
Any of these numbers could be wildly wrong, but skewed in different directions. Perhaps only 5% or 10% of high-income exercisers want in-home bikes or treadmills as their primary workout. Perhaps high-income households have a much higher skew of exercise enthusiasts. Perhaps Peloton can take 75% of its niche, or only 25%. The business might scale internationally to only 25% of U.S. units or to 100% of U.S. units.
CONSENSUS EXPECTATIONS LOOK HIGHLY UNREALISTIC
Although we can’t know how many units Peloton can sell before filling its unmet demand pipeline, we can know that consensus revenue estimates look highly unrealistic given the ranges of numbers above. We estimate that reaching consensus estimates for 2023 revenues of $4 billion requires Peloton to sell 3.5 million units. Peloton’s refusal to disclose unit sales or ASPs for equipment make it impossible to build the kind of bottom-up model we normally would build, but we can get close enough using the following approach:
Cumulative units sold are known -- 557k
FY17, FY18, and FY19 product sales revenues are known. We can assume FY14, FY15, and FY16 sales of $17m, $38m, and $83m (120% growth in each early year) to arrive at cumulative product revenues of $1390m through FY19. That implies a cumulative ASP of $2,400 for the 557k units, which fits well with the known price points.
Assume product sales growth of 70% in FY20, 40% in FY21, 30% in FY22, and 20% in FY21. The growth implies that each year Peloton sells about $500m more equipment than it sold the year before.
These numbers also imply that the installed base of equipment, measured by product revenue it generates rather than units, grows by 88%, 66%, 52%, and 41%
Assume that the ASP grows by $100 per year as the mix shifts modestly towards treadmills. Then the installed base of units sold grows by 85%, 62%, 48%, and 37%. This is the base for subscription revenue growth each year.
Assume total subscription revenue grows by 15% more than the equipment units installed base in FY20 and 10% more each other year, thanks to (1) slightly higher growth from “digital” (non-equipment-linked) subscriptions, (2) possibly some ARPU growth, although I wouldn’t count on it, and (3) the mathematics of revenue-versus-subscriber-count growth when growth rates are high. Thus subscription revenues grow by (85% + 15%) = 100% in FY20, and 72%, 58%, and 47% afterwards. Assume that “other” revenue (1% of total) grows in line with product sales.
Our subscription revenue estimates implicitly assume that churn does not increase much, which is almost certaintly wrong, for three reasons. First, the exercise world’s most enduring truths are that people don’t stick with their exercise programs, grow tired of an existing exercise mode and switch, and stop using their in-home equipment. Peleton’s equipment base is so new that people haven’t had time yet to lose interest or grow tired of it. But the signs are already there: In FY19, Peloton subscribers who joined in FY16, FY17, FY18, and FY19 did 8.4, 9.3, 10.4, and 13.7 workouts per month. The longer they have been members, the fewer workouts they do. Second, Peloton is signing up the hardest-core users first; new customers in later years are more likely to churn off. Third, very high subscriber growth depresses the reported churn numbers, and Peloton's normalized churn is already likely 2x the reported numbers. To simplify a bit, let's say it takes a year before new subscribers start churning off. When Peloton more than doubles its subscriber count, as it has every year, then half its sub base has 0% churn and the other half has double the reported rate. Thus with Peloton currently reporting 0.65% monthly churn = 8% annual churn, the normalized rate is closer to 16%. The reported rate will rise as the subscriber growth rate falls.
The key resulting outputs are:
Total revenues for FY20-23 of $1.6, $2.4, $3.3, and $4.2 billion. (In line with current consensus estimates each year. Slightly different from consensus because I used round numbers in 10% increments for my product sales growth inputs.)
Cumulative products sold of 1.1, 1,7, 2,5, and 3.5 million.
Then from June 2023 forward, to avoid having equipment revenues decline, Peloton must keep selling almost 1 million machines each year, forever.
If Peloton does get to a 3.5 million installed base by 2023, then its equipment revenue decline from there seems more likely to be >50% rather than the 24-27% experienced by Fitbit and GoPro. Alternatively, and more likely, Peloton isn’t going to get close to 3.5 million unit sales by 2023.
Especially because a recession will probably occur between now and 2023. Big-ticket luxury good sales don’t do well in recessions.
THE SUBSCRIPTION BUSINESS ISN’T ENOUGH TO SUPPORT THE EQUITY VALUE
As the introduction’s second sentence said, Peloton is the next Fitbit and GoPro, plus an OK but small subscription business, minus current operating losses (and other things). The plus from the subscription business isn’t nearly enough to justify Peloton’s equity valuation. If Peloton manages to hit the unrealistic numbers laid out above, FY23 subscription revenues would still only represent 35% of total revenues. Meanwhile Peloton is currently running -22% EBIT margins and will be burning hundreds of millions through 2023. Margins should improve over time, but they aren’t going to explode higher. Equipment gross margins are more likely to go down than up, given their high starting point, the increasing competition, and Peloton’s desire to grow the subscription business as quickly as possible. Subscription gross margins can probably increase, but the high variable costs will limit how much. SG&A as a % of sales should fall steadily after the initial post-IPO land grab of FY20. Even so, some of the profitability gains from SG&A leverage will eventually be eroded by lost gross profit from falling equipment sales.
With businesses whose futures are so uncertain, we find it more helpful to see what assumptions one must make to justify the current equity valuation, rather than anchor on any other assumptions and say “it’s worth $X/Y/Z in a bear/bull/base case.” We can see here that Peloton is worth something far less than its current price. Here is what it takes to get to a $23 or $24 per share stock value:
Revenue assumptions through FY23 as above -- $4.2B revenues in 2023
Product revenues stay flat every year after FY24 -- i.e., they actually sell the impossible 950k units each year thereafter
Subscription revenues beyond FY23 grow in line with the equipment installed base -- which works out to 27%/21%/18%/15%/13%/12%
EBIT margins grow steadily to 13% by 2029. (In our model, we have blended gross margins rising 300bps and opex/sales falling 3200bps.)
Almost no taxes until 2028, then 22% tax rate
Capex = D&A, NWC/sales stays constant
Short term: GoPro had a nice >100% run after its IPO before crashing because revenues originally beat expectations. Fitbit had a very brief 50% run before quickly crashing. If Peloton’s management has not tamped sell-side estimates down to something they can beat handily over the next several quarters, then they have failed in their duties as leaders of a newly-public growth-tech business. The longer Peloton “beats,” the more likely that the stock goes up before going down. Even six months ago, we would have waited to short this stock until after a quarter or two of results, because the “beat” could push the stock much higher. Upside risk exists from the other direction as well: One could argue that sentiment on this stock is already negative enough that it wouldn’t take much good news or non-bad news to send it higher in the short term. However, we are making the bet that investors’ observation of WeWork, Uber, Lyft, Slack, SmileDirect, RealReal, Lovesac, etc., has made revenue beats by companies with questionable TAMs or business models less dangerous to the upside. (Wayfair keeps beating on revenues and recently had a 40% stock drop.) We have sized our position smaller than our usual short for now, leaving room to add if it does rise on us.
One relevant and surprising tidbit relevant to the short-term risk is that the borrow cost is already down to 4%, less than I would have expected. Short interest is 12.7m shares compared to an IPO of 40m and a greenshoe of 6m shares.
Longer term: When we think about “what could go right for Peloton,” it almost has to be the development of another revenue stream to supplement the core spinning offering. We don’t think it likely that either a second equipment type for the same offering or the digital (equipment-free) subscription offering are likely to be that revenue stream, in the first case because no other equipment is suitable and in the second because of widespread competition. But it is possible that one of them works out. It is also possible that Peloton develops something we can’t currently envision.
In likely chronological order: