Note: Please see my updated thoughts on Etsy here
Matt Joass, Chief Investment Officer at Maven Funds and one of the subscribers, texted me in early September, “Have you ever looked at Etsy? Pretty interesting model. Growing like a weed at the moment.” (sharing with permission, and as of this writing, Matt has no position on Etsy)
Anything that is growing like a weed these days trades at multiples so high that only the bravest (and often the most cavalier) investors can dabble into these stocks. But as I am primarily interested in understanding different businesses (and buying good ones whenever the fundamentals appear better than expectations embedded in the stock price), the business seemed interesting enough for me to do a deep dive.
One initial surprise was Etsy currently trades at ~10x 2020E revenue which is, of course, “cheap” for a growth stock when we see so many others trading at 20-30x (or even higher) multiples. Perhaps one of the reasons it is not trading at egregious multiples is it is not a SaaS/cloud/EV company. After spending decent time on Etsy over the last month, I do think it is indeed cheap (this time without the quotation marks).
I know my readers are mostly not here for a stock pitch. However, this will often (hopefully) be a serendipitous by-product of my deep dives, and regardless of my bullish/bearish tones in these deep dives, my basic approach to understanding business will remain same.
First, I will delve into the Total Addressable Market (TAM) of Etsy. After discussing the size of the market, I will take a closer look on the buyers and sellers at Etsy as well as the competitive dynamics of the industry. Finally, I will explain how Etsy makes money and why I am comfortable with Etsy’s valuation.
Sizing Etsy’s TAM
Founded in 2005 in a Brooklyn apartment, Etsy is a two-sided online marketplace connecting buyers and sellers of primarily handmade goods, vintage items (>20 years old), and craft supplies. In August 2019, Etsy also acquired Reverb, an online marketplace of new, used, and vintage musical instruments. Given that Reverb is only ~5% of total Gross Merchandise Sales (GMS), our discussion primarily focuses on Etsy’s primary marketplace.
Even though Etsy attracts buyers and sellers from all over the world, it identifies six countries as its core geographies: United States, United Kingdom, Canada, Australia, Germany, and France. 36% of Etsy’s 2019 GMS of ~$5 Bn was considered international. An order is considered international if either the buyer or the seller is located outside the US.
The top six categories represent ~75% of the GMS of $6.9 Bn (LTM) at Etsy: Homewares and Home Furnishings ($2 Bn) , Jewelry & Accessories ($1.2 Bn), Craft Supplies ($826 Mn), Apparel ($774 Mn), Paper & Party Supplies ($316 Mn), Beauty & Personal Care ($431 Mn, excluding mask sales). If you haven’t gone to Etsy’s website yet, I would suggest you to explore the site to be more familiar with the types of products they sell.
So how big is Etsy’s market? Let’s start with management’s estimates. In 2019 “Investor Day” presentation, Etsy mentioned how they came up with their Total Addressable Market (TAM). First, they started with total retail (both physical and online) sales of $1.7 Tn for all relevant categories in their six core countries. ~15% or $249 Bn of these retail sales happened online in 2018. Since Etsy’s whole mantra is to get away from commoditization, management does not think all of this can be addressable. Etsy customers do not go to the site to find the cheapest items available. They are more interested in making a statement (fashion/supporting small businesses) through the kind of products they are buying or using.
Etsy wants to focus on this “special” segment, a segment of customers who want to express themselves differently from the crowd. Even though the desire to be different from the crowd has an innate appeal to the vast majority of people, not everyone has the desire or capacity to spend for it. Etsy indicated ~40% of US consumers are expressive/seekers. Therefore, ~40% of the online sales of the relevant categories or ~$100 Bn is Etsy’s TAM. Considering Etsy’s GMS in 2019, it implied ~5% market share.
Etsy says they came up with this 40% number from their customer surveys. It is difficult to evaluate the validity of such claim. Nonetheless, I think even if management is wrong in terms of what percentage of the consumers are expressives, or not overly price sensitive, shareholders need not lose their sleep over this.
For context, here are some “select category” company TAMs mentioned in their investor presentations/10-k/sell side reports:
Wayfair (online and off-line home furnishings): $430 Bn; Asos (online apparel in specific countries): $288 Bn; Farfetch (online personal luxury goods): $307 Bn; Blue Line (Global gems and jewelry): $443 Bn
From the very beginning, sizing Etsy’s TAM has been one of the biggest talking points for investors as many feared the actual TAM is lot lower than management thinks. Sizing the TAM is more art than science and perhaps there is a legitimate question whether it’s even a useful exercise given the difficulty of estimating TAMs.
As an investor, I am less married to the exact size of the TAM and more interested in having an informed opinion on management’s estimates and reasonableness of those assumptions. I believe if Etsy finds it difficult to grow in future, it won’t be because their TAM is smaller than expected. If Etsy experiences lackluster growth in future, the most likely rationale will be poor execution. The market is indeed big enough for them to continue to grow at double digit for this entire decade. Let me explain my reservoir of comfort when it comes to Etsy’s TAM.
E-commerce has been enjoying secular tailwind even before Covid-19 and the pandemic has just accelerated the trend even faster. Etsy estimated ~22% of total retail sales will be online in 2023, and the TAM for “special” or Etsy segments will be $170 Bn in 2023. Of course, Covid-19 has exceeded even the most bullish penetration expectation for e-commerce in 2020 and it reached 27% in April 2020. Even though we expect the penetration to come down a little after physical stores open, I think we all mostly agree at this point that e-commerce penetration will be 22% or higher in 2023. As an online marketplace, Etsy will be one of the biggest beneficiaries of this secular tailwind.
Etsy management estimated the TAM based on relevant categories. But this pandemic really tested the assumptions of “relevance”. Etsy sold $346 Mn of masks just in 2Q’20. With the benefit of hindsight, this may not be surprising anymore, but there are other categories that sold extremely well during this pandemic that surprised even Josh Silverman, CEO of Etsy.
“And an example I would give is bread and bread-making in the month of May. I never would have guessed that bread and bread-making products would be a big category on Etsy. Never in a million years would it have occurred to me. But in the month of May, we started to see bread shortages in stores, and a lot of people stuck at home decided to take up bread-making as a hobby. And so they turned to Etsy, and we started to see a surge in keywords, in demand, and literally within hours, bakers are coming on Etsy and starting to meet that demand. And supply and demand met effortlessly and almost instantaneously because the communities are paying enough attention and bakers started to talk to bakers and word just got out. So we actually didn't need to do anything in that case, and we had enough supply.” (KeyBanc Capital Market’s Technology Series, September 15, 2020; emphasis mine)
That quote perfectly encapsulates the power of marketplace business model. Etsy, as the owner of the marketplace, does not have to predict all granular details of consumer behavior. All they have to do is come up with some high-level rules and regulations in terms of what you can sell and what you cannot, and then empower the good old capitalism at work. Millions of sellers are experimenting every minute what sells and what doesn’t, and they are adapting every day to make sure they make products for which there are buyers out there. At times, it can be face masks, or even bread making products.
Trying to size the TAM of such a marketplace can, therefore, be notoriously difficult task although it does leave an impression that size is perhaps not a constraint. So, what are possible constraints for such a marketplace?
The biggest constraint for such a marketplace business is the lack of awareness of the consumer. To elaborate more, I will borrow from Nick Sleep, one of the value investors who truly understood Amazon. He explained in 2011 with such eloquence why internet retailing could grow for years that I would request you to read the following few paragraphs twice:
“Unlike traditional retailing, where growth is bounded by the physical constraints of a building’s size, location, and opening hours, internet retailing has no such constraints. We could all, for the sake of argument, decide to do all our shopping online next week, and Amazon et al would take the orders. Odd, then, that the growth rate in revenues at the main internet-based retailers is, very broadly, comparable to that enjoyed by a physically constrained retailer, say, Wal-Mart, at a similar stage in its development. So what is going on?
…our hunch is that the growth rate in online retailing is held back by consumers’ psychological biases. We are all creatures of habit, and most of us give up comfort blankets quite reluctantly. It therefore takes time for a new regime to be adopted and, for instance, to buy books online instead of buying them at the local store.
In high street retailing we can estimate, broadly, what revenue growth for each firm will be next year: it is a product of same store sales growth plus any new stores. We might not be precisely correct about either number, but the range is bounded and so we can make a reasonable, “generally-right” (as opposed to “precisely wrong”) estimate. When investors think about the future of a business, they often have in mind the assumption that growth rates slow with time, as competition ekes away advantages and marketplaces become saturated. Predicted revenue growth rates (used in valuation models) therefore start high and end low. This is especially true for firms that are quite large already.
However, if the rate of growth in internet retailing is a product of attitude, rather than assets, then, the fact that a firm is quite large already does not necessarily tell you that its growth rate is set to slow. The widely held presumption that regression to the mean begins the moment the analyst picks up their pen, risks being wrong footed as a result. Two years of forty percent revenue growth, for example, will result in revenues doubling in twenty-four months and regression to the mean-based estimates would be out by almost a factor of two! That did not take long. In other words, although some online retailing firms may be quite large, they may also be quite young. In our opinion, it is this realization that has partially driven the revaluation of internet retailers these last few years.”
Boy, was he right!! After almost 10 years since these paragraphs were written, Amazon’s retail business is still posting 20%+ topline growth with no sign of slowdown in sight.
In light of this rationale, now imagine the impact of Covid-19. Mother nature has forced us to try e-commerce in the last six months, and many of us finally realized there are so many things we could buy online. Once things are back to “normal”, we will most certainly go out far more often than we did in the last six months. But would we go back to visiting physical stores as frequently as we did before?
If new Etsy customers liked what they bought during the last few months, it seems extremely unlikely that they would forget Etsy anytime soon. What Covid-19 did for Etsy, it would otherwise perhaps take years as well as hundreds of millions of marketing dollars to recreate a similar impact. For example, Etsy grew active buyers and sellers on its platform by ~13 million and ~600,000 respectively in the last two years. To acquire these buyers and sellers, it spent ~$375 Mn in Sales & Marketing. On the other hand, just in the first 6 months this year, active buyers and sellers on Etsy platform increased by ~14 million and ~560,000 respectively, and it spent only $163 Mn in Sales & Marketing during this period. CAC has basically more than halved during Covid-19.
Etsy’s GMS is likely to grow to ~$8.5 Bn in 2020, ~75% higher than it was in 2019. Non-mask sales grew 93% YoY in 2Q’20. These numbers substantiate the larger picture to me: one of the biggest risks that investors have worried about Etsy for years has been significantly de-risked. These numbers should assuage any concerns out there about TAM being smaller than expected.
Moreover, Etsy enjoys very strong network effects which I will discuss more later. Please note management focused on core six geographies while estimating TAM, but there is hardly anything that’s stopping a seller in Finland or New Zealand to start their shop at Etsy. Given the international nature and appeal of handmade and craft items, the range of products remain quite diverse and if Etsy expands to other markets aggressively, it primarily needs to focus on the consumer awareness in those markets given the global seller base and already high “liquidity” on the platform.
Therefore, a TAM of $170 Bn in 2023 sounds very much reasonable and this implies only ~7% market share based on my 2023 GMS estimates which I will discuss later. It is, however, perhaps highly likely that management will increase its TAM estimates following its experiences during Covid-19 in the next Investor Day presentation.
Who shops at Etsy?
Melissa McCarthy. No, seriously.
Validation from such celebrities presumably at no cost can be a powerful thing. The most powerful and the least expensive way to grow and enhance your brand is word of mouth. This is even more evident from Etsy’s social media presence. Even though Etsy’s presence on Facebook is far from spectacular, its presence on IG trumps some of the other well-known e-commerce players.
I think it is at least partly, if not mostly, explained by the fact that Etsy is very “instagrammable”. Most people don’t buy something from eBay and flaunt it on IG. But Etsy shoppers are much more likely to do exactly that which creates word of mouth and brand awareness. 85% of Etsy site visitors come organically and only 15% come via paid traffic channel. This level of word of mouth helps to lower CAC.
Looking at customer cohorts (see below), we know Etsy loses ~55% of its customers from prior year. This is of paramount importance because customer expenditure on Etsy platform increases by 50-60% by year 2.
Repeat buyers, defined as buyers who ordered more than once in last twelve months, as % of active buyers have consistently hovered around 40%. Habitual buyers, defined as buyers who order more than 6 items or spent >$200 in last twelve months, are 5.5% of all buyers. Only ~5% of habitual Etsy customers are below 25 years old, and 41% of the habitual customers have more than $100k annual income confirming their relatively price insensitive status.
Therefore, there seems to be a massive potential opportunity if Etsy can somehow retain the existing buyers on its platform. After years of ~40% of repeat customers, it has increased by ~500 bps during 2Q’20. It is difficult to ascertain to what extent this is being benefitted just by the pandemic. We don’t know whether repeat buyers as % of active buyers will go back to ~40% level in a year or two.
Etsy did try to look into the pain points of one-time buyers. It decided to prioritize free shipping (orders above $35) in its search results since July last year. Shipping cost is certainly a friction in e-commerce and Etsy mentioned buyers are 20% more likely to buy items with free shipping. This may also increase their penetration among buyers with below $100k income. GMS per active buyer did increase by 3.8% in 2019, highest in last 8 years.
Overall, there seems to be a massive potential opportunity just by improving retention of current customer base, repeat and habitual buyers, and therefore increasing GMS per active buyers. There were 17.5 mn and 19 mn new buyers in 2018 and 2019 respectively. So customer base does not seem to be too close to hit a ceiling, but improving those retention metrics can truly do wonder for Etsy. If 5% bump that we saw during Covid-19 becomes permanent (again, difficult to know), there is a lot of growth left in this business.
Etsy sellers: a tailwind in the passion economy
Li Jin has been writing on the secular theme of “Passion Economy” for quite some time (feel free to explore her substack to dig more into Passion Economy). She eloquently argues while the gig economy came to the forefront during the last decade, the decade of ‘20s will be dominated by passion economy as all sorts of creators have exploded, fueled by platforms empowering these very creators. Jin writes in one of her posts, “New digital platforms enable people to earn a livelihood in a way that highlights their individuality. These platforms give providers greater ability to build customer relationships, increased support in growing their businesses, and better tools for differentiating themselves from the competition. In the process, they’re fueling a new model of internet-powered entrepreneurship.”
But how big this passion economy can really be? A Dartmouth study on latent entrepreneurship found that 71% Americans (110 Mn) would prefer to be self-employed. Therefore, supply of these potential entrepreneurs/” side-hustles” should be a secular tailwind. Another 2017 study on “Creative economy” showed there were already ~17 mn people in the US in 2017 who earned some income from selling their content/products in different platforms.
~85% of Etsy sellers are women, and 90% of the Etsy shops are just run by one person. As per my call with their IR team, Etsy marketplace tends to follow the pareto principle: ~20% of sellers sell ~80% of GMS, and ~30% of Etsy sellers consider this a full-time job.
In a two-sided marketplace, sellers are of paramount importance. A buyer may come to Etsy only a handful of times in a year, but a seller hopes to make a sale everyday and they are the ones who carefully gauge the trend to come up with products that can have demand from buyers. If there are more sellers on the platform and the average GMS per seller continues to increase, Etsy will inevitably be winner as well. So how big is the suppliers of sellers?
~30% working-age women i.e. ~23 mn in the US are currently homemakers. Among the OECD countries, this number is close to ~40%. Many of these unutilized/underutilized working age population can be a potential source of Etsy sellers.
Of course, not all of them are going to be successful as every marketplace has a Schumpeterian component in place. 45-50% sellers stop selling after year one. But as you can see from the data by seller cohorts below, among the sellers who survive and remain on the platform consistently enjoyed >40% CAGR growth over four-year period. After some lull in sellers growth during 2016-2018, the growth has accelerated again in 2019 with 20% growth in sellers and the pace has obviously picked up further during Covid-19 as number of sellers crossed 3 million on Etsy’s platform.
Given ~50% of sellers are based outside of the US, it is likely that seller growth may continue for quite some time. If a Moroccan rug seller can sell $5,000 worth of products on Etsy every year, that’s a pretty sizable source (assuming ~20% margin) of income for him/her (Morocco’s GDP per capita is ~3,300/year). During my call with Etsy’s IR, they mentioned they are really eyeing the Indian market for potential sellers. Coming from South Asian background, I know for a fact that handmade items are incredibly popular in this part of the world, and there is a pretty robust network of sellers (mostly women) which could be a good fit for Etsy platform. There are some structural headwinds such as payments that may need to be solved, but I think over time, these issues can be dealt with. Almost all of these South Asian countries have less than $2,000 GDP per capita (again, annually). Therefore, even if selling on Etsy can be a good side-income or a way to pass time for people in the US, it can potentially be full-time job for many international sellers and hence, active sellers growth on Etsy platform can possibly continue for years.
Can Etsy be toppled by another marketplace?
Etsy came to IPO on April 16, 2015. Six months later, Amazon launched Amazon Handmade. You know this drill; stock was cut in half three months after Amazon Handmade was launched. Etsy is almost ~20x since February 2016 lows.
I think the debate between Etsy vs Amazon Handmade should be over by now. To be clear, this is coming from an Amazon shareholder. Amazon is associated with low prices and/or faster delivery. “Handmade”, “Vintage”, or “Craft” are not the type of words most people associate with Amazon. Even though Etsy has emphasized on free shipping recently, Etsy management clearly mentions that they don’t pretend to be as efficient as Amazon is in terms of shipping/delivery. Etsy will always be behind Amazon in those metrics. Etsy’s motto is “Keep Commerce Human” and I think the type of consumers who buy vintage/craft items resonate with such messages more than fast shipping.
Of course, Amazon can fund Handmade and try to undercut Etsy by luring sellers to their platform. First, that doesn’t solve the demand aggregation problem as I indicated the vast majority of customers don’t really think about Amazon while buying the kind of products Etsy sells. More importantly, Amazon retail is truly a behemoth and I estimate Etsy’s GMS to be only ~1.5% of Amazon’s 1P+3P GMS. We may live in a ZIRP world, but the greatest capital allocator in the twenty-first century i.e. Bezos is unlikely to fund a segment to bleed money for years that simply doesn’t move the needle for them. I wouldn’t be surprised if Amazon handmade is shut down sometime in near future.
Moreover, experience on Etsy marketplace is reasonably differentiated from your experience in Amazon. ~30% items on Etsy can be customized which typically have ~30% higher Average Order Value (AOV). Sometimes, buyers initiate conversation with sellers to customize the item and conversion rate and AOV after buyer-initiated conversation is ~12x and 200% respectively.
It is difficult to replicate such experience in Amazon/other commoditized marketplaces. Although Etsy cannot compete with shipping time against Amazon, it has just come up with a new strategy: Etsy sellers can upload videos to show the buyers how their products are being made. After launching it in May 2020, ~700k videos have already been uploaded. This seems to be a clever strategy to change the nature of the game (you cannot compete against Amazon on shipping time) and re-emphasize the core message of Etsy: keep commerce human. As buyers see how their products are being made by the seller, buyers are much more willing to tolerate longer delivery time. This also reminded me of the famous elevator-mirror story. It, however, also raises the possibility of buyer leaving the marketplace and building a rapport with the seller which may prompt buyers to shop directly on seller’s website. Etsy does not seem to be too concerned with this as they believe buyers value the variety a marketplace offers which no one seller, however popular, can recreate.
It's not just Amazon, I do think other potential new or existing players are also likely to find it difficult to pose a potent threat to Etsy because of two factors:
First, Etsy enjoys cross-border network effects. Unlike Uber which enjoys hyperlocal network effects, increased buyers/sellers in one geography also enhances the network effects both intra and inter-countries. Second, its suppliers are reasonably differentiated which means increasing supply of sellers improves buyers experience further. There is no visible asymptote from buyers’ perspective with increased supplies of sellers. When Uber increases drivers on its platform, it reduces wait times for riders. But after a point, the utility of reduced wait times hits an asymptote and reducing wait times further does not really lead to greater utility/value for riders. However, Etsy’s two-sided marketplace does not have such a problem. If tomorrow 10 million more sellers start selling on Etsy, it will materially enhance product offerings and improve buyers experience on the marketplace. I believe Etsy is in the tier 1 of the marketplaces: highly scalable and highly defensible. (recommended reading on different tiers of marketplaces by Sameer Singh from whom I borrowed the below image as well)
Interestingly, when I spoke to Etsy’s IR team, they mentioned that from their surveys they came to know that of the people who are interested in handmade/craft/vintage items, ~45% shop from Etsy and ~30% go to local craft fairs/markets.
Therefore, the biggest competition for Etsy seems to be the local craft fairs or markets. There does not seem to be a credible threat from other marketplaces yet. If you want to explore other possible Etsy alternatives and their fees, click here to access a google spreadsheet that contains this information.
How Etsy makes money
Etsy segments its revenue in two ways: marketplace revenue, and service revenue which I discuss below.
Marketplace revenue (72.5% of revenues): Marketplace revenue consists of primarily two things: transaction revenue, and payments revenue. In 2019, marketplace revenue was 11.9% of GMS, up from 8.7% in 2015.
Transactions revenue: Etsy takes 5% of the selling price as transaction fee. It was 3.5% until July 15, 2018 after which it was increased to 5%.
Etsy sellers also pay $0.20/listing on the site. You can list as many products as you want for four months for $0.20/listing fees. If it’s not sold within four months, you can list them for another four months for another $0.20/item. The higher the number of sellers, the higher the listings on Etsy. In 2019, Etsy had more than 66 million listings.
Since August 4 2020, Etsy also increased its transaction fee from 3.5% to 5% on Reverb which, as mentioned earlier, they acquired in August 2019.
Although Etsy does not disclose revenue mix by transaction revenue and payment revenue within marketplace revenue segment, based on some reasonable assumptions (discussed later in the valuation section) we can conclude transaction revenue is ~75% of marketplace revenue.
Payments revenue: Etsy payments provides 10 different payment options to buyers while shopping on the site. Etsy charges 3% (including shipping fees and sales tax) +$0.25 on each sale made on the platform. After including exchange rate fees for international GMS which was 36% of overall GMS, the payment fees get closer to 4%. The cost for Etsy to process these payments is ~2-2.5% of GMS. Therefore, this is relatively lower gross margin (~40-50% vs 65-70% overall) business for them.
Until 2017, Etsy payments was known as “Direct Checkout” and at the end of 2016, 45.5% of active sellers used “Direct Checkout”. However, on May 17, 2017, Etsy rebranded “Direct Checkout” as “Etsy Payments” and made it mandatory for all active sellers on the platform. This is a general theme that I believe will continue to be a recurring theme going forward. Etsy would launch one of these “seller services”, invest on the service and patiently wait for adoption by critical number of sellers, and then make it mandatory to drive bottom line. In Etsy’s defense, it did mention that shops that allow Etsy Payments saw 49% higher sales than shops that didn’t during the time it was optional.
However, please note that Etsy payments is not available in all the countries. In 2019, 89% of all payments were processed via Etsy Payments. It was only 62% in 2015, so making it mandatory clearly worked in favor of Etsy.
Service revenue (27.5% of revenues): Service revenues consistently outpaced marketplace in terms of growth over the last five years. It consists of promoted listings or Etsy ads, offsite ads, shipping labels, Pattern etc. Again, Etsy does not disclose the revenue mix of these products within service revenue segment. In 2019, service revenue was 4.5% of GMS, up from 2.7% in 2015.
Etsy Ads: Etsy ads are based on auction for ad space on Etsy’s properties and fees are charged on Cost-per-Click (CPC) basis. In 2019, 16.6% of active sellers advertised on Etsy. This number has hovered around 15-17% in last 5 years, indicating only sub-section of the sellers choosing to use this service. Etsy ads are also known as “promoted listings”.
Etsy mentioned they want to keep a balance between paid and organic search within the platform and they do not expect no more than 30% of the sellers to use Etsy ads within the platform.
Revenue from promoted listings increased by 41.4% and 51.3% respectively in 2018 and 2019.
Offsite Ads: On February 26, 2020, Etsy declared sellers using Etsy ads would only be able to show their ads on Etsy and removed Google shopping from advertising tool. However, Etsy didn’t remove Google shopping altogether. Instead it launched “Offsite ads”.
What is offsite ads? Under this program, Etsy itself will advertise sellers' products on Google, Facebook, Instagram, Pinterest, and Bing. If a shopper clicks on any of the ads and buys product from the shop in the next 30 days, Etsy will take 15% of that. Sellers won’t incur any costs per click, but only for the actual sales. Sellers can opt out of Offsite ads program if they want.
The devil is, of course, in the details. If a seller sold more than $10,000 on Etsy in last 12 months, Etsy’s take rates from this product goes down to 12%. However, such a seller has no way of opting out of the offsite ads program.
I thought this was a clever way to increase take rates for Etsy. Etsy indicated ~10% sales will come from offsite ads. Offsite ads will be lower gross margin business compared to Etsy ads since it does take the risk of CPC if the customer doesn’t end up buying the product in which case Etsy pays for CPC but seller doesn’t pay anything to Etsy.
I think over time Etsy will get better at execution with more data and iteration. The better Etsy gets at promoting sellers’ products, the more sales will take place at Etsy and both sellers and Etsy will be better off.
Shipping labels: Shipping labels allow Etsy sellers in the US, Canada, UK, and Australia to purchase discounted shipping labels. Sellers can ship orders with USPS, FedEx or Canada Post from their Etsy Shop. Since Etsy has the scale, Etsy can purchase discounted shipping labels from delivery service companies and then it can resell to Etsy sellers at a markup which is still lower than what Etsy seller would be able to get on his/her own from FedEx/USPS. Although shipping labels likely contribute very little to topline, it is an extremely high margin business since revenue is recognized net of cost of purchasing the labels. Therefore, it’s likely to be >90% gross margin business.
They also have other services such as Pattern by which Etsy sellers can build their own websites/branded shop. Etsy mentioned 2.5% of active sellers used Pattern in 2017 after which they stopped disclosing numbers related to Pattern.
To understand the unit economics of Etsy platform, let me focus on 2016 buyer cohort. In 2016, 4.5 mn new active buyers were added to Etsy platform. So, what is the unit economics of a buyer on the Etsy marketplace?
As discussed in “Who shops at Etsy?”, we know their retention rate and how much this cohort spent on an average in the next four years. Given the stickiness of the habitual buyers/repeat buyers after the initial large churn in second year, I have assumed a large chunk of these retained customers will continue to shop for a long time. We know the take-rates (revenue as % of GMS) for 2016-2019 and I inserted my forecasted take-rates (discussed later in valuation). Gross margins were kept at 70% throughout the years.
Other costs e.g. G&A, R&D etc. as % of sales were gradually assumed to come down with scale benefits. SG&A line item requires some explanation.
To estimate the Cost to Acquire a Customer (CAC), I first started with the assumption that in 2016 17.5% of marketing expense was maintenance and the rest 82.5% was growth marketing expense, which is the relevant number for estimating CAC metric. Since Etsy is a two-sided marketplace, not all growth marketing is spent on acquiring buyers, and some are also spent on driving sellers to the marketplace. I assumed 70% of growth marketing was spent on acquiring buyers. For example, Etsy spent $82 mn on marketing in 2016, 82.5% of which i.e. $68 mn was growth marketing. Of the growth marketing, 70% or $47 mn was invested to acquire 4.5 mn buyers in 2016. Dividing $47 mn by 4.5 mn, we get CAC of $10.5/buyer. I assumed maintenance marketing expense for this cohort as 10% of sales going forward.
Based on these assumptions, I could see $24 value add per customer over next 10 years. Since some of these customers will probably continue to shop on Etsy even after that, lifetime value is even higher. Again, the sum of PV below is NOT lifetime value of the customer; it’s just a customer economics of a particular cohort for next 10 years.
How about ROIC?
Margins never tell the whole story. At the end of the day, ROIC matters much more than just mere margins. There is a common (mis)perception that marketplaces are low capital-intensive business. That may seem true for Etsy as well since capex as % of sales was 1.9% in 2019. Of course, a two-sided marketplace does not invest through their balance sheet. If you are looking for how they use their capital to grow their business, you need to peek through their Income Statement.
If you follow me on Twitter, you know I am a big Mauboussin fan. Mauboussin and Callahan recently wrote a paper on how to calculate ROIC for the companies who primarily invest through their Income Statement (Click here to read the thread). I have incorporated my learnings from the paper and then made several adjustments to calculate Etsy’s ROIC.
The big idea is fairly simple. Etsy spent ~26% of its revenue for marketing in 2019. While some of it is spent to reach out to existing buyers and sellers, the vast majority is spent to acquire new customers to grow the platform.
So, the idea is to segment marketing expense into two categories: maintenance marketing, and growth marketing. For 2015, I assumed 85% of marketing expense as growth marketing and the rest as maintenance marketing. While the paper assumed a fixed mix between maintenance and growth marketing, I gradually increased maintenance marketing (25% in 2019) and decreased growth marketing (75% in 2019) to reflect the reality that as the marketplace gets bigger, you have lot more incentive and reasons to reach out to your existing customers first than trying to continuously run after new customers whose CAC is likely to increase in future.
Similarly, I assumed 10% product development i.e. R&D expense as maintenance and the rest as growth R&D in 2015. For G&A, it was 30% maintenance, and the rest are growth G&A in 2015. I have gradually increased maintenance expense for these line items as well. If you are slightly confused, I recommend reading the thread linked earlier. It may also be more clear after going through my valuation discussion. If you are still a bit confused, I suggest you download my model and dig further. I am sure you will not remain confused after that. But if you do, send me an email and we can discuss further.
After capitalizing these expenses, I amortized them in the following schedule: marketing (2 years), product development (6 years), and G&A (2 years). I added back amortization to the earnings and added the capitalized intangibles, net of amortization to invested capital.
To simplify the tax calculation, I just assumed a normalized 21% tax rate for all the years.
After making these adjustments, Etsy’s ROIC came to ~20% in 2018 and 2019. It is likely to shot to ~40% in 2020.
So folks, looking at ROIC, we know we have a good business at hand. Now we just have to pray that the business is trading at a reasonable valuation. Okay, let’s get to it then.
If you have read my deep dive on Uber, you already know I approach valuation from “expectations investing” perspective. But I suspect many will be first time readers on my website, so let me briefly illustrate my approach again.
I do not believe in price targets, nor do I think I have any special insight on risk-free rates, equity risk premium, or many other factors that goes into the blackhole of calculating discount rate. Given that 10-year treasury is yielding 0.6% these days, we have to be realistic about our equity return expectation. If you keep dreaming about the Pollyanna days of double-digit return in the equity market, you may sit in the sidelines for a long time.
So, what is my IRR expectation? If I am looking at a company that has a long-dated senior unsecured corporate bond (5-10 years) trading on the market, I look at the Yield-to-Worst (YTW) of that bond and add 300-400 bps to account for the fact shareholders come last in the capital structure waterfall.
Etsy does not have a long-dated corporate bond trading on the market. In 2019, it had a convertible bond due 2026 (the conversion price has already been exceeded by current market price following the massive Covid-19 rally). The effective interest rate on Etsy’s debt was ~4%. Adding 300-400 bps gets us to 7-8% expected IRR.
Now what? I basically built my Etsy model and made assumptions in a way to generate ~8% IRR. Finally, I eyeballed those assumptions to see whether I am comfortable with these assumptions and whether I feel a margin of safety embedded in my assumptions. Obviously, this is more art than science.
As you already know, since I bought Etsy (full disclosure: ~10% of my personal portfolio at an average cost of $111/share bought around third week of September) I am quite comfortable with the assumptions I had to make, and I believe IRR potential is likely to be double digit over the long-term. Let me elaborate further on my assumptions and explain my source of comfort.
Etsy model, assumptions, and my source of comfort
Modeling Etsy’s revenues: Etsy’s revenue is driven by GMS. GMS is driven by the liquidity of both buyers and sellers on the platform. I am giving you another reminder that I was less interested in forecasting these drivers myself and more interested in building the model in a way to generate ~7-8% IRR, but there are indeed shades of my views embedded in some of the numbers.
If you want to see how big of an impact the pandemic had on Etsy, look no further than GMS growth in 2020 which may exceed ~70% YoY growth. Please note that although we have 2020 numbers till 2Q’20, I did not just annualize it since Etsy enjoys seasonality in 4Q (~30-35% of annual numbers). Moreover, although it may appear 2021 numbers to be too low, I have considered the fact that mask sales were $346 mn till 2Q’20. Assuming ~$500 mn mask sales in 2020 and zero mask sales in 2021 (perhaps not realistic since some mask sales likely to persist beyond 2020), the implied adjusted GMS growth in 2021 is ~14%.
While growth of active sellers is assumed to go down over the years, it may prove to be more robust than assumed in the model. As Etsy has been looking into the Indian market, this can potentially unlock millions of sellers on the platform in future.
Marketplace revenue is driven by payments revenue and transaction revenue, which includes transaction fees and listing fees. I have gradually increased share of payments processed by Etsy payments in the platform to 95% (currently 89%) and the share of international GMS is also expected to continue to gradually increase. Although domestic payment processing fees is ~3.5% of GMS, with more international GMS, effective payment fees get close to ~4% because of the extra fees related to international exchange rates.
As explained earlier, Etsy takes 5% of GMS as transaction fees. This used to be 3.5% until mid-2018, and I do expect them to increase this rate again sometime in future. I asked Etsy IR whether they want to keep raising transaction fee on the marketplace. One possibility they hinted is instead of raising transaction fees percentage across the board, Etsy might try to raise it in some categories in future. For example, they may keep transaction fees for vintage items at 5% but raise the fee to 10% for homewares and home furnishings. The effective transaction fees can, hence, increase for the overall platform. Therefore, I gradually increased the transaction fee percentage for the platform from 5% in 2019 to 7.5% in 2030.
Another component within transaction revenue is listing fees ($0.2/listing per quarter). In 2019, there were more than 66 million items for sale on Etsy marketplace. Listing fees are driven by both the number of active sellers as well as the number of listing they upload on the marketplace. Although $0.20/listing may appear very insignificant, it really adds up and contributed significant amount of revenue to Etsy. This also allows Etsy to generate revenue even from the failed sellers on the marketplace.
Service revenue, which includes Etsy Ads, Offsite ads, shipping labels etc., has been high growth segment for Etsy. Etsy has been consistently introducing and exploring seller services to drive more revenue on its platform. In 2020, Etsy introduced Offsite Ads and I expect them to continue to come up with new products going forward. Etsy expanded service revenue as % of GMS by ~180 bps in last 5 years, and I assumed ~300 bps expansion over the next 10 years.
Etsy defines its take rates as “revenue as % of GMS”. This means everything that goes into the revenue as described above goes into their take rates calculation. As per this definition, in 2019, Etsy’s take rate was 16.5%, an expansion of ~500 bps over the last 5 years. As per the model, there is another ~630 bps expansion of take rates in next 10 years.
I know exactly what you are thinking: Is this a rake too far? (highly recommended reading). Let me address this concern.
There is no standard definition of take rates. Each marketplace defines it differently which makes it incredibly difficult to make apple-to-apple comparison. If you see the below image, you can see commission rates/take rates of different marketplaces/platforms (if you want to explore this more, click here).
Let’s take Amazon for example. Although Amazon typically does not disclose this, 2018’s 10-k intentionally or unintentionally left some clues and we can calculate Amazon’s take rates on 3P retail was 26.7%. This still excludes advertising fees 3P sellers spent on Amazon marketplace. When you add all that it perhaps exceeds ~30%. In fact, ~30% seems to be fairly common take rates for most of these marketplaces/platforms.
While there has been a deafening debate around “fairness” of take rates, the status quo indicates assuming ~23% take rates (that too a very liberal definition of take rates) 10 years from now for Etsy is perhaps far from aggressive. I have a simple framing in mind: In 5-10 years, is Etsy marketplace more valuable for sellers? I think the answer is extremely likely to be “Absolutely YES”. If that’s the case, will (and should) the owner of the marketplace enjoy a greater fruit of the pie?
I want to, however, make it clear that there is likely to be increasing tension between sellers and Etsy if they continue to increase take rates. How Etsy management nurtures the relationship with sellers while increasing take rates may well be the defining driver for the stock in the long run. We should expect a lot of press about dissatisfaction of Etsy sellers regarding increasing take rates in future. Let me elaborate with a personal example why I won’t be too concerned reading those reports.
The vast majority of my readers/subscribers (~70-80%) comes from twitter. If there were no twitter and I had no following, I would be forced to increase my subscription price >100x or close this website altogether. I cannot possibly overstate my dependence on twitter to acquire new subscribers/readers. This may sound like a philosophical argument borrowed from John Rawls’ “veil of ignorance”, but if Twitter, hypothetically speaking, wants a 30% cut from my revenues, it is not obvious to me why it should be inherently unfair. Now rest assured, I am quite accustomed to get my subscribers from twitter for free, and you can even expect some rage tweets from me about how Jack and his team is stealing from small, independent analysts like me if they ever could come up with a model that forces me to share a percentage of revenue. But what can I do?
I would suggest you to mostly ignore those tweets.
Again, granted that there is certainly a ceiling, however blurry, in terms of what marketplaces can charge as take rates, but I think I am comfortable assuming Etsy’s take rates ~700-800 bps lower than the status quo take rates of many other marketplaces.
Cost structure of the business: Gross margin for Etsy expanded by ~350 bps YoY in 1H’20, and I kept it 70% for each of the next 10 years. There might be some further tailwind for gross margin given the fact that the cost of offsite ads (previously Google shopping ads) was accounted in the “Cost of revenue” line whereas from the second quarter of 2020, these costs are now accounted in “Marketing expense”.
For the rest of the cost structure, as explained earlier, I segmented into two categories: maintenance expense and growth expense (which is essentially investing and hence was capitalized while calculating ROIC). With the passage of time, the growth expense is likely to come down as the marketplace becomes more mature. If you look closely, marketing expense as % of revenue is not expected to come down despite the massive revenue growth in 2020. This is due to the fact that Etsy management has decided to double down with marketing dollars to drive awareness and traffic to the marketplace. This makes perfect sense to me. Instead of making some eye-popping free cash flow margins, it makes sense to grab this Covid-19 opportunity with both hands and make sure everyone knows about the brand. The next job then would be to retain as many of these pandemic-driven customers as possible in 2021 and beyond.
When I spoke with the IR team, they mentioned Etsy barely spent any money on marketing dollar to acquire buyers during 2005-2012. The entire marketplace was basically driven by word of mouth.
Even in last few years, they spent mostly on bottom-of-the-funnel marketing, and during this pandemic, the management decided to spend more dollars on top-of-the-funnel as well to drive awareness of the broader brand. Etsy launched a recent TV ad to execute this strategy. Earlier, if a customer was looking for a vintage scarf and she googled the item, Etsy’s products would show up and in many cases, that customer would come back to Etsy only when she’s thinking about buying her next “vintage scarf”. Etsy says they found many customers were not aware of how broad the product selection is on Etsy marketplace and they wanted to build a broader marketing strategy to address these gaps.
Valuation: Once I have my three statements, I got the Operating Cash Flow (OCF) from Cash Flow Statement and deduct both capex and tax-adjusted Stock Based Compensation (SBC) to calculate Free Cash Flow (FCF). As Etsy has ~30% FCF margin, I used this FCF to buyback shares to make sure cash just doesn’t keep piling up on balance sheet. The buybacks may prove to be a bit more aggressive than what may happen in reality, especially in out years. They may also use this cash to make acquisitions which is of course not forecastable.
To get to ~8% IRR, I had to use 19x FCF multiple in 2030. Of course, predicting what the appropriate multiple is 10 years from now is a fool’s errand. But if Etsy continues to execute well on its marketplace, it will enjoy an unassailable lead on this space in 10 years. Then think about the excellent economics (both margins and ROIC) Etsy already has proved and the growth trajectory it may have going forward (notice double digit revenue growth even in 2030). If interest rate indeed remains lower for longer, do you really believe you will get this business at ~5% FCF yield in 2030? I don’t think so. It’s more likely to be ~25-30x FCF if those factors all go in the right direction. At 25x FCF, we will get 10.3% IRR and at 30x, IRR will reach to 11.8%. If you allow me to be a bit greedy, Etsy may also grow faster than assumed in my model. Remember Nick Sleep? Although he cautioned about assuming regression to mean while forecasting internet retailing, my model sort of implies some regression to mean. If Mr. Sleep’s insights prove to be right once again for Etsy as well, perhaps Etsy will make a mockery of this model and handily beat the projections. If that happens, IRR will…you get the idea!
Management quality and their Incentives
When Etsy named Josh Silverman as the new CEO on May 2, 2017, the stock fell 14% in after-hours trading to $11.39/share. Etsy went public in 2015 at $16 per share. Etsy is a 10-bagger since Silverman was declared the CEO.
I have alluded to this before and John Doerr would agree: execution really matters. This is a good business with excellent economics, but any business can suffer without relentless execution. Look at GMS growth pre and post Silverman. The way he and the team scaled Etsy is indeed a thing of beauty.
A Stanford MBA, Silverman was also the former CEO of Skype. Before you raise your eyebrows thinking Skype’s monumental failure, Silverman joined at Skype in 2008 when it was valued <$2 Bn. He left Skype in September 2010 and few months later, Skype was sold to Microsoft for $8.5 Bn. Before Skype, he was at eBay for ~4.5 years.
Current Chairman of the board is Fred Wilson. Wilson, founder of Union Square Ventures, backed Twitter, Tumblr, Zynga, Coinbase, and Etsy. He joined the board back in 2007 and has been serving as Chair of the board since 2017.
One thing that stood out to me at Etsy is how diverse their team is. Personally, I do not particularly look for ESG factors while investing but generally acknowledge that it’s difficult to maximize shareholder wealth without simultaneously taking care of the stakeholders. While every tech company in the Valley talks about diversity, Etsy does seem to walk the talk. Half the board is female and as of 2018, female engineers were 38% of the tech team, ~2x the typical number in Silicon Valley.
In terms of executive compensation, Etsy tracks GMS, revenue, and adjusted EBITDA. I would love to see ROIC and FCF in the mix, but given my comfort with Etsy’s economics, I do think growing GMS, revenue, and adjusted EBITDA margin (don’t worry, unlike Uber, they mostly just add back SBC for adjustments) will naturally lead to superior FCF generation. Having some sort of ROIC metric as part of the comp mix would be better though as that would mean management/board would think long and hard before going for an acquisition.
What to do going forward
I wish to be long-term shareholder of Etsy unless, of course, I discover some nasty surprises in near future. Next steps are fairly easy: just listen to the earnings call (expect me to write twitter threads after each quarterly earnings), and also follow (mostly casually) the sell-side conferences the company attends.
As always, please leave your feedback in the comments or via email. I love hearing from my readers. I am thinking about covering a non-tech name for my next deep dive.
If you are still not a subscriber, click here to make sure you do not miss the next deep dive. There is currently a free trial period till November 15 after which you will be billed $10 on a monthly basis. Future deep dives after November 15 will only be available to subscribers.
There are plenty of recommended reading materials I have mentioned in the body of the piece. I haven’t discussed Reverb at all, so I will leave you with a good piece on the revival of guitars: Guitars Are Back, Baby!
Disclaimer: All posts on “MBI Deep Dives” are for informational purposes only. This is not a recommendation to buy or sell securities discussed. Please do your own work before investing your money.