Infrequent Investment Ideas Vol.1: Why Buy Redbubble Limited’s Stock (ASX: RBL)?

Redbubble is an Australian company, which operates two online marketplaces, Redbubble.com and TeePublic.com, where artists can upload their graphic designs and customers can shop for various items on which these designs are printed. The printing is done by third-party fulfillers using digital printing technologies, which make printing small volume or even single unit orders economical. Artists receive royalties on the sale of each product bearing their design. Redbubble is the largest online marketplace that focuses solely on this niche industry, known as “print-on-demand.” Redbubble’s marketplace model is based on a flywheel effect, described in Redbubble’s annual report thus:

…the more artists in the marketplace, the more relevant content and so more reason for customers to come. The more customers the better the fulfilment network becomes, and this in turn brings back more customers. And with more customers more artists are attracted.

The pandemic was a boon to the business, and in calendar 2020, the company’s unique customers (unique e-mail addresses) and 659,000 selling artists (who have made a sale in the past 12 months) grew 63% and 59%, respectively. The value of products sold through its websites grew 77% to A$620m, of which Redbubble’s revenue was ~A$522m, and ~A$99m was paid out to artists. The business achieved a pre-tax margin of 6.7%, an EBITDA margin of 9.5%, and generated A$79m of free cash flow, a record amount for the company.

Yet, Redbubble’s stock has sold off significantly over the past 4 months – it’s now ~45% off its all-time high reached in February – first after the company reported its 2Q21 results in February, when it posted somewhat disappointing gross margins and decelerating sales growth relative to the previous quarter (yoy growth was still 84%). More importantly, the newly appointed CEO, Mike Ilczynski, told investors and sell-side analysts that he would focus on scaling the business, which would require investments, and short-term profitability would likely suffer, but he was short on detail. Then on the 3Q21 results call in April he laid out investment plans in somewhat vague terms, set financial goals for “CY24+” of A$1.25bn+ of sales and 10-15% EBITDA margins and said annual EBITDA margins would be cut in half, “in the short-term.” The CEO’s plans lacked specifics on the investments, but his goals were made clear – to scale the marketplace by growing the number of participants on both the artist and buyer side, as well as building loyalty/increasing retention rates for both groups. Essentially, Mr. Ilczynski will go for winning mindshare from both customers and artists.

Online marketplaces which become the salient option/the go-to destination for a particular set of products or services for a large enough number of customers and suppliers tend to dominate their markets. These marketplaces become a solution to the overabundance of choice but must first grapple with this overabundance to attract and sustain the attention of potential marketplace participants. Therefore, I think investing in marketing and in the improvement of both the customer and the artist experience will increase Redbubble’s chances of ultimately reaching that critical point of salience, and, hence, the expected value of the company.

The timing of the sell-off suggests the reasons for it were the lack of visibility and the sacrifice of short-term profitability to long-term oriented investments. This means the recent stock price drop has provided a time arbitrage opportunity. Surely, there are risks ahead, namely capital allocation mismanagement and some emerging legal risks to the print-on-demand industry, both overlaid by the constant struggle with competitors. I think the company’s competitive advantages, capital structure and economics, and the optionality embedded in its business model skew the range of the company’s intrinsic value heavily to the upside relative to its current valuation.

What does Redbubble do?

Section summary: Redbubble is an Australian company, which operates two online marketplaces, Redbubble.com and TeePublic.com, where artists can upload their graphic designs and customers can shop for various items on which these designs are printed. The printing is done by third-party fulfillers using digital printing technologies, which make printing small volume or even single unit orders economical. With A$666m of sales made on its websites over the past 12 months, Redbubble is the largest online marketplace that focuses solely on this niche industry, known as “print-on-demand.”

Redbubble is an Australian-based company, which operates two three-sided online marketplaces, Redbubble and TeePublic, connecting artists, customers, and fulfillers. An artist can upload a graphic design on redbubble.com or teepublic.com (the artist side); a customer can browse the design selection on redbubble.com or teepublic.com and buy a variety of items ranging from t-shirts and smartphone cases to bath mats, on which the chosen design is printed (the customer); once a customer makes a purchase on the website, the order is sent to one of Redbubble’s partners’ 41 fulfilment facilities, which print the design on the chosen blank item and ship it (in Redbubble or TeePublic-branded packaging) to the customer (the fulfiller side). For each item, Redbubble sets a base price, which includes the cost of printing and shipping and Redbubble’s gross margin. The artist then determines what percentage of the base price he or she would like to receive for each item bearing his or her design listed on redbubble.com, while teepublic.com offers fixed pay-outs depending on the item. The final price listed on the websites is the sum of the base price and the artist mark-up, so the artist pay-out is entirely passed on to the customer. For example, if the base price of a t-shirt is A$10, and the artist sets his or her mark-up/royalty at 20%, the final price of the t-shirt listed on redbubble.com will be A$12. The artist pays nothing to upload or maintain designs on either website. The fulfilment partners are paid for their services but do not pay Redbubble any up-front fees. Redbubble recognises revenue net of payments to artists, and only once the item has been delivered to the customer, typically in a few business days, depending on the shipping option chosen by the customer. Redbubble derives 69% of its revenue from North America, 14% from the EU, 11% from the UK, 6% from Australia and New Zealand, and 1% from the rest of the world. 

Redbubble was founded in 2006 to make use of advances in print-on-demand technology. At the turn of the century, the progress in digital printing had enabled economically feasible production of low volume or even single item orders in the printing industry. The economics made sense partially because digital printers do not require much of a set-up for the design to be printed, as compared with screen printers, resulting in much higher throughput. This obviated the need to accumulate inventory of finished goods in anticipation of demand and enabled low-volume production as demand came in, or “on-demand.” Exploring various print-on-demand business models, Redbubble’s founders eventually reasoned enough customers would buy apparel with different designs that “express who they are” or are perceived as “something unique,” to quote co-founder and ex-CEO Martin Hosking, rather than mass-produced brand-name clothes, and that print-on-demand technology could meet that demand. Redbubble was not the pioneer of print-on-demand e-commerce: CafePress, Custom Ink, and Spreadshirt had been founded a few years earlier but initially focused on allowing customers to upload their own designs, which would then be printed and shipped to them, rather than aggregating and selling third-party designs. These companies were essentially offering the same services to consumers that promotional products/corporate swag companies like 4imprint or Vistaprint offered to corporations, but without the minimum order quantities. The websites of the first consumer-focused print-on-demand e-commerce websites featured third-party designs as well, but the focus was on allowing customers to customize their own designs.

After initially dabbling with the concept of customer customization as well, Redbubble decided to create a catalogue of independent artists’ designs, from which customers can choose. Thus, should enough artists’ designs be uploaded, a sufficient amount of demand for very specific searches (“long-tail” in the industry jargon) would be met. This aggregation of artists’ designs, customer demand, and a high number of small volume orders for printers results in a meaningful value proposition to each side of Redbubble’s three-sided marketplace. Customers derive utility from purchasing products that “express who they are.” Artists get to sell their designs on three continents essentially effortlessly relative to the alternative of establishing their own shop (online or physical), managing fulfiller relationships or taking on fulfilment themselves, and everything that these endeavours entail. Redbubble directs volume to printers, thereby assuming the marketing, sales, customer support, and payment processing costs that the printer would otherwise have to incur to secure that volume. The marketplace model is based on a flywheel effect, described in Redbubble’s annual report thus:

…the more artists in the marketplace, the more relevant content and so more reason for customers to come. The more customers the better the fulfilment network becomes, and this in turn brings back more customers. And with more customers more artists are attracted.

What do the business’ economics look like, and what is the main driver of value creation (for all marketplace participants and shareholders)?

Section summary: As an online marketplace operator, Redbubble requires little capital expenditures, and its negative working capital cycle is a source of financing. Its fulfilment costs are mostly variable in nature, and the artists’ payments are mark-ups determined by the artists and entirely passed through to the customer. Marketing is one of the company’s largest and most important investments, albeit accounted for as an expense. Still in a period of rapid growth (greatly helped by the pandemic), Redbubble pays to acquire a large portion (40%) of its revenue through targeted advertising, but the majority (60%) of sales come through free (“organic”) channels. Therefore sales made through organic routes are significantly more profitable on a gross profit after paid acquisition (“GPAPA”) basis. The remainder of Redbubble’s costs – software, labour, leases – are more fixed in nature in any given period. Hence, Redbubble’s operating leverage and cash generation are a function of GPAPA, which, in turn, will be determined by Redbubble’s ability to build brand awareness and customer loyalty i.e. mindshare, in order to grow organic sales in its revenue mix.

Over the past twelve months, Redbubble took A$0.34 off every Australian dollar of sales (net of payment platform fees) made on its website, while the artists and the fulfillers took A$0.16 and A$0.50, respectively. Supplier payables are typically settled within around 4 weeks, and artists are paid monthly, but only once their earnings exceed A$20 (TeePublic doesn’t have a minimum). Since the production is outsourced to its fulfiller partners, Redbubble holds no inventory, and its operations require no meaningful investments in property, plant, and equipment. These features of the business model result in a negative working capital cycle and require negligible capital expenditures (mostly capitalized software development) relative to the entire cost basis.

Redbubble only recognizes A$0.84 as revenue, referred to as “marketplace revenue.” Artists’ payments as a percentage of value of merchandise sold on Redbubble’s websites (“gross transaction value” or “GTV”) are likely to be fixed and could even grow, but they are mere (artist-determined) mark-ups on the base price of a product. Considering artists are paid monthly, an increase in artists’ payments would actually enhance the company’s liquidity in any given period. Fulfiller costs are dependent on revenue mix but are likely to decrease with GTV growth, as Redbubble commands volume discounts from its partners.

Outside of fulfiller costs, the bulk of Redbubble’s costs are marketing, employee and contractor expenses, software services (Redbubble uses AWS), and lease payments. Employee and contractor costs have grown at a lower rate than GTV and revenue, thus providing some scope for operating leverage. But it is marketing expense that is the crucial investment, whose success will determine the business’ economic future.

A significant part of software development costs are capitalised, and, since Redbubble operates two online marketplaces, these costs’ accounting treatment is an accurate representation of economic reality in that it results in the creation of an asset to the business. Accounting conventions fall short when it comes to their treatment of marketing expenses, however – marketing is vital to the success (or failure) of Redbubble, yet marketing costs are expensed. As a result, Redbubble’s balance sheet showed an invested capital base of negative A$4m at the end of 1H21, yet, the company had spent A$68m on marketing in the prior 12 months.

As an online marketplace, Redbubble is only successful in attracting customer traffic if it has a large enough selection of designs, and it is only successful in attracting artists to upload their designs if the websites have enough customer traffic. Once Redbubble attracted enough artists to the platform to spark customer interest, more artists joined the platform, and the expanded design selection brought in more customers, thus creating the flywheel effect. But those customers did not all come to Redbubble through long-tail Google searches. In fact, organic/free traffic (unpaid search, e-mails, direct website visits) has accounted for only ~60% of Redbubble’s revenue almost every year since it became a publicly traded company in 2016 (it was 70% in FY16). The rest of the revenue was generated through paid advertising. In calendar 2020, Redbubble spent A$68m in paid acquisition costs, generating A$248m of GTV. Assuming the overall artists’ payments percentage and gross margin that Redbubble reported for the period applies to this “paid” GTV, the company’s marketing investments generated A$85m of gross profits (132% of the cost of marketing) – paid GTV was profitable on a GPAPA basis.

Somewhat counterintuitively, paid acquisitions are the crux of organic/free GTV growth and, hence, the flywheel effect. Organic GTV comes primarily from either search engine-directed or direct website visits. Direct website visits come from customers who are already familiar with Redbubble or from returning customers. These customers also likely have higher conversion rates, compared to customers who have arrived at redbubble.com or teepublic.com through a search engine – there is a higher probability that a customer who has searched for an item on Google will end up purchasing from one of the other vendors that came up in the query. Therefore, these visitors are not only likely more profitable in aggregate but have also yielded some “share of mind” to Redbubble; that is, Redbubble is a pre-existing option in their mind when they feel the urge to buy a particular product or just browse through the offerings. But to win this mindshare in an increasingly competitive battle for customers’ attention, Redbubble needs more than skilful SEO for long-tail searches. It needs more targeted advertisement and general marketing to bring customers to the website and then provide an excellent customer experience.

Etsy, a formidable Redbubble competitor, has managed to achieve this on a large scale – as of 1Q21, it had 44.3m “repeat buyers,” which Etsy defines as buyers who have made 2 or more purchases over the past 12 months, of which 7.9m “habitual buyers,” buyers with 6 or more purchase days and over $200 in spend in the trailing 12-months. Redbubble doesn’t disclose equivalent metrics but as of 1H21 only had 9.2m unique customers, defined as “an unique email address; does not account for overlaps between Redbubble and TeePublic.” Redbubble defines repeat customers as “customers who have previously purchased, regardless of the date of their initial purchase,” and those have consistently made up ~40% of GTV over the past few years. Etsy.com is now in the top five most visited e-commerce websites in the US and in the top ten globally. Redbubble’s ranking in this particular category is more difficult to find, but, for comparison’s sake, Etsy.com is ranked as the 32nd most visited website in the US (across all categories), while Redbubble is ranked 761st by similarweb.com.

Etsy spent ~$501m in total on marketing in its FY20, or 29% of revenues, and this proportion has increased almost 10 percentage points since Etsy’s IPO in 2015. Redbubble’s marketing spend has also trended up as a proportion of revenue over the period but from ~9% to ~13%. This difference is all the more remarkable considering that Etsy has generated about 80% of its revenues from organic channels since it became public (this is partly a function of the fact that Etsy merchants likely spend considerably more time and effort on marketing than Redbubble artists for reasons to be discussed on pages 6 and 7).

Even only comparing Etsy’s performance marketing efficiency to Redbubble’s overall return on marketing, Redbubble emerges as a significant outperformer, or perhaps under-spender. FY20 Etsy spent $333m on performance marketing to generate ~$1.9bn of GTV (it calls it “GMS”). Overall, Etsy’s take rate was about 17%, and its gross margin was 76% in FY20. Assuming this take-rate applied to the paid GTV as well, Etsy generated $339m in revenue through performance marketing. Etsy’s COGS (“cost of revenue”) include cloud-hosting and other technology costs, which Redbubble accounts for in its operating expenses, so Etsy’s like-for-like gross margin would be even higher. Regardless, even if the $339m of revenue was generated at no cost, Etsy’s FY20 performance marketing would still only have generated 102% of its cost in gross profit, compared with Redbubble’s 132%.

This comparison is not made to imply that Redbubble should bring its marketing as a percent of revenue to Etsy’s level – part of the reason Redbubble has managed to achieve higher returns on marketing is that its paid acquisitions focus on cheaper, long-tail search ads – but to highlight that marketing is vital to the successful scaling of an online marketplace.

The “Investment in Marketing” section of Etsy’s FY20 10-K starts with the following sentence: “We are focusing on initiatives to drive traffic to Etsy and shape perceptions of the Etsy marketplace as the go-to shopping destination for everyday items that have meaning and those ‘special’ purchase occasions that happen throughout the year.“ Although Redbubble is not aiming to be the go-to shopping destination for “everyday items,” but perhaps for “unique” designs printed on various items, it does broadly share these goals, and Etsy’s management believes that their achievement requires significant investment in marketing. If marketing is a requirement for Etsy’s business, it is vital for Redbubble’ growth.  

Redbubble can be a very high returns business if/when it scales. Over the past 12 months, Redbubble’s GTV grew 82% to A$666m, payments to artists (“artists’ revenue”) grew 84.5% to A$107m, and the company’s marketplace revenue grew 82% to A$107m. Payments to fulfillers were A$333m (up 73% yoy), bringing the business’ gross margin to 40.5% (37.4% in the prior period). Marketing expenses were A$72m (doubling yoy), which resulted in a GPAPA margin of 27.7% (25.7% in the prior period). Other operating expenses were A$108m (A$96m in the prior period), bringing pre-tax income to A$47m (-A$17m in the prior period). As the company is growing and only recognises revenue upon delivery, revenue in a given period understates the cash inflows from sales. Most of the marketing costs are recognised as incurred, however; hence, the income statement understates the company’s free cash flow, and the cash return on marketing is even higher than described above. Redbubble’s fiscal year ends in June, and the company was exempted from publishing its quarterly 4C reports containing cash flow statements in October 2019. Therefore the full cash flow statements for the last twelve months are unavailable. They are available for the first half of FY21 (the six months ended March 31^st, 2021), which show that Redbubble generated A$79m of FCF (including lease payments; the company is debt-free), while net income was A$41m, with the difference mostly due to receipts from customers yet to be recognised as revenue. In FY20, Redbubble generated FCF to equity of A$37m on GTV of A$416m and customer receipts of A$472m.

What does the competitive landscape look like and what is Redbubble’s competitive advantage?

Section summary: In a broad sense, Redbubble is a retail company. Like other online retailers, it greatly benefited from lockdowns over the past year, as evidenced by its accelerating sales growth – sales rose 83% yoy on an LTM basis compared with 36% in FY20 – a quarter of which came from masks, a non-existent category prior to the pandemic. As its markets are lifting lockdowns, Redbubble will be facing increasing competition from consumer services and physical retail, not only for dollars, but also for the attention of consumers who are presumably spending less time online. Of course, this is a headwind that every online retailer is facing, but it will probably be felt more strongly by smaller players who are fighting for consumer mindshare. More specifically, in the category of non-branded apparel, homeware and accessories, on which various designs are printed “on demand,” Redbubble is facing competition from three broad categories of competitors, discussed below. Redbubble has competitive advantages against each category on either the artist or seller side, or sometimes both. 

Print-on-demand online marketplaces

There are many print-on-demand online marketplaces competing with Redbubble (Society6, Zazzle, Teespring, Spreadshirt, Threadless, Minted, CafePress, the list goes on) with different offerings: third-party art, third-party art with customizable components, blank items and printing services allowing you to upload your own design and have it printed and shipped. Redbubble is the largest among these online marketplaces specialising in print-on-demand in terms of revenue, traffic, buyers, and selling artists.

The flywheel effect is the driver of growth of every online marketplace and is sought-after by all of Redbubble’s marketplace competitors. However, flywheel effects are a function of size, and network heuristics would suggest that this relationship is not linear. Etsy’s spectacular growth illustrates this and seems to be the latest example of the “winner takes most” dynamics of online marketplaces. Consequently, Redbubble’s size relative to this subset of competitors is a significant competitive advantage. It’s difficult to think up ways in which competitors can displace Redbubble: its large art catalogue and significant SEO investments, both accumulated over a longer period than most competitors, ensure that Redbubble.com is among the top results in search queries (and Google shopping); despite the Etsy example, Redbubble’s marketing spend most likely eclipses that of its pure print-on-demand competitors, which means Redbubble has an advantage in attracting traffic through paid search and other paid channels as well; this peer-leading level of traffic makes Redbubble the go-to destination for art monetization for artists.

Many artists sell on other platforms as well, but, if Redbubble is the platform with the most artists, there also must be many who don’t consider uploading their art on lower traffic websites a worthwhile endeavour. Some competitors have focused on certain niche blank products (such as Society6 with Deny Designs in home accessories), and niche focus is how Redbubble built its marketplace and got the flywheel in motion in the first place. But at this stage of the game, intense focus on a particular niche is at odds with the “winner takes most” dynamics – it is unlikely that the industry will end up segmented in many small players who become the “go-to destinations” for various small niches (e.g. the go-to for printed shower curtains); or at least it will be a precedent in online commerce where large aggregators have so far been able to “take most.”

Etsy and Amazon

There are plenty of merchants who sell quirky designs printed on various items on Etsy. Large print shops such as Gelato, Printifi, and Printful have also developed partnerships and integrations with Etsy, allowing these merchants to print on demand and drop-ship products to customers (i.e. completely outsource production and fulfilment). Etsy is a much larger marketplace – its ~160m buyers (90m who made a purchase in the past year) dwarf Redbubble’s 9.2 million unique customers. If traffic was the only factor attracting artists, Etsy would be a no-brainer first choice for artists, or maybe second to Merch by Amazon. Artists can upload designs on Merch by Amazon; Amazon takes care of printing, shipping, and listing the product on amazon.com. However, one needs to apply to be accepted by Merch by Amazon, and demonstrating sales on other platforms is viewed quite favourably in the application process. Merch by Amazon is also somewhat limited in its product range, offering primarily apparel. Other than the application, Merch by Amazon requires a significantly lower investment (of time and money) than Etsy. Amazon doesn’t require any up-front fees and it handles all of the communication with the customer, including returns. Etsy has a $0.20 listing fee for every item you list on the website, and it’s the merchant/artist who handles communication with the customer, as well as returns, even if he or she uses a print shop’s drop-shipping services (they might accept the returned item but refunds and replacements are to be handled and paid for by the merchant). Amazon has other drawbacks from an artists’ perspective, however, as the listed item is often perceived as simply another Amazon product, and therefore it is difficult for artists to establish a customer following.

Considering Etsy’s and Amazon’s scale, size is seemingly a competitive disadvantage for Redbubble against these e-commerce giants. If customer traffic is the driving force behind the artist side of the flywheel, Redbubble doesn’t stand a chance. But I would argue that Redbubble has a different value proposition for artists, particularly compared to Etsy, which serves as a competitive advantage. Setting up an account on Redbubble and uploading a design takes a few minutes, and that is about all of the engagement required in the sales process. Of course, artists can market their work through other channels. But the customer communication, returns, production, and shipping is all handled by Redbubble. However, you can’t get something for nothing, and this lack of effort required from the artist is reflected in the seemingly small artist earnings – Redbubble’s average annual artist payment (annual artists’ payments divided by the number of selling artists) has averaged only ~$120 since FY15. Nevertheless, like with most online marketplaces, the distribution of artists’ payments is heavily skewed, as about 80% of sales are made by 12% of sellers. The average annual payment for this group of artists has averaged ~$807. Further down the tail, 4% of artists account for 60% of sales on Redbubble. For these top-selling artists, the average annual payment has been ~$1,816.

It is worth comparing the economics faced by an Etsy seller and an artist using Redbubble. Over the past twelve months, Etsy’s average GTV per seller has been $2,483. Etsy’s take rate has averaged about 17% over the period, leaving $2,061 for the merchant on average. Etsy sells significantly pricier items (furniture, jewellery etc.) compared to Redbubble, which is evident in Redbubble’s lower average LTM GTV per seller of $684. Accordingly, merchants who sell printed products directly competing with Redbubble’s probably sell below the Etsy average. Etsy is also not a typical online marketplace, since it focuses on crafts products, which are the antithesis of mass production. That is, by definition, the craftspeople who sell on Etsy can only scale their production so much before it seizes to be a crafts product. As a result, Etsy’s GMS is not quite as skewed towards its most successful sellers as is the case with Redbubble and most online marketplaces. Etsy doesn’t explicitly disclose this distribution, but it does disclose some figures on seller retention, and one can work out that the longest-surviving cohort in those disclosures, those sellers who have been actively selling on the site for 4 years, accounted for 20% of the total number of sellers and 51% of Etsy’s GTV in FY20. The average GTV per seller in this successful subset of sellers was $6,538 in FY20, and it had grown at an astounding 41% CAGR since they started selling on Etsy.

The limit on production scale might not apply to an artist using a drop-shipping printing service because the large print shops integrated with Etsy have sufficient capacity and large geographic footprints. However, selling in multiple geographies is probably quite costly for small merchandisers, in terms of payments to fulfilment partners. Maybe there is a volume discount for orders aggregated by Etsy, but I doubt that aggregation is as seamless as with Redbubble. Redbubble paid 50% of GTV to its fulfilment partners over the past 12 months; and Redbubble shelled out ~$5.5m to its average fulfilment partner during the period. So, factoring in a 17% average take-rate from Etsy and more than 50% to fulfillers (let’s say between 53% and 63% to round up the numbers), an Etsy merchant selling printed t-shirts, for example, took in 20%-30% of sales he or she made one etsy.com. Assuming his or her sales were around the overall Etsy GTV per seller average of $2,483, this seller made ~$500 and ~$745 over the past year. Assuming the seller is an average representative of the 20% of sellers who have sold on Etsy for 4 years, and generated sales around this group’s average GTV of $6,538, he or she would have taken in between ~$1,300 and ~$2,000. These amounts are optically significantly higher than the average take of a Redbubble artist and even the average earnings of the top 12% of Redbubble artists (but not for the top 4% on Redbubble). But these numbers might exclude other costs and certainly exclude time and effort/opportunity costs – Etsy surveys of its sellers suggest that its sellers spend a bit more than half of their time creating their product, with the rest dedicated to inventory management, shipping, customer communications, admin etc. Inventory management is not an issue when drop-shipping, but they have to manage the relationship with the fulfillers. Selling on Redbubble requires none of these efforts. It’s not at all obvious that the incremental effort required to sell on Etsy rather than Redbubble would be worth it (the above comparisons were made after converting Redbubble’s artist metrics into USD).

Furthermore, many artists might have different – asdf asdf motivations when beginning the proces0073 of monetizing their designs and might not even consider Etsy as an alternative to Redbubble (or vice versa), depending on these motivations. According to Etsy’s 2021 survey of their sellers, “69% consider their Etsy shop to be a business,” “55% are multi-channel sellers,” and “30% of Etsy sellers were pursuing their creative business as their sole occupation.” I don’t think Redbubble attracts many of that cohort of sellers; perhaps it is a more suitable option for some of the 65% who “started their Etsy shop as a way to supplement income” (obviously some of these intersect with the other responder subsets). Neither does it strive to: Redbubble doesn’t market itself as a platform on which one can build a business, but rather as a platform to display and monetize one’s “art,” which is evident when the company talks of “artists” when Etsy talks of “sellers.”  

Merch by Amazon requires about the same amount of effort to set up as Redbubble (once your application has been approved). Artists’ payments are fixed and vary between 15% and 30% on Merch by Amazon, depending on the product. Amazon certainly offers more customer traffic, a wider geographic reach, better shipping terms and, thus, better customer experience. But building a following of customers who found that your designs and general style “express who they are” through Merch by Amazon is virtually impossible. Products created through Merch are listed on amazon.com as “Brand: Unknown.” In contrast, on redbubble.com, there is a very visible “created by” line under each product. One can: 1) click on the artist profile and follow the artist to be alerted whenever he or she posts a new design; 2) add their designs to “favourites”; 3) message the artist if the artist has allowed that option; 4) see their social media profiles if they have provided links in their profile page; 5) see who else follows the artist and what other artists the artist follows. TeePublic.com offers all of these options as well and goes a step further – some artists have “Hire This Artist” options on their profiles, allowing customers to hire that artist for projects. TeePublic makes it explicitly clear that it is in no way involved in this relationship. In effect, TeePublic has allowed artists to advertise their work on their website for free; in fact, artists can make money selling their designs in the process.

On the other hand, Merch by Amazon applications are more easily approved if the artist already has an Amazon seller account or demonstrates success in selling on other platforms. Consequently, although Amazon doesn’t disclose Merch metrics, it’s quite possible that artists earn substantially more on Amazon than on Redbubble and that they only see Redbubble as a way to get approved by Merch. This is purely conjectural, and, even if true, having a bunch of artists flock to Redbubble in order to get access to Amazon’s print-on-demand services is not much of a problem, flywheel effect and all.

Ultimately, Amazon and Etsy have reached a critical point of salience in consumers’ minds. Many millions of loyal customers visit them directly with a specific purchase in mind, with the intention to discover something to purchase, or just to browse. Competing for customer mindshare with these giants directly would be futile and foolish of Redbubble. However, Redbubble can successfully compete with them for artists’ mindshare. In fact, focussing on attracting artists first is how Redbubble approached the chicken-or-egg problem of every online marketplace. And, at the risk of belabouring the point, having enough long-tail artist content is vital to attracting customers who want something unique. This is how Redbubble successfully competes with Etsy and Amazon for customers.   

Small businesses with online shops

Large print shops have also developed partnerships and integrations with Shopify, Wix, and other e-commerce platforms to allow individuals and small businesses to set up their online shops and use the print shops’ services fairly seamlessly. As with Etsy sellers, these shops could in theory scale globally using the print shops’ global networks, albeit at a significant cost. Setting up one’s own online shop with these e-commerce platforms could also be costlier than selling on Etsy, but this is not the major competitive disadvantage that these small businesses have when facing Redbubble.

Unlike physical shops, for which location is a major determinant of foot traffic, online shops need to find a way to attract customers’ attention in an ocean of competitors. And it’s usually online marketplaces such as Redbubble rather than search engines that serve as the conduit because of the prohibitively high costs of marketing for a small business. Artist profiles on Redbubble and TeePublic allow artists to feature a link of their own websites, which could be Wix or Shopify sites. Artists can also link their Instagram, Facebook, Pinterest, DeviantArt, and other social media pages to their Redbubble and TeePublic profiles. These pages can then lead customers to the artist’s own online shop (Pinterest Pins) or allow them to shop directly (Instagram Shopping or DeviantArt). So Redbubble can lose sales from successful artists who have enough of a following to migrate to these alternative sales channels, but as long as Redbubble continues to attract new artists, it will win in the aggregate.

What does the risk/reward look like, and why does the opportunity exist?

Section summary: Redbubble’s shares have sold off significantly over the past 4 months (the share price is now ~50% off its all-time high reached in February), first after the company reported its 2Q21 results in February, when the company reported somewhat disappointing gross margins and decelerating sales growth (but it was still 84%). Perhaps more importantly, the newly appointed CEO, Mike Ilczynski, told investors and sell-side analysts that he would focus on scaling the business, which would require investments, and short-term profitability would likely suffer, but he was short on detail. Mr. Ilczynski provided a bit more detail on the 3Q21 results call: he laid out investment plans in somewhat vague terms, set financial goals for “CY24+”, and said “short-term” EBITDA margins would essentially be cut in half from the CY20 company record of 9.5% to “mid-single digit range over an annual period.” The CEO’s plans lacked specifics on the investments, but his goals were made clear – to scale the marketplace by growing the number of participants on both the artist and buyer side, as well as building loyalty/increasing retention rates for both groups. Essentially, Mr. Ilczynski will go for winning mindshare from both customers and artists. The timing of the sell-off suggests the reasons for it were the lack of visibility and the sacrifice of short-term profitability to long-term oriented investments. This means that the recent stock price drop has provided a time arbitrage opportunity. Surely, there are risks ahead, namely capital allocation mismanagement and some emerging legal risks to the print-on-demand industry, both overlaid by the constant struggle with competitors. However, I think the company’s competitive advantages, capital structure and economics, and the optionality embedded in its business model skew the range of the company’s intrinsic value heavily to the upside relative to its current valuation.

Legal Risks 

Redbubble has been hit with multiple lawsuits, in the US and Australia, alleging liability for copyright and trademark infringement stemming from content posted on redbubble.com. The company has mostly been successful in its defence, winning summary judgments or case dismissals on the basis that it is a merely a platform (like Amazon or eBay) not liable for the content that its users share on that platform, invoking section 512 Digital Millennium Copyright Act (DMCA) and Section 230 of the Communications Decency Act (CDA). However, in 2021, Redbubble faced two (potentially three) unfavourable rulings, which are a cause for concern. Before discussing them in more detail, I want to emphasise that I have no legal training.

In 2017, Ohio State University (OSS) brought a case against Redbubble, alleging that Redbubble is liable for infringing OSU’s trademarks by selling products bearing OSU’s trademarks without permission and violated OSU’s right-of-publicity under Ohio state law for selling products bearing the image of former OSU football coach Urban Meyer, who had assigned his publicity right to the school. In April 2019, the Ohio district court granted Redbubble summary judgment in Redbubble’s favour, on the basis that Redbubble is not the “seller” under a legal standard and, hence, doesn’t “use,“ under a legal standard, the trademarks and right-of-publicity in question. OSU appealed, and, on February 25^th, 2021, the Sixth Circuit of the United States Court of Appeals, reversed and remanded the district court’s decision, and the case is back in court for further fact-finding.

In 2018, Atari, funded by a litigation-funding company called Legalist, filed a lawsuit against Redbubble in the Northern District of California, alleging Redbubble is liable for selling trademark-infringing products. On January 21^st, 2021, the court granted in part and denied in part summary judgments in favour of Atari. The case is still ongoing, since the judge found that Atari didn’t properly establish the amount of damages it was seeking, and there are case management conferences scheduled for July.

In 2019, apparel company Brandy Melville filed a lawsuit against Redbubble in the Central District of California, suing Redbubble for trademark infringement. The judge denied summary judgments and brought the case to a jury, which, on June 23^rd, 2021, found that Redbubble wilfully contributorily counterfeit several Brandy Melville trademarks and awarded statutory damages for those infringements of $500,000. The jury also awarded Brandy Melville $20,000, representing “the total amount of Defendant Redbubble’s profit attributable to Redbubble’s contributory trademark infringement.” Redbubble immediately released a statement:

Redbubble believes that certain critical findings were not supported by the evidence offered at trial and will be asking the court for relief from the verdict on that basis. Redbubble remains confident in its position and will continue to vigorously pursue its defence of the claims.

What these three cases have in common is that Redbubble’s argument that, as a platform, it was a mere intermediary in the transaction, which absolved it of liability since it wasn’t the seller of the products, was rebuffed to some degree by judges. In all three cases, the facts that Redbubble arranged for the display, manufacturing, and shipping of products that effectively did not exist prior to a customer making a purchase on redbubble.com, that these were packaged in Redbubble-branded packaging, that Redbubble marketed these as “Redbubble products,” and that Redbubble would be responsible for handling the returns and refunds were brought up by the judges to conclude that Redbubble could be considered the “seller” of the products under a legal standard, which makes Redbubble ineligible for safe harbour under the DMCA and CDA.

If Redbubble is unable to obtain relief from the Brandy Melville case verdict, and if it ultimately loses the OSU and Atari cases, these cases could become dangerous precedents. None of the three cases is likely to directly result in significant financial damage to Redbubble, but a wave of lawsuits encouraged by these precedents could spell the end of the company. I realise this statement implies that Redbubble is awash with infringing content, and, given some of the details revealed in these cases, it’s quite likely that there is a sufficient amount of infringing products on the company’s websites to cause a real headache.

In the Brandy Melville case, Redbubble tried to argue that it took sufficient measures to prevent intellectual property infringement. From the case’s Civil Minutes, July 10^th, 2020, my emphasis:

Redbubble argues that it takes a broad range of steps to both prevent infringement generally and to prevent it against Brandy Melville in particular. These include requiring sellers to confirm that they own or have rights to the content they post for sale, disabling the accounts of users who repeatedly infringe, and employing proprietary software and a thirteen-person marketplace integrity team to evaluate and remove suspect listings.

These measures were clearly not enough to prevent repeat offences in either the Brandy Melville or the OSU case. In both cases, the companies notified Redbubble of infringing content on its website, which ought to have prompted Redbubble to remove offending content. In the Brandy Melville case, Redbubble did just that but apparently failed to prevent further offending content to be uploaded, as Brandy Melville continually found such content and filed the suit. In the OSU case, after receiving a cease-and-desist letter, Redbubble asked OSU to identify the offending items. From the Sixth Circuit’s decision, February 25^th, 2021:

So OSU sent Redbubble a letter containing photos of nine offending items. But Redbubble told OSU that pictures weren’t enough to identify the offending products, asking for URLs or other identifying information. After this, communication halted between the parties. In the end, Redbubble did not remove the offending products from its website. After the communication breakdown, OSU sued Redbubble in December 2017.

The “communication breakdown” strikes me as a sign of complacency – it seems Redbubble was confident enough in its legal position that it didn’t find it necessary to respond, or its small marketplace integrity team was just overwhelmed. Either way, Redbubble seems to have been confident enough in its legal defence to not address its limited ability to effectively police content on its website. Brandy Melville brought to the judge’s attention that the “My Work Was Removed” section of Redbubble’s help page ended with, “… content is only removed when it has been specifically identified as infringing in a legally valid takedown notice. We generally don’t go looking for similar works to remove from the marketplace.” This sentence is no longer on the website.

Redbubble still has legal options, and the ultimate resolution of the cases might take some time, but this is no consolation for a long-term investor. Even if Redbubble ultimately prevails in the three cases, there is always the risk of changing regulation, which could remove its legal defences. Furthermore, as the business grows, in the absence of an effective monitoring system, infringing content will also grow, which poses a risk to the Redbubble brand – it’s not the company’s goal to be perceived as a marketplace for knock-offs. Of course, perfection on this front seems unattainable – even Amazon sells knock-offs – but there is significant room for improvement in Redbubble. A quick search for famous global brands on redbubble.com yields mostly harmless but also some suspicious results (e.g. “Just do it” stickers).

Possible remedies

Outside of favourable court rulings, investments in technology seem to be the only remedy. I’m not a software engineer, but a software solution involving image recognition that would cross-reference a huge database of IP (e.g. USPTO) seems like a daunting, if not impossible, and prohibitively expensive task. Furthermore, Redbubble’s legal troubles stem from the lack of further prevention, after takedown notices had been served, rather than an ex-ante capability (the “communication breakdown” in the OSU case was pure negligence). A system to effectively prevent further infringements of a particular trademark once a notice has been received seems more achievable – there are companies specialising in reverse-image searches, which are used by copyright holders to trawl the web for potential infringements. AWS, of whom Redbubble is a customer, offers image recognition instances and APIs, and so do the other major cloud vendors. These tools are only going to get better with improvements in image recognition AI. Redbubble has also not disabled keyword search for trademarks it has not licensed, which could serve as a useful guideline as to what content to monitor.

With no software or legal expertise, it would be futile for me to try to estimate the costs of such investments. The Tiffany vs. eBay case (in 2010) revealed some interesting details about eBay’s fraud prevention efforts. eBay said it spent $20m a year on fraud prevention on reimbursements for counterfeit products, 200 employees exclusively on combating infringement, automated “fraud engine” software, manual listing, and a take-down system assuring removal of the claimed infringing product within twenty-four hours. This is a very significant expense relative to Redbubble’s opex (A$108m on an LTM basis) and headcount (with around 300 employees in total), but it was made by a company with $9.2bn of sales at the time. eBay also faces tougher challenges, as it sells a much larger variety of products, and image recognition might not suffice. Finally, this example is a decade old, which is a huge amount of time in the context of image recognition software development.

Of course, eBay couldn’t prevent the listing of infringing products ex-ante. What mattered for eBay (and for Amazon and Etsy) is that it was ultimately found not to be the “seller” of the counterfeit goods. The investments described above could make Redbubble more effective in preventing infringing content uploads once served a takedown notice, which could in turn prevent future litigation. However, if the parties claiming infringement decide to litigate without first serving a takedown notice, or litigate anyway, and the courts find that Redbubble is the “seller” of infringing goods and therefore liable, as a famous IP lawyer put it when discussing Redbubble’s cases, “… then Redbubble bursts.”  I cannot emphasise enough my lack of legal training.

What I find most striking about these legal issues is that they get almost no attention on earnings calls from management, sell-side analysts, and shareholders alike. According to Redbubble’s FY20 annual report, management does not consider the then-current legal action “likely to have a material adverse effect on the business or financial position of the Company,” and I couldn’t find a conference call transcript where the issue is discussed by any call participant. Management is no doubt better educated on these topics than me, and so are its large shareholders, some of whom have taken very large concentrated bets (relative to their portfolios) in Redbubble. The stock hasn’t sold off immediately following the unfavourable outcomes discussed above, so there could be some crowd wisdom to take note of here. But investment is ultimately about making one’s own conclusions, and I consider this to be the biggest risk to Redbubble’s business by some margin. The company’s IR department, to my pleasant surprise, responded very quickly to my questions but, as expected, only highlighted Redbubble’s commitment to working with rights holders to protect their IP and to going above and beyond legal requirements for online marketplaces in its various jurisdictions.  If I’m wrong about the severity of the risk, all the better for the investment case.  

The risk of mismanaging capital allocation in a period of significant investment

This risk overlaps with the assessment of management, and in Redbubble’s case, there’s already been a period of apparent mismanagement. However, this past episode seems to have been a case of underinvestment in initiatives to boost Redbubble’s mindshare among customers and artists, in order to maintain short-term profitability. The new CEO’s plans indicate he understands that investing to grow the number and loyalty of participants from each side of Redbubble’s marketplace is more important to shareholder value creation than short-term profitability. Surely there is a risk of overspending on bad initiatives, but I would be more concerned with the risk of investing too little in good ones for the sake of maintaining short-term margins.

Mr. Ilczynski was brought in in January of this year to replace Barry Newstead, who was appointed as CEO to replace retiring co-founder Martin Hosking in the summer of 2018. Prior to that, Mr. Newstead had spent 5 years as Redbubble’s COO, directing the company’s operations during a period of rapid growth – the company had almost quintupled GTV over the period. Yet, despite this apparent success as COO, his CEO run was disappointing. The board fired him in February 2020, and Martin Hosking came out of retirement to fill the CEO role while the board was head-hunting.

During Mr. Newstead’s short-lived CEO reign, Redbubble’s growth slowed down significantly – the company’s marketplace revenue grew ~41% in FY19, but more than half of that growth was contributed by TeePublic, which was acquired in November, 2018 (2Q19). A big part of the reason was Google’s algorithm change in October, 2019, which caught Redbubble off guard and the search portion of Redbubble’s organic sales channel was affected. But more importantly, the unpaid and paid GTV mix was kept roughly unchanged and on earnings calls, Mr. Newstead continued to emphasise the stable profitability of the marketing spend as measured by the ratio of gross profit to performance marketing. While the company’s growth was languishing, Mr. Newstead appeared to lose focus and launched the Fan Art/Partner Program (which I don’t think was a bad idea but maybe shouldn’t have been high on the priority list at the time) and was even developing strategies for Redbubble to launch a physical retail footprint.

Reading the company’s results, it’s not that obvious that Mr. Newstead mismanaged Redbubble’s capital: the company did not burn cash and margins slightly improved, but the company wasn’t focusing enough of its spending on growing the number of artists and buyers. This interpretation might be biased because it’s made with 2020 in hindsight, and maybe, absent the exogenous shock of the pandemic, the business was just facing a slowdown that no CEO could have remedied. But Martin Hosking became interim CEO at the very start of the pandemic and immediately boosted marketing spending and continued to do so even as it became apparent that lockdowns and mask demand could really drive organic growth and, hence, significantly improve margins. If value creation for Redbubble is synonymous with scale, then Mr. Newstead’s capital allocation error was one of underspending.

The silver lining is that Mr. Hosking showed he looks out for the interest of shareholders, of whom he is the largest. He currently sits on the board and still owns ~16% of the shares (he donated about 2% in October, 2020). Although he is retired, he is still relatively young (61 years old), as demonstrated by his ability to successfully manage the company through the pandemic and could step back in the CEO role if needed.

Mr. Ilczynski’s track record is difficult to assess. From 2006 to 2019, he held various executive roles at SEEK, which operates a number of online job marketplaces and education websites globally, culminating with the position of CEO of one of SEEK’s two business units, which he held for 18 months. SEEK has been quite successful in scaling online marketplaces, and its market cap had grown from about A$900m to A$7.5bn during Mr. Ilczynski’s time there (an 18% CAGR). He certainly has more experience in scaling online marketplaces than Mr. Newstead, but there is no clear track record of his capital allocation decisions. We’ll need to trust Mr. Hosking and the other board members in their choice.

The capital allocation risk with Mr. Ilczynski’s investment plans is that he spends excessively, and there would be plenty of opportunities to do so. For example, Apple’s and Google’s moves towards increasing privacy options for its customers could really limit the effectiveness of targeted advertisement (I’m no expert on digital advertising). The risk of overspending is partially mitigated by the company’s capital structure (no debt), its ability to finance itself through working capital, and what seems to be a “fail fast” mentality by Mr. Ilczynski, who on multiple occasions during this year’s earnings calls stressed that large investment ideas will be tested in small experiments first.

The growth opportunity and embedded optionality

I don’t think the Redbubble business lends itself to detailed bottom-up modelling, so it would be useful to use the financial goals set out in the CEO’s April 2021 letter to shareholders as a reference point. Mr. Ilczynski’s investment plan has an aspirational goal for “CY24+” of A$1.5bn of GTV, A$250m of artist payments and A$1.25bn of marketplace revenue, with 25-30% GPAPA margins and 10-15% EBITDA margins, or EBITDA of A$125m-A$188m. This target, assuming it’s achieved in CY24, implies a revenue CAGR of 24% from CY20, or 19% if achieved in CY25. Either one would mark a significant deceleration from the growth seen over the past twelve months of 82%, 61% excluding masks. Presented differently, this financial goal could be reached if Redbubble customers double (to 18m), the average order value grows at 2% from the historical average of ~A$50 (to $A55), and customer loyalty increases to bring the average annual number of orders per customer to 1.5 from the historical average of 1.3. The targeted margin structure doesn’t envision much improvement in GPAPA margins and, hence, operating leverage, relative to CY20 despite the targeted sales growth and customer loyalty improvements. This seems quite conservative and very much achievable if the revenue target is reached.

To assess how plausible achieving the revenue target is, one can try to establish some base rates/reference points. Redbubble.com and TeePublic.com had about 236 million visits over the past 6 months. The traffic to the US domains of Zazzle, TeeSpring, Society6, and Spreadshirt over the period was approximately 242 million visits combined. Assuming similar conversion rates (about 2%), there was the same number of purchases from just four of Redbubble’s largest competitors (more if one factors in country-specific domains in Europe). The purchases wouldn’t represent unique customers as some customers could have purchased on multiple sites and/or could have made multiple purchases per site, but every purchase made on competitors’ websites is a potential new or repeat customer for Redbubble. Moreover these four competitors clearly do not command 100% of the market for printed products (just think of Amazon), and the above-mentioned number of visits came almost exclusively from the US. So there is plenty of room for share gains.

Furthermore, I think the overall market for Redbubble’s and its competitors’ print-on-demand products will continue to grow. The narrative that demand for these products is driven by people’s desire to be different is fairly clear. Custom-printed products have been around for a while, but seamlessly accessible online aggregators of designs from hundreds of thousands or millions of artists available to be printed on hundreds of products have not. Demand for these products isn’t new, but the technology on the supply side is. Even if it’s a mere shift from brick-and-mortar retail to e-commerce, there is still a long runway ahead, especially because there will be more “canvases” to be added on the supply side.

The pandemic created the conditions for print-on-demand technology to truly shine in the market for face-masks, which exploded overnight and was a huge sales and profit booster for print-on-demand companies. However, the mask market was ideal for print-on-demand because of its extreme homogeneity – prior to the pandemic, face-masks were fairly bland, and the physical dimensions of the design can’t really be varied much, so sawing masks out of colourful fabrics or printing designs on them were really the only options for differentiation. Relative to masks, there is huge variation in the designs of homeware and clothing, further differentiated by brand, material etc., hence plenty of opportunity for expression. Yet the release of new items available for print on Redbubble’s sites has always resulted in incremental sales, so wacky pictures, quotes etc. are a desired further form of differentiation.

Although Redbubble offers more than a hundred different items, there are plenty more opportunities. For example, according to the ex-CTO of Kornit, the leading manufacturer of digital printers, moisture-wicking athletic apparel hadn’t been viable for print-on-demand due to the difficulty of digital printing on polyester. 4imprint, one of the leading suppliers of promotional products and corporate swag, offers a much wider variety of products (thousands of them) than Redbubble. Some of these certainly won’t have much success with consumers (e.g. tape measures), but there are plenty that will resonate with consumers (e.g. beer mugs, basketballs). 4imprint and its B2B peers have minimum order quantities, but if volume economics or printing quality is hindering the introduction of certain items, I’d bet it’s only a matter of time before printing technology makes them viable for single item orders or achieves acceptable printing quality.

Finally, I see two potential “options” embedded in the business. The aggregation of large numbers of artists and customers on one platform could create other ones that I haven’t thought of as well. First comes the Redbubble Partner Program, which has yet to make a meaningful contribution to the business although it was started in 2019. Redbubble has reached licensing agreements for 60 brands, primarily for TV shows. Artists can create designs using the licensed content and submit it for approval by the brands’ IP owners. The financial terms of the deals are not disclosed, but the brands’ owners receive a royalty on products sold. They also receive designs and an audience of millions of customers or potential customers for free. They are getting paid to receive creative content, which will then be used for essentially marketing their brands. The artist gets exposure to potential employers and also increases his or her chances of selling designs on redbubble.com as his or her design is attached to a popular search item. Redbubble faces the royalty costs but also gets organic traffic and sales from fans searching for their favourite movies, music artists etc. There are many brands out there for Redbubble to license, and the addition of brands can enhance the flywheel effect. Many probably wouldn’t want their brand associated with all the other brands and content on Redbubble (Redbubble can do a better job of removing sexually explicit content), but clearly some don’t mind.

The second option is a variation of the Partner Program, but with product (mainly apparel) brands. Today Redbubble’s fulfilment partners only use non-branded “blank items.” Promotional products companies like 4imprint mostly use blanks as well, but also have licensing agreements with apparel brands, on whose products they can print their customers’ designs. I think the demand for, say, a Nike shirt with a drawing of an otter by fans of both Nike and otters is self-explanatory. Reaching an agreement on pricing so that both the brands and Redbubble are satisfied with the margin might be difficult, but there are precedents of success in the corporate swag industry. Large apparel companies might also not want to have their brands associated with some of the content on Redbubble, even if they set implementable guidelines on what can be printed on their products.

In summary, I think Redbubble has plenty of room for growth. If Mr. Ilczynski’s investment plans succeed in leveraging and enhancing Redbubble’s competitive advantages described in the competition section, the company’s CY24+ aspiration is very much within reach and could easily be surpassed.

So what’s Redbubble worth?

Gun to my head, I would say A$1.5bn-A$3bn.

Redbubble has managed to finance itself without major equity raises since its IPO, with the exception of the equity issuance to finance the TeePublic acquisition, and is debt-free, hence no interest expense. The recognition of revenue upon delivery rather than payment and the company’s negative working capital cycle contribute significant cash inflows not reflected in the income statement during periods of rapid GTV growth. Capex mostly goes towards intangible assets (software) and has averaged 4% of marketplace revenue since the company’s IPO. Net, in periods of rapid growth, EBITDA will likely understate FCF, as was the case in both fiscal and calendar 2020. Therefore, even when factoring in the full Australian corporate tax rate, the target CY24+ EBITDA will likely understate FCF, unless Mr. Ilczynski’s investment plans include a significant increase in capex relative to revenue. To be on the safe side, we can assume that EBITDA will be a good proxy for FCF. This means the CEO is aiming for A$125m-A$188m of FCF in CY24+.

If Redbubble manages to deliver this result in CY24/CY25, an investor who wants an FCF yield of 7-10% on his or her investments will be willing to pay 10x-14x FCF (capitalizing at 7-10%) for the company, or an EV range of A$1.25bn-A$2.6bn (applying 10x to A$125m and 14x to A$188m). Discounting this future value back at 7-10%, a very reasonable if not ambitious rate in the current environment, suggests a present value range of ~A$850m-A$2bn if discounted back 4 years (discounting $A1.25bn at 10% and A$2.6bn at 7%). If Redbubble achieves its CY24+ aspirations, it would have grown sales at a CAGR of 19%-24% (depending on when A$1.25bn is achieved) since CY20, so, in reality, the company will be expected to continue to grow, and these multiples will be extremely low as they don’t embed any growth expectations; it would likely trade north of 20x FCF (or EBITDA as a proxy in this case). Etsy’s EV/consensus NTM EBITDA has averaged 35x over the past three years, while Redbubble’s has averaged 38.5x. So the multiple could be north of 30x. Repeating the above calculation with multiples of 25x-30x yields a range of present values of A$2.1bn-A$4.3bn. Redbubble’s current EV of A$936m suggests the market’s odds of Redbubble living up to its CY24+ aspirations are quite long.

I think Redbubble faces better than even odds that it will reach its CY24+ aspirations. In my view, a valuation of 25x-30x that aspirational FCF would be reasonable, pointing to a present value range of A$2.1bn-A$4.3bn. Assuming a 2/3 probability for this scenario suggests an expected value of Redbubble’s EV of A$1.4bn-A$2.9bn, or an MV of ~A$1.5bn-A$3bn, after factoring in the company’s net cash position of A$120m. The other 1/3 I’d give to the scenario where the legal risks described earlier unfold in the worst possible way and render the company worthless. There is no obvious catalyst for the first scenario. The company just needs to execute to a point where investors start believing the CY24+ goals can be easily achieved and start betting that way. It’s unlikely it will take four years for the verdict to come in. Speaking of verdicts, the obvious catalyst for the doomsday legal scenario would be just that, a few or only one milestone court decision that would determine that Redbubble is not merely a platform but the “seller” under a legal standard and, therefore, liable for IP infringement. If the cost of a solution to deal with infringing content is too high or Redbubble is inundated with litigation before it can implement it, the company will probably be worthless. I think this is an extreme and improbable, but not impossible outcome, and I only ascribe it such probability to illustrate the asymmetry I see in the bet. I don’t know enough about digital marketing to determine whether Redbubble’s websites have value to provide some margin of safety by virtue of their traffic alone. 

Redbubble appears attractive on a relative basis as well. It currently trades at a large discount to Etsy and 4imprint, two adequate comps in the absence of pure competitors. Both Etsy and 4imprint trade at 37x EV/consensus NTM EBITDA, while Redbubble trades at 19x, significantly below its 3-year average of 39x. GPAPA is a key metric for not only Redbubble but also 4imprint and Etsy. The valuation gap on GPAPA basis is even larger: Etsy trades on 34x EV/CY20 GPAPA basis (I don’t have consensus GPAPA estimates), 4imprint trades on 15x, and Redbubble trades on just 6x.

Etsy’s premium is justified, as it has more than doubled revenues in 2020 off an already high base relative to Redbubble, it has consistently delivered high returns on capital, and the company seems to have further room for growth, all a function of its success in gaining consumer and seller mindshare. 4imprint is a B2B company, but it shares many similarities with Redbubble: 4imprint is effectively an online marketplace connecting businesses who want to purchase various items bearing their logo and printing companies. 4imprint does have some internal printing and sawing capacity, but it mostly drop-ships products to customers through its large fulfiller network. As a result, it also has nil or negative working capital investments and negligible capex requirements. It operates with lower gross margins and higher marketing expenses relative to sales, so it has lower GPAPA margins – 15% on average during FY14-FY19 compared with 21% for Redbubble during the same period. 4imprint has a lower opex base, and the net result is an average EBITDA margin of 7% during the FY14-FY19 period, or slightly above what Redbubble’s CEO guided in the short-term. 4imprint is also essentially debt-free but has a lower cash position ($40m). There is also another interesting similarity between 4imprint and Redbubble – 4imprint has an aspiration of $1bn in revenue, or A$1.3bn, essentially Redbubble’s CY24+ marketplace revenue target, though 4imprint hasn’t shared its margin targets or a timeframe.

There are also notable differences between the two companies. 4imprint’s working capital inflows are not as large relative to revenue to offset capex, and it pays a higher tax rate, so its EBITDA understates FCF. Another difference from Redbubble is that 2020 was a very challenging year for 4imprint and the entire industry – business travel and conferences where promotional products are exchanged came to a halt and 4imprint’s sales fell 35%, which was partially offset by masks. 4imprint barely broke even in 2020, and the future of the promotional materials industry has rarely been more uncertain. Yet 4imprint rerated up, while Redbubble rerated down. 4imprint does have a longer track record of profitability, but investing is about the future. I’m not suggesting that Redbubble is undervalued on the basis of 4imprint’s valuation – maybe 4imprint is overvalued – just providing another example of how low the probability of success of Redbubble’s plans is, as implied by the current market valuation, relative to another available investment.

Ultimately, at the current EV, I’m willing to underwrite Redbubble’s risk/reward proposition. I realise that I could lay out many more outcomes in some sort of probability distribution, and I haven’t even included the optionality I described earlier in the valuation, but I don’t think those exercises would illuminate any further my view that the risk/reward is skewed to the upside.  

Martin A.