Retail Revolution vs. Evolution: Part I
Why Some DNVBs Are Opening Stores Faster Than Others
This is Part I of the final output of a Spring Semester Independent Project, tackling the role of vertical retail in scaling Digitally Native Vertical Brands (DNVBs). I set out hoping to answer questions like: “Why is the rate of retail expansion so different, even for DNVBs at a similar stage?”“Why are some stores guideshops, while others hold inventory?” “How are stores performing overall?”
I am in the second year of my MBA at Harvard Business School, working on this project under the advisement of Finance faculty member Bob White. In completing this work, I interviewed operators, investors and real estate developers who have built, financed and / or negotiated leases with brands including Allbirds, Casper, Chubbies, Harry’s, Peloton, and Warby Parker. My research culminates with the following exploration of the role of retail for DNVBs in the business of big-ticket / durable products versus softline / consumable products, plus an in-depth case study on Casper with some concluding thoughts.
What is a Digitally Native Vertical Brand (DNVB)?
- DNVBs are born on the internet and targeted at digital natives
- DNVBs provide superior experience via customer data and end-to-end brand control
- DNVBs’ core assets are their brands; they should be valued like retailers, not tech companies
- DNVBs will generally be unprofitable until consumers catch up to the reality that they are selling
- DNVBs are fundamentally attractive due to their use of vertical commerce (gross margins of 65%) not e-commerce (gross margins of 30%)
- DNVBs are born digitally, but need not remain online-only
This final idea is the topic of my research. DNVBs now account for over 600 vertical retail locations across the United States, larger than the footprint of Whole Foods. Retail experimentation has evolved aggressively from the early days of brands maintaining a single pop-up shop on each coast. As evidenced below, DNVBs have come to embrace locations across red and blue states, coastal and non-coastal cities, in and out of malls.
Big-Ticket Durables vs. Softlines and Consumables
A closer look reveals that the majority of this retail expansion is being driven by a specific sub-set of DNVBs who share two attributes. Brands with products that are big-ticket (median price of a typical purchase is ~$995) and durable (built to last and intended as a one-time or very infrequent purchase) have, on average, opened ten times the number of stores than that of DNVBs operating in softline and/or consumable categories, which have a median per-item price point of just $95. A sample of ten brands from each group illustrates this key difference in retail strategy.
Big-Ticket Durable Brands
Softline and Consumable Brands
Some of this discrepancy can be explained purely through consumer intuition. Big-ticket, durable items are highly considered purchases; it is no surprise that consumers would prefer to touch and feel these products before making any big decisions. Many purchases of this type can also be enhanced through expert consultation. Even though at-home try-on options are available, some consumers would still feel more comfortable interacting with a brand associate when having their orthodontics fitted (Candid, Smile Direct Club) or their bespoke suit measured (Indochino, Knot Standard), explaining the rise of the non-inventory holding “guideshop” model that connect consumers with services, then ship products to home.
Even so, my research suggests that the real driving force behind the frenzied pace of retail expansion has less to do with customer experience, and more to do with customer acquisition costs relative to lifetime value. Exploring this relationship begins with a primer on customer economics.
Customer Economics: A Primer
The success of a direct-to-consumer (DTC) business model depends on a company’s Lifetime Value to Customer Acquisition Cost ratio (LTV:CAC). This relationship summarizes the total business value created by a dollar of marketing spend. Total business value can be thought of as the amount of profit that a customer generates for the company after the customer has been acquired. Ignoring time value of money discounting, a single customer’s contribution can be expressed as:
t = the number of months the customer has a relationship with the company
n = the number of times per month the customer makes a purchase
P = the average amount that the customer spends on each purchase
m = the profit contribution of a given purchase, based on margin structure
Conventional wisdom in the Software-as-a-Service (SaaS) industry, which popularized customer lifetime economics, is that companies should aspire to an LTV:CAC ratio of 3.0. For these types of businesses, n (the number of times per month that a customer makes a purchase with the company) is typically fixed, as the customer is billed according to a monthly subscription plan. However, SaaS companies still have several other, important levers to pull to optimize this ratio:
Increase t: By making their service “sticky,” and providing a superior product versus competitors, SaaS businesses can grow LTV by keeping their customers for as long as possible and thus billing them more months in total.
Increase P: By increasing prices for the same service, or convincing customers to sign up for incremental services, SaaS businesses can increase the amount that the customers regularly spend during each monthly billing cycle.
Increase m: By improving product margins, SaaS companies can ensure that a higher proportion of revenue is realized as profit. Because SaaS companies have largely fixed cost structures, margins improve as the company scales and spreads its development costs over a larger base.
Minimize CAC: Finally, on the other side of the LTV:CAC equation, minimizing customer acquisition costs also improves the ratio. Pulling this lever means finding equally effective, yet cheaper, means of reaching customers (such that marketing spend is reduced, but total number of customers acquired stays constant), or maintaining marketing spend but allocating it towards more efficient channels (such that marketing spend is held constant, but total number of customers acquired increases). Attempting to reduce CAC by cutting marketing spend will not improve the LTV:CAC ratio, if it simply results in fewer customers being acquired.
Customer Economics in the Context of DNVBs
The set of levers described above is specific to the SaaS industry. While the economics of a DNVB are also governed by the LTV:CAC relationship, most DNVBs opening retail locations have non-subscription business models, and thus the following set of levers available:
Increase t: Like SaaS companies, DNVBs are incented to remain relevant to a customer for as long as possible, in order maximize their share of wallet when the customer spends in their given product category. However, unlike subscription businesses, many DNVBs are not guaranteed regular subscription revenue during the length of their relationship with a given customer.
Increase n: Unlike SaaS companies, which typically can’t increase purchase occasions above and beyond their monthly billing cadence, DNVBs will seek to increase purchase velocity via marketing. For example, makeup brand Glossier might attempt to move a consumer from quarterly to monthly beauty purchases, by introducing her to exciting new product formats.
Increase P: Similar to SaaS companies, DNVBs can increase what consumers spend by raising prices or growing basket sizes with a push into adjacent categories (for example, luggage brand Away now offers dopp kits, garment bags and packing cubes). However, these behavioral changes are not guaranteed to persist in the future, as consumers are not locked into a subscription plan.
Increase m: DNVBs inherently benefit from the higher margins of their direct-to-consumer relationships, bypassing the cut taken by third-party retailers. However, unlike SaaS players, DNBVs have more variable cost structures, because manufacturing and shipping costs scale somewhat linearly with revenue, curtailing their ability to improve product margins indefinitely.
Minimize CAC: DNVBs will also seek to minimize their customer acquisition costs through the effective allocation of marketing dollars. However, one critical twist is that DNVBs must also continuing marketing to their existing customer base. A customer who purchases a pair of Allbirds, for example, is not guaranteed to buy his next pair of sneakers from the same brand; similarly, the Outdoor Voices shopper has a plethora of options when it comes to replenishing her athletic wardrobe. Thus, marketing costs risk scaling linearly for non-subscription DNVBs, which must aggressively monetize their base while simultaneously acquiring new customers. This optimization is made all the more challenging by the exploding costs of digital acquisition channels, including a 70 percent year-over-year increase in Facebook CPM, and 30 percent year-over-year increase in Google Paid Search CPC.
LTV:CAC for Big-Ticket Durables
DNVBs selling products in softline and/or consumable categories can leverage their ability to increase consumers’ purchase velocities and basket sizes. This behavior can be driven via performance marketing, but ideally, it’s the result of successful, organic brand-building over time.
But what happens when a brand plays in a category where growing customer value over time simply isn’t an option? Most consumers aren’t in the habit of regularly replacing their couch or carry-on luggage, buying a second stationary bike for their home, or getting another set of invisible aligners after they’ve already had their teeth straightened. For brands that play in big-ticket, durable categories, most consumer sales will end up being one-time transactions, with limited potential for future economic value.
What do the LTV:CAC optimization levers look like for this type of brand? Minimizing CAC is the only option available, if purchase frequency, price and margin are all assumed to be relatively fixed. Paradoxically, while these brands should thus care the most about minimizing CAC, they are also most likely to overspend to acquire a customer, lured by the potential of a $1k+ transaction.
These optics attract a multitude of competitors to the space (mattress brand Casper now competes against over 100 bed-in-a-box fast-followers), who further drive up acquisition costs through bidding behaviors on online platforms, motivated by the threat of customers not re-entering the product category for another decade. These incremental players exacerbate what is best summarized as “dis-economies of scale”– while the first cohort of core customers might have been attracted relatively cheaply, future customers will behave differently and be progressively more expensive to acquire.
The remainder of this post dives into operational metrics. All of the data that I’ve used in my analysis has either been publicly reported (including estimates made public by well-regarded media sources), or is publicly available using freemium tools that are able to estimate proprietary site analytics. Two such freemium tools that I will refer to often are SimilarWeb (which provides estimates of monthly site traffic, as well as the upstream source of that traffic) and SEMrush (which provides estimates of the search volume and CPC for various search keywords). When necessary for analysis, I’ve also made my own assumptions; these inputs are denoted in blue text. No data contained therein was shared with me privately. Freemium data is also not guaranteed to be accurate, and my analysis should be treated as an outsider’s exploration based on the best public data available to me.
My methodology starts with understanding how much annual site traffic two DNVBs get, based on SimilarWeb estimates, and then using publicly reported sales data to estimate their online conversion rates, using my own assumptions about average basket size. Then, based on SimilarWeb data about the percentage of site traffic coming from various upstream sources, I examine the relative cost and effectiveness of three marketing channels (Paid Search, Social Media and Referrals) to better understand what is driving CAC.
Traffic: Brand Building vs. Performance Marketing
The top-of-funnel marketing activities that drive sustainable advantage for strong brands make less sense in the short-term world of big-ticket durables. An analysis of the website traffic sources of two well-known DNVBs — Casper, which sells mattresses, and Glossier, which sells makeup and skincare products — reveals how these strategic differences manifest.
Glossier receives, on average, 25% more site traffic per month, despite Casper doing three times more online revenue in 2018 (Glossier 2018 online revenue estimates are based on a reported figure of “over $100 million” in sales, assuming that 95 percent of sales transact on Glossier.com; Casper 2018 online revenue estimates are consistent with a 2018 revenue report of $373 million, assuming 77 percent of sales transact on Casper.com). Per my average basket size assumptions below, the two brands would have implied online conversion rates of 1.3 percent and 9.5 percent, respectively.
Comparing the conversion rate of Casper versus Glossier is somewhat apples-to-oranges, given major differences in product category and basket size. It is more important to understand how each of these brands acquire traffic — and why this leads to dramatically different consumer behaviors, and thus CAC.
Casper over-indexes significantly in paid traffic sources — SEM, referral traffic, display advertising, and Facebook ads — with only about one-third of traffic coming from direct load (consumers who navigate directly to the Casper website due to prior awareness). Glossier, on the other hand, earns nearly half of their traffic from direct load, while simultaneously over-indexing on word-of-mouth and influencer-heavy platforms, like YouTube and Instagram. These findings are immediately consistent with the statistic that re-marketing to a previous customer can be as much as 90 percent cheaper than acquiring a first-time customer — something that Casper must do effectively every time it makes a sale.
Deep dives into specific acquisition channels reveal further sources of Glossier’s CAC advantage relative to Casper.
12.4 percent of traffic to Casper.com originates from Google paid search ads. Two keyword strings, “casper” and “casper mattress,” are estimated to account for nearly 60 percent of this volume, which translates into 1.7 million site visitors per year who clicked on a paid search ad for one of these two terms. The cost of this traffic? Nearly $6.8 million per year, based on cost-per-click estimates of $3.90 and $4.32 respectively for these two terms alone.
Organically, these two terms generate results for 96.5 million unique URLs (“Results in SERP”), making paid advertising necessary to rise to the top. Prices are astronomical because Casper competitors view search advertisements as a way to intercept a consumer in the mattress marketplace who hasn’t yet made a purchase. Four other brands (Nectar Sleep, Tuft & Needle, Leesa, Allswell) can be observed purchasing ads against the term “casper mattress.” Notably, CPC prices are highest for terms that might be used by an undecided shopper, giving brands the opportunity to win a sale (“casper mattress reviews” currently auctions at $10.05 CPC).
In contrast, only 5.9 percent of paid traffic to Glossier.com originates from Google paid search ads. The term “glossier” is the only paid keyword that drives >1% of search traffic and is responsible for approximately 1.3 million visits. At a cost of $0.23 cost-per-click, defending this term costs the brand just $306k per year, and there is virtually no observable competitive bidding.
It can be difficult to disarticulate whether social media platforms are paid or organic acquisition channels. Most sites offer a mix of organic publishing and paid advertising products; furthermore, even seemingly “organic” content is often produced and disseminated by well-paid influencers, who are compensated by the brand for access to their audience. To oversimplify:
- Facebook: An entirely paid platform, where brands earn negligible organic reach (even those who previously built large “communities”).
- YouTube: Pre-roll video ads are paid, while native content can be a mix of organic and influencer-backed. Outside of earning a channel “Follow,” brands don’t have a captive audience on the platform, and are thus reliant on search optimization (including paid ad products) to drive visibility.
- Instagram: Remains a rarity, in the sense that brands can build a meaningful, captive audience that will see most or all published content in their personal feeds. The platform is a free, direct pipeline to engaged fans, and thus investing in building a community still makes sense, for now. Organic reach can also be supplemented with paid advertising products.
Brands like Glossier have benefited enormously from the close consumer relationship offered by Instagram, scaling their 1.9 million-strong community to the point that they no longer find it as necessary to engage paid influencers to boost their reach. For a recent launch, Glossier eschewed their typical influencer outreach strategy, instead choosing to gift the new product to 500 of their most engaged customers, seeding grassroots word-of-mouth.
Casper, on the other hand, is unable to participate in the same way. While Glossier superfans derive enormous value from following the brand — getting information on new products, beauty inspiration, tutorials, and other content to inform future purchases — even a satisfied Casper purchaser might see limited value in following the brand on Instagram, given that they won’t have a near-term future opportunity to engage (i.e. buy product). This fundamental difference is reflected in relative community size. Notably, sub-brand Glossier Play, launched in March 2019, accumulated in under a month nearly the same number of followers (120k) as Casper (148k) on Instagram.
Casper is much more reliant on Facebook to drive social traffic, and doesn’t earn meaningful volume from Instagram. Glossier skews more heavily towards YouTube, with Instagram a meaningful part of the mix.
A significant portion of Casper’s brand equity comes from the fact that the company purported to usher in a new era of transparency in mattress shopping, rendering obsolete the commission-seeking store salesperson who would falsely recommend the mattress with the best profit margin. But, as many journalists have already pointed out, the rise of the bed-in-a-box economy has given rise to a new type of “salesman, ” specifically the mattress review website.
These sites profit from affiliate fees, effectively commissions earned when a review site reader clicks a link that directs to a mattress site, and subsequently makes a purchase. Per Recode, direct-to-consumer mattresses are uniquely attractive for this type of model, as “a high-priced item that results in a large commission, coupled with a heavy consumer reliance on reviews, since many of these new mattress brands are not widely sold in physical stores.”
The economics of the affiliate cut can vary significantly from site to site and brand to brand, depending on the importance of the relationship. And while these sites claim to be unbiased, powerful incentives might still shape the contents of reviews. MemoryFoamTalk, one such site, recommends brand Nectar as their top choice for mattress buyers; site owner Andrew Levy has gone on record revealing that Nectar pays his site $150 per successful referral, more than three times the industry going rate of around $50 per sale. Other industry insiders have reported commission fees as high as $250 per mattress.
The CAC implications of these affiliate relationships are crippling, but the downside of not actively participating is apparently even steeper. Casper has filed suit against three such sites who recommended other brands over their own without sufficiently forthright disclosures about their commercial relationships with these brands, claiming millions of dollars of lost sales. Today, nearly 6 percent of Casper’s site traffic comes from online referrals, more than it garners from display advertising, Facebook or YouTube. While difficult to estimate what Casper pays for this traffic, a single, high-value affiliate relationship can amount to over $1 million per year in referral fees.
Casper seems to have paid a control premium to more deeply embed itself. In 2017, after an acrimonious lawsuit with Tier 1 review site Sleepopolis, Casper provided financial support for JAKK Media LLC (owner of Slumber Sage and Mattress Clarity) to acquire the site. The founding team is no longer involved. Sleepopolis subsequently published an updated and more favorable review of the brand, and offered customers a $75 coupon towards a future Casper purchase. Today, the site is Casper’s fourth most-productive referral channel; it’s estimated that the acquisition cost between $3 and $5 million. Given the importance of favorable reviews in driving referral traffic, Casper’s expense to finance the acquisition could be thought of as a component of their CAC.
Glossier has largely avoided this arms race, mostly by way of an auspicious origin story. The brand’s top upstream referral site, beauty blog Into The Gloss, is an immensely popular content destination that was also the precursor to the spun-out Glossier brand. Thus, 25 percent of Glossier’s total referral traffic, or an estimated 22.0k visitors in March 2019, is driven free of affiliate fees through a wholly-owned subsidiary. For comparison, Sleepopolis drove an estimated 3.8k visitors to Casper.com in the same month.
Implications for Funding Needs and Valuation
Although neither Casper nor Glossier publicly reports annual marketing spend, VC funding provides another window into understanding differences in CAC for these two brands. Founded within a year of one another, Glossier has raised $186 million in total venture funding, to Casper’s $340 million.
Chamath Palihapitiya, former VP of Growth at Facebook, observed that $0.40 of every VC dollar raised tends to go straight to customer acquisition. If this is true for these brands, Casper has spent almost twice the amount of Glossier on acquisition (and acquired far fewer customers in the process, per my earlier assumptions).
This fundamental difference in CAC helps explain why when, coincidentally, both brands led $100 million Series D rounds in March 2019, they were each valued at just over $1 billion — implying a 12.0x Revenue / EV multiple for Glossier, versus just 3.0x for Casper.
Rent Is (A Component Of) The New CAC
The results above begin to clarify the question that I set out to answer: why are DNVBs in the business of big-ticket durables opening stores at a rate of 10x their softline and consumable peers? My answer begins with the fact that these brands are paying more and more to acquire new traffic that converts at a lower and lower rate, due to their fundamental inability to retain previously acquired consumers in the product categories in which they play. If rent is re-imagined as just another input to customer acquisition cost, and DNVBs’ stores remain as highly productive as in the past (see below), a large store base could improve overall LTV:CAC ratios, as revenue from retail increases relative to revenue from e-commerce. This idea will be further tested in a follow-on to this post.
Casper, who recently announced plans to open 200 stores over a three-year period, is the perfect case study to pressure-test this hypothesis with some assumption-driven modeling. You can review my analysis here. It is also worth noting that, despite reporting impressive average retail sales of $1,500 per square foot, Casper pales next to even more productive peers. The leaders, which span hardline and softline categories, include:
- Glossier: $5,300 / sq ft (source)
- Everlane: $4,500 / sq ft (source)
- Away: $4,000 / sq ft (source)
- Warby Parker: $3,000 / sq ft (source)
- Casper: $1,500 / sq ft (source)
These results verify that it is still well-worth the effort of DNVB softlines, like clothing brand Everlane, and consumables, like makeup brand Glossier, to get deeper into the business of retail as well. However, absent the burning platform of a near-breakeven LTV:CAC ratio, as big-ticket durables face, retail buildouts will likely be more incremental in these categories. My research has also led me to the following conclusions:
- Retail is challenging, and not a panacea: Retail operations are difficult and require organizations to build a new set of skills, ranging from site selection to actual execution. There are both direct and hidden costs that stem from a retail strategy. Even if stores offer the promise of an improved LTV:CAC ratio in some categories (still to be further explored), this potential is still highly dependent on successful execution.
- Execution risk depends on store type: Although big-ticket durables have opened stores at a higher rate, many still only operate limited-SKU guideshops that don’t hold inventory. This inherently simple format allows stores to be built smaller and more cheaply. In contrast, the execution risk of opening a full-fledged apparel shop, for instance, is much more significant. Brands must manage far deeper assortments, including correctly forecasting demand across all styles and sizes, and clearing seasonal inventory. While big-ticket durable brands are perhaps opening stores to capitalize on opportunity, it’s also possible that brands who have chosen not to aggressively pursue this path are in fact minimizing risk.
Thanks for reading. I welcome your feedback at email@example.com. The link below directs to Part II of this series, a more in-depth analysis of Casper.