Firehose #154: 😴 Sleep no more. 😴

A constructive look at Casper's IPO, games eat everything, the reproducibility crisis, and more!

One Big Thought


Casper ($CSPR) and One Medical Group ($ONEM) both filed to go public in the last week or so. The "cashtags" in the prior sentence link to Twitter discussions on both companies. Both companies have many happy customers (including me), but, as you can see, the conversation around CSPR in particular has become a bit caustic. 

I chose to stay out of the fray on Twitter. Much of the discussion was unhelpful and mean-spirited. That's unfortunate because S-1 filings illuminate some of the most interesting questions on business model, finance, and strategy. Respectful, constructive conversations around them are useful for the entire operating and investment community.

Thankfully, a few tweets like this one from @post_market really summed up the situation with CSPR:

"So while we debate the merits of $CSPR and its peers. The reality is that they provide a strong customer value proposition, but its more difficult to say if they will ever capture the value that they destroyed for the incumbents"

CSPR gets credit for creating a category. Its innovative product offering ("bed in a box") reduced an overly complex legacy shopping experience to a radically simple e-commerce purchase. Incumbents took it on the chin in response.

This idea was so good that it quickly attracted reasonable alternatives into the market. Many of these businesses benefited from the customer education that CSPR provided with its aggressive marketing dollars. Customers can now shop around between providers and, while CSPR still leads the market it created by some measures, it is by no means alone. In fact, as @post_market points out in another tweet, its competitor Purple ($PRPL) is already public, growing at double the rate with positive free cash flow and more revenue than CSPR.

Because of these competitive dynamics, the "mattress space" has become a punching bag for skeptics of vertical brands in general. A typical playbook for one of these companies may be to (1) launch a "hero" product that provides branded product innovation, (2) at high gross margins, (3) enabled by an owned & operated e-commerce experience, (4) with high marketing efficiency, usually driven by viral or word-of-mouth campaigns. Success in #1-4 for several years creates the foundation for the company in question to (5) upsell customers into other product adjacencies, thereby growing LTV in a march towards category dominance.

The theory goes, if you can get to #5 quickly, competitors will need to start with a lower LTV, since they've only done #1-4, and you can therefore outspend them for some period of time to drive them out of the market. A larger product assortment also gives customers more potential entry points into your brand, providing a marketing advantage over competition. That's why vertical brands with high margins, high marketing efficiency, and extensible product offerings tend to consolidate markets.

CSPR seems to have gotten stuck in the middle of this process. Competition intensified before CSPR was truly ready to roll up the "sleep economy." The S-1 states:

From 2017 through September 30, 2019, while growing our e-commerce channel, we have maintained 'first purchase profitable' e-commerce economics. 

My "between the lines" read of this sentence is that contribution profit on first purchase is just greater than cost of customer acquisition. A rough approximation of this ratio is the 1.4x spread between gross profit and sales & marketing cost in the last 9 months.

In addition, cohort value is likely a slow build for CSPR. The company doesn't disclose cohort data, except for one mention in the below paragraph: 

From Casper's beginning through September 30, 2019, we have seen more than 16% of customers who have purchased at least once through our direct-to-consumer channel return to purchase another product. Importantly, 14% of our customers returned within a year of their original purchase (excluding those customers whose original purchase date was less than one year prior to September 30, 2019 and had not made a repeat purchase as of that date). Further, 20% of customers in our direct-to-consumer channel during the first nine months of 2019 were repeat customers.

Let's say 14% of customers repurchase in a year and all of them show up to buy every year for 3 years in total (an aggressive assumption). That's only 1.3x purchases on average in that cohort over a 3-year period. In addition, subsequent purchases are likely to come from lower AOV (perhaps lower margin) items like pillows and sheets, which will do less to move the needle on LTV than the original mattress purchase used to acquire the customer. A quick SecondMeasure query confirms this to some degree with 6% average annual dollar retention. Most of CSPR's contribution profit, therefore, appears to come from each customer's first  transaction.

The lack of additional cash flows from customers creates a drag on overall profitability because, without them, operating leverage is difficult to achieve. Every cash flow dollar must instead come from customers acquired in the current period vs. those from prior periods.

We can think of operating leverage as the difference between gross profits after marketing and the other categories of OpEx (in this case, G&A). For the 9 months ended Sept 30th, the former grew 82% from $23M (2018) to $42M (2019). Over the same period, the latter grew 20% from $88M (2018) to $106M (2019). While the former is growing faster than the latter (good!), the absolute dollar gains in one offsets the other almost exactly. In other words, CSPR added $19M of "operating leverage" in 2019, but its increased G&A consumed it entirely.

With the benefit of hindsight, CSPR could have grown more efficiently if it had been able to attach more product to customers earlier in its cohort cycle. I'm sure the company knew this. It made a choice to instead double down on what was working: shipping mattresses hand over fist. 

These short/long-term tradeoffs are really hard choices for operators. For context, CSPR was one of the fastest growing e-commerce businesses around its founding. When a product is flying off the shelves, all you can do is scramble to sell more of it. Moreover, so much of the CSPR story early on was the simplicity of selling a single mattress. I can see the team asking themselves the question, "Is it worth the brand confusion and distraction from something we know is working to push lower dollar value items like sheets and pillows? Can't we wait and solve LTV issues at scale?"

The lesson from CSPR may be to continuously ask yourself whatever version of this question is relevant to your business. Ultimately, success at scale in any consumer product company means owning the category and establishing the default shopping behavior for it. Sacrificing some growth in the short-term can be worth it, if doing so builds a long-term moat.


Tweet of the Week

Links I Enjoy


The doctor will see you now. →

One Medical Group revealed some interesting data in its S-1 filing. It has 77 physical offices, including some on-site at employers. Consumer retention is high at 89%, and business customer retention is even higher at 97%. "Care margin" (essentially gross margin excluding D&A) is 40% in the most recent period.

Member growth is relatively linear at 4-5% per quarter, consistent with a model that requires the deployment of retail "boxes," although the company has diversified away from pure consumers to businesses and health networks.

A business case study shows 8% annual savings due to fewer urgent and specialty care visits. Membership fees are only ~20% of revenue base. Most of revenue comes from billing insurance for services rendered.


Games eat everything. →

Media blogger Matthew Ball wrote a very thoughtful piece arguing that TV has peaked and that gaming will eat its lunch.

My favorite point he makes is #3: "Gaming has unprecedented content leverage." Because of content replayability and multiplayer games' social use case, the hours spent per dollar of content development are simply unachievable by traditional media.


Silicon Valley timeline. →

I enjoyed flipping through this timeline of innovation in Silicon Valley, which goes back to the late 1800's. A few of my fav's: first radio station in San Jose (1909), Bill Hewlett and David Packard start HP and sell their first audio-oscillator to Disney for the production of "Fantasia" (1939), IBM's San Jose labs invent the hard disk-drive (1956).


The reproducibility crisis.

One of the most boring, but most consequential controversies in science is the fact that most published experimental results can't be reproduced by others.

Less than 20% of researchers are ever contacted about having their work reproduced. Reproducibility rates across various fields are estimated between 10-40%. Yet most researchers think that at least half the papers in their field can be trusted.


One shot secrets.

Having young kids means that every movie theater trip is rare and precious. This behind-the-scenes video on how they made 1917 look like one continuous shot may be the closest I get to seeing it in theaters!

Enjoyed this newsletter?

Getting Drinking from the Firehose in your inbox via Substack is easy. Click below to subscribe:

Disclaimer: * indicates a Lightspeed portfolio company, or other company in which I have economic interest. I also own stock directly in AAPL, ADBE, AMZN, CRM, FB, FTCH, GOOG/GOOGL, NFLX, SHOP, SNAP, SPOT, SQ, and TWLO.

Header image credit: