Firehose #158: 🍰 Capital stack. 🍰

Debt for startups, cheap-fil-A, YouTube's big reveal, Bernoulli's blunder, and much more!

Image result for layer cake

One Big Thought

Alex Danco wrote a provocative article this week on the changing capital stack of venture-backed startups. His summary view is that founders of businesses with contractual, recurring revenue are diluting themselves unnecessarily with successive equity rounds. Instead, he predicts that, in the near future, startups will use debt capital to finance the acquisition of these relatively stable cash flows.

David Haber subtweeted the post with a good summary of the key question it raises:

Selling equity is, indeed, an expensive way to finance a company’s growth.

The below is a list of software companies that went public in 2019, how much equity they raised prior to IPO (according to Pitchbook), and the disclosed equity stake of all named founders in the companies’ respective S-1 filings:

  • Cloudflare (NET): $404m raised, 27% founder-owned at IPO

  • Livongo (LVGO): $237m raised, 12% founder-owned at IPO

  • Medallia (MDLA): $338m raised, 30% founder-owned at IPO

  • Slack (WORK): $1.2b raised, 12% founder-owned at IPO

  • Fastly (FSLY): $220m raised, 16% founder-owned at IPO

  • Zoom (ZM): $161m raised, 22% founder-owned at IPO

  • PagerDuty (PD): $173m raised, 7% founder-owned at IPO

These are all world-class companies, and yet, on a dollar-weighted average basis, founder ownership in this group is only 17% at IPO. Employees may hold an additional 10-20% of the cap table in stock options or RSUs, but, in nearly all cases, investors own a majority of the business at IPO.

Debt may seem like an attractive solution to excessive dilution — more so when companies have built up cohorts of reliable historical cash flows from customers. For example, PagerDuty (PD) had nearly 10,000 customers at IPO and a dollar-based retention of 135%. Its cohort chart looks like a beautiful layer cake.


Alex goes on to quote Jonathan Hsu of Tribe Capital, who sees opportunity for debt capital in such a layer cake chart:

When you acquire some customers and they start yielding revenue that behavior sounds an awful lot like buying a fixed income instrument and there is a lot of sophistication around how to value those cash flows. In some sense, what we’ve seen over the last decade is that software enables a whole new business model – recurring revenue – which is both good for customers and is good for investors. It’s good for investors because it becomes more “predictable” in the sense that it starts to look more like a fixed income yielding asset and thus more amenable to traditional financial techniques and thus potentially “in scope” for a wider set of investors.

In the case of PD, the equivalent annual yield is something like 35%. Even “mezzanine” debt at a 12% cost of capital would leave an ample surplus. Further, if this “SaaS-debt” could be bundled together across a mixture of companies with different customers and securitized against their respective contracts, companies could lower the cost of debt capital even further through diversification.

And yet, even writing these paragraphs makes me nervous. The wonderful thing about equity is that it aligns incentives. Debt tends to do the opposite. Debt holders get paid even if the business is floundering. They do have an incentive to keep the company alive, so that they can continue to earn a yield, but they will also be the first ones out the door when something dramatic changes.

Startups need a large margin of safety when utilizing debt. Productive uses of debt include lines of credit that are asset-backed, non-recourse to the parent company. If a set of assets (e.g. receivables, inventory, etc) go bad, the debt providers can only lay claim to those assets, and not the company’s IP, other assets, or anything else. The cost of capital tends to be higher for these facilities, but you’ll sleep better at night knowing the damage can only be minimal if the world changes.

Perhaps that framework can also be extended to the receivables on a SaaS contract. However, I worry that companies tend to overweight the stickiness of their own products in the long-term, and underweight potential disruption by a new competitor. High net dollar retention for your expensive SaaS product is cool until, for example, an open source upstart commoditizes your product and forces you to drop prices. Some businesses are more defensible than others, but no revenue stream in software is forever static and dependable. Gavin Baker made this point clearly:

I suspect we shall see debt used increasingly in venture-backed businesses. My hope is that operators will constrain its use to areas where debt is a more natural fit. Personally, I’m not there yet on its application to SaaS revenue.


Tweet of the Week

Links I Enjoy


Profitable poultry. →

Chick-fil-A sells over $10 billion of product each year, putting it just behind Starbucks and McDonalds as the country’s 3rd largest restaurant chain. Part of its success is due to its unique franchise model. While most fast food chains require owners to pony up $1-2 million in capital per location (including franchise fees and up-front opening costs), Chick-fil-A starts its partners at only a $10K fee.

This lower entry point allows Chick-fil-A to select from a broader swarth of potential owners, which it whittles down through an intensive screening process.

In lieu of wealthy investors, Chick-fil-A selects franchisees who are involved in their local communities. The company’s aim, says a spokesperson, is to find people who are willing to be “highly involved” in day-to-day operations.

That position in local communities drives regular traffic into the stores, which gross an industry leading $4.2 million each. However, Chick-fil-A charges a much higher rake of gross revenue (15% vs. 4-8% for its competitors) and also takes half of the franchisees’s profits. In exchange, the company finances much of the up-front costs. In this way, Chick-fil-A employs a more capital intensive business model, but in doing so unlocks more effective business operators in its network.


The big reveal. →

Tech journalists have long speculated how big a business YouTube is inside of the Alphabet family. On this week’s earnings call, the company revealed that YouTube booked $15 billion in 2019 revenues, or ~10% of the overall Alphabet portfolio. Moreover, that amount does not include YouTube Premium, for which it disclosed 20 million subscribers, or YouTube TV, for which it disclosed 2 million. At $12/month for the former and $50/month for the latter, these other offerings contribute over $4 billion of additional revenue.


All aboard the neural nets (video). →

The Lumière brothers pioneered the art of the motion picture in France the late 1800’s. Their famous film “L'Arrivée d'un Train en Gare de la Ciotat” reportedly startled audience members, who ran for the exits thinking the train was actually running into the movie theater.

Recently a YouTuber named Denis Shiryaev used neural networks to “restore” old videos like this one. If you’ve ever seen “L’Arrivée”, you’ll get a kick of watching it now in 4K resolution!


A theory of lift. →

I am sitting on an airplane right now reading about how no one actually knows why airplanes fly. (Eeeek!)

I thought Bernoulli’s original explanation was simple and sufficient, but apparently now scientists have identified holes in that theory. While we know the equations of laminar airflow around an airfoil, those equations are not necessarily explanations.


A synthesizer

MIDI 2.0. →

Your music may sound different in coming years thanks to the first update to the MIDI protocol since 1983!

The biggest development is the expansion from 7-bit values to 32-bit values. Mike Kent, one of the technical leaders in creating MIDI 2.0, says this is like going from the resolution of a 1980s television to the high-def televisions of today. It means that instead of 128 steps for features like volume, there will now be billions. An area where producers think this might be particularly helpful is allowing for subtle “pitch bend” (see the video below) and controlling how much bass and treble are emphasized in every note.

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: