Portfolio Theory In The Wild: Funding YouTube Creators

More life-applied portfolio theory

The Business Breakdowns podcast interviewed Aaron DeBevoise, found of Spotter.

Link (with transcript): https://www.joincolossus.com/episodes/20544486/debevoise-spotter-funding-youtube-creators?tab=transcript

Spotter is a private company that provides knowledge and capital to YouTube creators including MrBeast.

There were 2 great examples of “portfolio theory in the wild”. I will relate them back to where I’ve covered these ideas before and their implication.

1) Optimize the asset

Aaron describes the funding model:

We’re actually not equity for creators. We don’t have a say in their business. There’s no feeling of obligation outside of continuing to do what they do. Loans are really scary for creators because the whole issue with creators is that their business is actually fairly volatile…our approach was, hey, we’ll purchase or license the back catalog, the catalog — the videos that have already been uploaded all the way through the deal and we know their information. It’s almost like postbox office financing for movies. You already know what the box office is and therefore, you can predict what the future might hold and the creator gets to keep 100% of your future videos’ revenues. So if you have 1,000 videos today, we’ll license all 1,000 — the revenue from that, the revenue stream from those videos, the first bit is you upload after the deal, 100% of the revenue goes to you, and we keep 100% of the past. So it’s really not a loan, it’s not equity. It’s more of a cash-flow acquisition financing.

Host Ali Hamed correctly compares it to the music royalty business where investors buy the cashflow an artists’ back catalog. Aaron zeros in on the key differences (emphasis mine):

That’s a great, great comparison. I mean, it’s the first one that always comes out as the music royalty financing business is very similar to this. But there are some really specific differences. So in music, 20 years ago, a lot of the financings that were happening were discounts on future potential cash flows, but those cash flows were fairly determined. It was literally like where are you going to get distributed on radio or on other platforms, but it wasn’t that you were going to go optimize those music assets too much. So it was — we should expect you to make $1 million, we’ll pay $800,000 today. And then the multiple started to expand because the opportunity to optimize those assets, the music royalties, distributing on Spotify and other platforms became so much bigger that, that’s why the multiples grew plus competition. It’s fundamentally different and the optimization of YouTube assets is really around advertising. In our current business, we don’t really buy distribution rights to other platforms. We really focus on the YouTube opportunity, which makes it a really clean deal. So our focus is ad optimization. We can get more dollars out of an individual video than an auction would be delivering. At some point, as we get better at that ad optimization, can pay creators a premium to what they would otherwise make.

The second thing that’s really different, and I think is the most important difference is the motivation behind why people sell royalties to music versus what someone can do on YouTube. So typically, the music royalty licensing business or acquisition business is an exit strategy for most musicians. It’s not really, hey, I’m going to sell $0.5 million worth of my royalty so I can market my next album with a lot more money. You see Sting selling his past catalog for $300 million, and he’s not going to make six more albums out of that. Whereas in YouTube, the ability for that creator to immediately deploy that capital and grow their business is really amazing. And so that’s why we see exponential growth in creators that we’re giving money to. They’re taking the money. And yes, they might buy a house or something. But most of the time, they’re investing in themselves because the growth opportunity is there. So we’ve actually seen deals where we’ll do a deal that’s — the first deal is $4 million where the next deal might be four times the size of that in, like, span of eight months.

Aaron explains how they “optimize the asset”:

The auction itself, like I said, has millions of advertisers. The way that they buy is really based on audience. They think, hey, I have a product that fits males 18 to 34, who like gaming. And then they buy that way. That’s really effective to buy at scale for advertisers, but it doesn’t actually specifically pick out content that is either suitable or aligned with the brand’s initiatives. So he can’t go out and say, hey, I’m going to buy MrBeast because I love what he stands for and his audience has much higher engagement. And therefore, I think I can get more effective click-through rates. What we do is we’re able to say, hey, actually, I know you might be landing on MrBeast, right? Or you might be landing on some other channel, but you don’t know that you are. And we, because we own those assets can actually say, we will guarantee that you land on these assets. And the reason we’ve actually invested in these assets is because they have high level of engagement and high levels of engagement leads to predictable behavior.

It’s not that we went out and just said, here are bunch of underpriced assets that we can now sell at a higher rate. You use a bunch of valuable assets that you’re not realizing that they’re valuable as an advertiser until we tell you. And once you do that, when you put your ads on our content, it’s actually a way more effective ad buy, meaning that in the auction, you probably have to serve two or three ads for every one ad you serve on the content that has high engagement. There was this opportunity to optimize for the advertiser, and therefore, they should be willing to pay you more for those ad units.

Their ability to do this rests on a data advantage which acts as a moat:

First and foremost, it’s a data-driven business. We have 10 years of historical data around all sorts of data points around viewership and monetization that helps us be able to predict the future pretty well. And that has proven to be very challenging to do just if you wanted to start today, it would be really hard. And even we’ve gotten way better at it as we financed more content. So actually financing more content has allowed us to get more accurate, and pay higher prices. So the better you get, the more you can afford to pay because the more accurate you’re going to be and continuing to build that moat where we’re seeing second deals and third deals. The reason we can get those second and third deals to much higher prices is not only because the creator has done a better job or has become more popular. It was because we’ve learned that catalog over time. And so when we look at another video on top of a catalog that we already own, we have a way higher confidence in our role in the performance of those assets.

2) This is a portfolio business

Aaron:

Any one channel can have a downturn and other channels have an upturn…It is a portfolio play. So the more you deploy on YouTube, the safer you are, which allows us to be more aggressive in our pricing with creators because we can assume that we’ll deploy $1 billion. And that’s obviously safer thousands of channels than doing any kind of single event, right? Single-picture financing on movies is one of the riskiest things you can do… And this is why creators going to banks by themselves is really hard. An individual creator can literally have a terrible 12 months and then become a hit again. Never going to be a business where banks can individually loan out capital to creators based on their YouTube revenue streams.

Relating This Back To Portfolio Theory

By figuring out how to “optimize the asset” and diversify properties that on their own highly volatile, Spotter becomes the most efficient bidder for the financing deals. since they are able to earn the highest returns on the assets both from playing offense (optimizing the asset) and defense (diversification), they are better suited to own the risk and therefore capture the return. And since they can be the highest bidder, they get first look at upcoming deals.

The idea of “optimizing an asset” is similar to a familiar theme in the rise of large tech companies. In recent decades has been their ability to deploy capital safely because of the insights afforded from virtuous data loops. In A Former Market Maker’s Perception of PFOF, I describe the dynamic in the context of market makers paying for order flow:

These [prop] shops came of age at the same time as the giant tech firms. This is a hint of how much they have in common. The difference is the size of the relative opportunities, but the tactics are similiar.

It started with skill and luck. The early big bets on talent and technology meant they were bringing guns to a knife fight. SIG wasn’t know as the “evil empire” on the Amex just because of the black jackets we wore. They understood the risk-reward was completely outsized to what it should be 25 years ago. They were amongst the first to tighten markets to steal market share. They accepted slightly worse risk-reward per trade but for way more absolute dollars. They then used the cash to scale more broadly. This allowed them to “get a look on everything”. Which means you can price and hedge even tighter. Which means you can re-invest at a yet faster rate…The parallels to big tech write themselves. A few firms who bet big on the right markets start printing cash. This kicks off the flywheel:

Provide better product –> increase market share –> harvest proprietary data. Circle back to start.

The lead over your competitors compounds. Competitors die off. They call you a monopoly.

In Making Uncommon Knowledge Common, Kevin Kwok describes a unifying theme behind the success of 3 companies Rich Barton founded (Zillow, Glassdoor, and Expedia!):

The Rich Barton Playbook for winning markets through Data Content Loops…In order to grow their demand high enough to become a beneficial flywheel, Barton’s companies use a Data Content Loop to bootstrap their demand and create unique content and index an industry online (homes for Zillow, hotels and flights for Expedia, companies for Glassdoor). These Data Content Loops help the companies reach the scale where other loops like SEO, brand, and network effects can kick in. Barton’s companies then use this content to own search for their market. This gives them a durable and strong source of free user acquisition, which enables them to own demand…The Data Content Loops of Barton’s companies let them be the authoritative public source on a subject at scale and low cost. The ultimate purpose of the “Data Content Loops + SEO” strategy of Barton’s companies is to own the demand side of an industry…Barton’s companies take industries that are low frequency and use their Data Content Loops and SEO to acquire users for free and engage them more frequently…Owning demand ultimately becomes its own compounding loop since becoming a trusted brand builds its own network effects.

The upshot of all of this: competitors who don’t have the same information or capabilities, will think Spotter is overpaying.

This same dynamic shows up in trading. Your default response to a strange-looking price shouldn’t be “that’s dumb”, it’s “what am I missing?”

Recall this section from You Don’t See The Whole Picture:

When You Don’t Understand The Price You Don’t Understand The Picture

Price is set by the buyer best equipped to underwrite the risk.

A market-maker example:

If X is willing to pay me a high looking price for a stock or option, what’s the probability they are selling something else to someone else such that they are happy to pay me the “high” price? 

Let’s say a call overwriter sees a modest surge in implied vol and is happy to collect some extra premium. Except he’s selling calls to a Citadel market-maker who’s happy to pay the “high” price because her desk is selling index vol. In fact, they are selling index implied correlation at 110%. You might be happy selling the calls for 2% when they are usually worth 1%, but if the person buying them from you knows they are worth 3% at the time you sold them then make no mistake, you are playing a losing game.

However, if your professional edge is in deeply understanding the stock you are selling calls on, then you might be the one capturing the edge in the expensive calls. You are capturing it ultimately from the fact that index volatility is ripping higher and market makers are simply capturing the margin between the weighted option prices of the single stock in proportion to the index volatility. So you, the informed single stock manager, is making edge against the index volatility buyer who set off the chain of events.

The decomposition of the edge between you and the market maker is unclear. But the lesson is you must know where you stand in the pecking order.

An option relative value trade example:

 If volatility surges in A but not in B and they are tightly correlated let’s look at how 2 different market participants might react.

Naive

The naive investor is not monitoring the universe of names. They do not think cross-sectionally. They see a surge in A and decide to sell it. It may or may not work out. It’s a risky trade with commensurate reward potential.

Sophisticated

The sophisticated trader recognizes they can sell A and buy B whose option prices are still stale (perhaps there has been a systematic seller in B who has been price insensitive. Maybe from the same class of investor our friend “naive” came from. They don’t look at the market broadly and realize the thing they are selling is starting to “stick out” as cheap to all the sharps).

Here’s the key: the sophisticated trader will do the same trade as the naive one but by hedging the vol with B, they can do the whole package bigger than if they simply sold A naked.

The sophisticated traders are the ones who see lots of flow. They “know where everything is”. While in this example, sophisticated and naive both sold A there will be times when sophisticated is lifting naive’s offer. Sophisticated has sorted the entire market and is optimizing buys and sells cross-sectionally.

Are you the fish at the table?

Flow traders and market makers are always wondering if their counterparty is legging a portfolio that they’d like to leg themselves if they saw the whole picture…Mathematical expectancy, like a house’s edge, is priced by its most efficient holder. If prices are always being set by the party who most efficiently underwrites/hedges/prices the risk and you know you are not one of those parties then you should wonder…am I being arbed?

Conclusion

Spotter digs a moat between itself and its competitors by:

  1. narrowly focusing on one channel — YouTube
  2. this allows it to optimize the assets within that investing landscape
  3. which allows it to build the most efficient portfolio; and therefore bid the highest

This series of actions gives Spotter “first look” at the deals in the space further reinforcing the loop, allows it to deploy capital faster, and ultimately have a small advantage compound until it creates a meaningful gap between itself and the competition.

Further reading

Portfolio Theory And The Invisible Option On Hobbies (7 min read)

It talks about how, now more than ever, activities you may do for fun might turn into opportunities.