“What is Big Machine actually worth?” Billboard asked in the wake of the brouhaha over BMLG’s acquisition by Scooter Braun’s Ithaca Holdings. The analysis in the article is detailed and thought-provoking — and it reaches the conclusion that the deal was much, much richer than the reported $300 million. This makes Scott Borchetta look like a super-savvy businessman, compensating for all that bad PR as that guy who sold Taylor Swift’s back catalog to a man who’d bullied her.

It also raised so many questions in my mind that I pulled up a spreadsheet and started running numbers. I think it’s very possible that the total value of the deal is around the $300 million reported.

If we want to value Big Machine at $480 million and up, we have to accept three assumptions that ought to be raise an eyebrow:

• BMLG’s leak/PR chose to modestly understate the value of the deal, even though ordinarily M&A announcements brag on total deal size.
• BMLG is in better financial shape in 2019, without Taylor Swift, than it was with her in 2015, when Universal refused to pay $250 million.
• The leaked “more than” $100 million revenue and “north of” $40 million EBITDA accurately predict future performance without Swift’s new releases.

If we question these assumptions, we can justify a roughly $300 million deal that’s worth roughly $300 million, including Borchetta’s equity stake in Ithaca.

This is an exercise in critical thinking, to construct a plausible alternative to Billboard’s analysis. Unless one of the players chooses to release more detailed numbers, there’s no way to know the true valuation of BMLG.

What do these numbers mean?

The three numbers we have — more than $100 million revenue, more than $40 million EBITDA, more than $300 million deal — tell us very little, and I’d much rather see a cash flow analysis and some actual deal terms. But these few data points are what we have and what Billboard has — so let’s poke them with a stick.

First question — what period do these numbers cover? The most plausible answer is fiscal year 2018. So January 1 2018 to– no, wait, this is a record company. Record companies sometimes skew the fiscal year so that Black Friday isn’t at the end of the year. So these numbers could cover 1/1/18 to 12/31/18… or they could cover 10/1/17 to 9/30/18 (the way Warner Music does) or some other date range.

Why does this fiddling with the calendar matter? Because Taylor Swift’s Reputation was released on November 10, 2017, sold over 1 million copies in its first week, and was certified platinum on December 11, 2017. And that’s just in the U.S. Worldwide, sales have topped 4.5 million.

• Let’s assume 2 million of those sales took place in November and December 2017.
• Figure a 15-track album costs around $15.
• Of that $15, the label gets about 60%, or $9.
• That adds up to about $18 million… and that’s before accounting for collector items like zines (yes, there were zines and books!).

If Taylor Swift’s Reputation launch is included in the $100 million revenue, then normal revenues in a Swift-less world might be closer to $80 million. That’s a 20% difference in revenue! Heck yes, it will affect the value of the company!

(And there’s still another 2.5 million in Reputation sales that would be mostly in calendar year 2018, so figure another $20 million or so in revenue. Heck, we can assume annual Swiftless revenue is no more than about $85 million — the real question is whether it’s closer to $70 million.)

But what if revenue was $190 million? Nope, I doubt it.

Human psychology comes into play here:

• If revenue was more than $150 million, someone would have used “$150 million.”
• Any ordinary person, without the faintest P.T. Barnum qualities, would feel justified in rounding up to $150 million from $140 million.
• So actual revenue probably wasn’t higher than $135 million, and it might not even surpass $110 million.

Similarly, EBITDA of $40 million is probably no more than $43 million — because if it were $45 million, you’d say so.

That awesome EBITDA margin of 40% —

–is exactly the kind of weird number that requires some probing. Only a handful of companies in similar industries are that high. Universal Music Group’s EBITDA margin is less than 17%, while Warner’s is less than 12% — wait, what the heck is an EBITDA?

EBITDA is Earnings Before Interest, Taxes, Depreciation, and Amortization. To get a sense of EBITDA, take your revenue. Now subtract:

• direct cash costs of making and shipping the product
• marketing costs
• salaries and payments to contractors or consultants
• rent, office supplies, all the stuff you use to run a business

EBITDA is a non-GAAP (Generally Accepted Accounting Principles) number, so you can go hog-wild with taking out one-time and non-cash costs. Everybody does so, and the way record labels work, oh, that allows us to have real fun with math. We’ll get to that in a minute.

First, though, depending on what the real revenue and EBITDA numbers are, EBITDA margin could be anywhere from about 43.5% (unlikely) to about 32%. It’s easy to make a case for something around 35%.

That’s still really high. Is Big Machine habitually twice as efficient as traditional labels? Borchetta would love you to believe this, especially in the wake of closing the Dot sublabel and purging the roster of artists his team couldn’t figure out how to develop. Who wouldn’t? However, there are at least two other ways to get to a super-high EBITDA margin.

It might not be typical. A Taylor Swift launch should show a higher margin than a newbie launch. Swift is a huge star who generates tons of free press (smaller marketing spend in proportion to revenue) and who can move lots of deluxe and add-on product (generally higher margin products).

There may be aggressive accounting for advances. The way music is made, artists get an “advance” that is supposed to cover living expenses while making the album and may also cover recording costs and some marketing costs, depending how the deal is structured. Why does this matter?

Advances are partly “capitalized.” However much the label can reasonably expect to make back through sales, it capitalizes. That part is not treated as an expense that affects EBITDA. Instead, it’s an asset on the balance sheet, and is amortized over years. (Remember, amortization is excluded from EBITDA.)

The more optimistic you are in your expectations of making back the advance, the better your EBITDA looks. Here’s a round-number, loosey-goosey example based on Warner Music’s financials — notice how EBITDA margin improves as we capitalize more of the A&R costs.

Here’s where it gets fun. First off, you know and I know that Big Machine can reasonably expect to make back Taylor Swift’s advances, so all that money goes nowhere near EBITDA. On the income statement, Swift is practically generating free revenue! (Remember: her back catalog will continue to generate money for the buyer, but there will be no new Swift records.)

Second, there’s nothing stopping Big Machine from being optimistic about its artists and capitalizing a larger portion of advances that most labels would. Dream big! Be bold! (Because let’s face it — to insiders, that 40% EBITDA margin has to be saying: “our A&R is brilliant, our artists coin money, yay us!” Looking over the roster, I’m skeptical that it’s consistently better-performing than the rosters of other labels — which isn’t a slam on the artists, who have less control over that than one might think.)

If this were a publicly traded company, an accountant would be tasked with telling BMLG’s Chief Financial Officer “fuck no, you should be more conservative, you will get sharp letters from the SEC.” But BMLG doesn’t have that oversight.

However, the accounting team doing due diligence on the BMLG-Ithaca deal would have said “fuck no” and recalculated a pro forma EBITDA based on more standard and realistic assumptions. (This happens in valuing privately held companies, so it wouldn’t be shocking or even reflect badly on BMLG.) A more conservative EBITDA calculation in a non-Swift year might plausibly be closer to 25%, and this would still be pretty darn fabulous.

Be fruitful and use multiples

The “12 to 16 times EBITDA” calculation for a sale price, in the Billboard article, is a back-of-the-envelope way of estimating the value of a company. Glancing at the S&P 500 (because it provides lots of data for us to compare) entertainment-related companies are being priced at about that multiple of EBITDA and maaaan, is that high. Entertainment is being priced like the snazzier bits of health care. Record labels are being priced for rapid growth, as music industry revenues have turned and are climbing.

Meanwhile, if we use a $40 million EBITDA, Big Machine is priced at 7.5x EBITDA, which puts it in the cellar with oil & gas and metals & mining — both, coincidentally, industries that tend to have high EBITDA margins because they capitalize a ton of equipment costs. Hmmmm…

Now let’s say a more normal (Swiftless) Big Machine EBITDA is 25% margin on revenue of $85 million, or $21.25 million. On a $300 million deal, that’s 14x EBITDA, or right in line with industry norms.

The $300 million price is also a price/sales multiple of about 3, which is, if anything, on the high side for an entertainment company that will almost certainly see a substantial drop in revenue from having no more new Swift material.

A big question we should have is “were multiples lower in 2015?” The difference between a 12x and a 14x multiple on EBITDA of 21.25 would turn a $250 million deal (the price Borchetta wanted in 2015) into a $300 million deal. For entertainment companies as a whole, no, multiples were about the same in 2015 as now.

There’s no easy way to compare just record labels, as none are public (Warner Music Group, which has financial reports available, was acquired in 2011 and is not publicly traded, so there’s no pricing for the “price to” part of the ratio.) Music revenue was just hitting bottom in 2014, and it’s a guess whether multiples were dropping due to despair or already rising because insiders saw the turnaround coming.

So no equity stake for Borchetta?

Eh, I think Borchetta probably does get an equity stake in Ithaca. Figure it’s 15% of Ithaca’s $800 million valuation, or $120 million.

Let’s say the “more than $300 million deal” is a $330 million deal because that’s a more plausible rounding difference than calling a $480 deal “over $300 million.”

• $120 million of that is Borchetta’s equity stake in Ithaca.
• That leaves $210 million, which is more than Universal offered back in 2015, when Swift was still under contract and Florida Georgia Line were near their peak (and also not looking to bail).
• A $330 million valuation on $21.25 million EBITDA is 15.5x, toward the high end for record labels.

Is this exactly what happened with the deal? Heck, no — it’s very back-of-envelope and doesn’t delve into important issues like debt (is there any? at what interest rate? due when?), a proper discounted free cash flow analysis of the label’s catalog (where Swift’s old albums doubtless dominate), or the number of albums left on the contracts of BMLG’s best sellers (which will have a major impact on future revenue, as replacing Taylor Swift isn’t easy or even likely). Lots of factors could make BMLG look somewhat more valuable or somewhat less valuable.

What it gets us is a plausible scenario for how a company can be worth more after losing the future revenues of its biggest star than it was before. I welcome additional information that would make it possible to fill in gaps, refine estimates, or even disprove my estimates.