December 19, 2023

Axel Springer Makes an AI Deal

Welcome to the very last Tuesday edition of A Media Operator for 2023. I suspect no one is going to care about what’s going on in the media world next week, so I’m going to use this as a week to get everything squared away for 2024. If you missed my announcement, I am officially going full time on AMO and ready to get to work!

I have received so many emails over the last week now that I am going full time, so I wanted to call out a few things:

  1. Become an AMO Pro member to start receiving the second edition of AMO on Fridays. This is where I dig in deeper on specific tactics and strategies for media companies. Plus, there’s a Slack channel where you can chat with other operators.
  2. If you are looking to get in front of 5,000+ media professionals to promote your product, shoot me an email so we can work on an advertising plan.
  3. I’ve received a few questions about whether I am consulting at all now that I am going solo. The short answer is that I am open to advisory sorts of relationships, so if you’re looking for that, hit reply. For the bankers and PE reading, I’m also open to deal diligence arrangements, so reach out. However, most of my time will be spent building AMO, so all of this is very limited.

I hope you have a wonderful rest of your 2023 and I am pumped for next year. With all of that out of the way, let’s jump in.

Axel Springer x OpenAI

OpenAI and other generative AI platforms need a ton of high quality content to train their artificial intelligence models. Axel Springer has loads of that high quality content that it wants to be compensated for. The two sides have reached a deal whereby OpenAI will be able to use Axel’s portfolio of content. (Disclosure: I am an employee of Morning Brew until the end of the week. Axel owns Morning Brew).

According to the Financial Times:

Under the agreement struck with ChatGPT-owner, the German group will earn tens of millions of euros a year to allow its news content to develop generative AI technology that can create text, images and code that are indistinguishable from human creations.

Axel Springer will also allow near real-time access to its news stories to allow the AI platform to provide current answers to questions from its users.

The deal, which FT reports is “an annual ‘eight digit’ sum” is broken into two parts. The first is all of the historical content across various Axel Springer portfolio companies, including Bild, Politico, and Business Insider. Then there is ongoing compensation for all content that is published henceforth.

I have a number of thoughts on this.

First, it is always better to be paid for something than to give it away for free, so it’s good that Axel and OpenAI were able to make a deal. Considering much of the content prior to this deal was used without compensation, this is a way for Axel to get something for its hundreds of millions of dollars invested in content.

Second, this might be a good way for OpenAI to protect itself from any sort of retroactive copyright lawsuits. Now that it has paid for access to Axel Springer’s archive—which you can be sure was used to train OpenAI’s model—it won’t have to worry about any sort of legal concerns. I suspect this is how OpenAI will approach many other large content companies, including News Corp, IAC, The New York Times, etc.

Third, Axel Springer can continue going back to the well for additional payments from other generative AI models. OpenAI paid? Great. What about you, Google? Do you want to pay for the same exact archive of content? And now that the deal has a price, Axel can use it as precedence for pricing other sorts of deals. That puts it in a good position as more of these large-scale generative AI platforms come. And the money can be rather lucrative.

Finally, and this point may shock you, but while the money sounds good, this may be a short-term win and a long-term pain. In July, Axel Springer published a press release talking about some of its financials:

Compared to the first half of the previous year, the company was able to increase pro forma sales by 1.4 percent to almost EUR 1.9 billion. Adjusted EBITDA increased by 12.8 percent in the same period. The share of digital business models in revenue increased to 85.5 percent (previous year: 84.9 percent). The company is following up on a successful 2022, in which revenues of around 3.9 billion euros had been nearly 13 percent higher than in the previous year.

Converting EUR to USD, that’s just over $2 billion in revenue in the first half of the year. If we assume that revenue is flat in the second half of the year, the company finishes the year at $4 billion revenue. Let’s assume it got $50 million from OpenAI. That’s 1.25% of its total revenue. If we’re generous and it got $100 million (obviously high since it only got 8 figures), then we’re at 2.5%.

There are two things worth discussing here, though, and we’ll use the $50 million number for simplicity’s sake.

First, the nice thing about licensing dollars is your costs are effectively zero. And so, that $50 million goes immediately to the bottom line. I don’t know Axel Springer’s EBITDA, but let’s assume it’s a 20% margin business. That means it might generate $800 million in EBITDA this year. $50 million a year matters a lot then. That’s over 6% of your annual EBITDA. That’s not insignificant.

Second, while $50 million in EBITDA is nice, it puts the $800 million at risk. According to a November TechCrunch story:

ChatGPT now has 100 million weekly active users, OpenAI CEO Sam Altman announced on Monday at the company’s first developer conference in San Francisco. The service released nearly a year ago and garnered an estimated 100 million monthly users within just two months of launching and set a record for fastest-growing user base.

Altman also shared today that over two million developers use the platform, including more than 92% of Fortune 500 companies.

If I am one of those OpenAI active users, I could go to ChatGPT and ask it the latest headlines and a short summary of stories from Politico (Axel Springer portfolio company) and get them fed to me without ever having to visit Politico’s website. Imagine that across 100 million people. Ironically, Axel Springer is providing said summaries.

And that’s the risk here. OpenAI is more than happy to pay for access to all of this content because it then gets in the middle of the user and the publisher. I don’t need to visit the various publisher sites to get information; I can just go to OpenAI. And when News Corp, The New York Times, IAC, and various other large content library owners do these deals, OpenAI only gets stronger. And that only makes its ability to play middleman that much stronger.

The best way to monetize a user is to get said user to your site. Licensing is great for incremental revenue that you might not get otherwise e.g. doing an international local language deal. But selling all of your content to a platform that wants to play middleman is a risky business. And that’s why I call it a short-term win, but possibly a long-term pain. The only way to prevent that pain is to maximize the clickthrough from OpenAI to Axel’s sites.

The deal provides OpenAI with summaries and then whenever that content is used, OpenAI must show a link to the user. And so, if the summaries can be generated in such a way that it leaves the user wanting more, it might turn out to be an audience development tactic. It’s ultimately a user behavior question. Will the summary be enough or will the user want to click for more information?

We should expect to see more of these deals happen because the money is there and they’re so EBITDA-friendly. But we should be careful blindly celebrating them because it cuts us off from the reader. The money is nice, but we need to figure out ways to acquire users and own the means of communication with them.

But this is the reality one way or the other. OpenAI and others were already getting access to the content. Might as well get paid for it, right?

Punchbowl going down funnel

Going deeper and offering more context than your competition on a specific coverage area is a sure-fire way to build a lasting and successful media company. And that is exactly what Punchbowl News is trying to do with Capitol Hill.

Last week, Punchbowl News announced that it was acquiring Electo Analytics. The New York Times wrote about the deal:

Punchbowl News said Thursday that it was acquiring Electo Analytics, a company that provides data to help decipher and analyze legislation. The stock deal values Punchbowl News at more than $100 million, according to two people familiar with the transaction.

The company plans to increase its annual subscription by $50 in January and will also eventually offer a higher-cost subscription that includes data from Electo Analytics for more than $1,000, but the final price hasn’t been set.

Punchbowl News is one of the most promising entrants to the digital news industry in many years. Ms. Palmer said the company generated $20 million in revenue this year and is profitable. Though the company is primarily focused on subscriptions, she said it is nearly halfway to its projected ad revenue target for 2024. The advertisers include Meta and Exxon Mobil.

I have not hidden my admiration for Punchbowl. CEO Anna Palmer spoke at the AMO Summit in October and she alluded to this deal happening when she said, “we really shifted in year two and three to make this a b2b strategy. So, getting a company in, having a bunch of slots. We’re tinkering with enterprise—you’re going to be hearing more about that at the end of the year.”

Well, we’re hearing about it.

Here’s why this deal makes sense. The Punchbowl News team has spent the better part of three years building an audience that cares about the inner workings of Capitol Hill. And we’re not just talking about politics. We’re talking about understanding the inner workings of government where policies and bills can mean millions or billions of dollars to companies around the country (and the world).

And now it can take that audience and push it down funnel, exposing all of its subscribers to the policy-tracking that Electo Analytics provides. This can be potentially very lucrative for two reasons.

First, it increases the potential LTV of a subscriber significantly. If it can convert some percentage of its $300/year paying audience to $1,000+ a year subscriptions, that’s a big win.

Second, it introduces this concept of negative customer acquisition costs. For example… imagine the Punchbowl News team uses information from Electo Analytics in its reporting with link to Electo. That same story has a sponsor on it, so readers are being monetized with the ads. When a user signs up for Electo from one of these ad-supported stories, Punchbowl News was effectively paid to promote Electo. That’s a negative CAC.

I’d like to see more b2b media companies do these sorts of deals. But they’re not easy to do because there is a delta in acquisition multiples between media and SaaS. Software simply trades for a higher revenue and EBITDA multiple. In essence, the media company could, potentially, have to dilute itself pretty significantly to afford this sort of acquisition. But, if it’s the right asset, you have a built-in audience.

The total addressable market for influential people with budgets that also care about Capitol Hill is not massive. And so, for Punchbowl News to continue growing, it needs to make sure that it can offer that audience all of the products they need to make better decisions when it comes to Congress. This deal is one way to do it. I would expect to see more.

Thanks for reading today’s AMO. If you have thoughts, hit reply or become an AMO Pro member for an invite to the AMO Slack and to start receiving the second edition each Friday. I hope you have a great week and holiday.