Are AI Platforms’ Rev Shares a Good Move for Publishers?

By: Jacob Cohen Donnelly

What if publishers could share in the revenue that AI platforms generate in exchange for a license of their content?

It’s a question that some AI platforms—Perplexity and ProRata.ai—are asking. Rather than forking over an upfront licensing fee, such as the ~$16 million a year that OpenAI is giving to Dotdash Meredith, the platforms will take the content for free, but then pay out as they make money. In the case of ProRata, it’s a 50/50 revenue split.

According to an Adweek story (which is part of both Perplexity and ProRata’s programs):

Gross (founder & CEO), who has been credited with inventing the pay-per-click keyword ad model, the backbone of the digital advertising industry—sees ProRata as a way to solve the issue of AI firms siphoning off media traffic and ad revenue. Unlike traditional media types like music streaming, which track revenue via views or streams, generative AI lacks a clear metric for attribution. Gross said ProRata’s model is similar to how Spotify tracks streams to distribute revenue.

It’s a good plan for the AI companies, but what about for publishers? Revenue share programs are far lower risk for the AI companies because they only have to pay out money based on usage. If no one uses the platform, it’d suck for them, but they wouldn’t lose any money on licensing the content. Compare that to OpenAI, which has to give publishers money irrespective of usage.

As for publishers, the risks are equally as low so long as they read the terms & conditions. Does it matter that a publisher has licensed content to a platform if no one uses it? Theoretically, no. But it becomes problematic if your content helps the platform indirectly monetize—perhaps by licensing to other platforms—and none of that indirect money goes to publishers. I haven’t looked at the agreements publishers are signing, but they need to be very careful here.

For argument’s sake, let’s assume that there is no nefarious language in the agreements. Are these sorts of agreements worth exploring? Is there real money there?

It really depends. Gross’ comparison to Spotify is an interesting one because Spotify has become an incredibly lucrative part of the music industry’s monetization scheme. According to Billboard:

At $9 billion a year [2023], Spotify accounts for a significant share of the global music market. While it’s too early in the year for 2023 global industry data to be released, rough estimates based on existing data give a clue about Spotify’s immense value. The global value of music copyright was $41.5 billion in 2022, according to music economist Will Page. If the global value of music copyright rose 10% in 2023 — not an outlandish assumption given recent trends — Spotify would have accounted for roughly 20% of label and publishing income worldwide.

If we look at Spotify’s most recent quarterly financials, it reported 252 million premium subscribers and is generating roughly $5 in monthly revenue per user. Spotify also has an ad-supported version, which is relevant to our analysis. In the quarter, it had 402 million monthly active users on the ad plan and generated $472 million in revenue from them—or $0.39 in monthly revenue per user.

So, 252 million premium subscribers and 402 million ad-supported users. This means about 38% of the users are on the paid plan. But that’s at a $5 ARPU. ProRata wants to charge $20 per month. If we use a very basic price elasticity model against Spotify’s numbers, the percentage of people that would want to pay is 9.4% (38% divided by 4 since ProRata’s subscription is 4x Spotify’s ARPU). This is likely too high considering OpenAI reportedly has 300 million weekly users, but as of September, only had 11 million paying users.

If ProRata can get to 1 million total users in its first year—and if we assume that free-to-paid ratio from above—that would imply about 94,000 paying subscribers and the rest on the free plan. Here’s what that looks like financially:

• 94,000 paid subscribers * $20 subscription price = $1.88 million
• 906,000 free subscribers * $0.39 (Spotify ad ARPU) = $353,000

Every month, ProRata would generate about $2.2 million in revenue and publishers would get half. As ProRata grows, publishers would continue to see that number go up. But as more publishers join the network, the money gets split amongst more partners based on how much of their content is used. Therefore, this is unlikely to become a massive revenue driver for the average publisher.

Must be a waste of time then, right? Actually, no. First, the user experience that Adweek showed in its article is very publisher friendly. Between links in the body of the AI-generated result and then sources on the right rail—at least on desktop—it should be a very “click friendly” alternative. If a publisher can generate money while also driving traffic, that’s incredibly helpful.

Second, this revenue should be in addition to anything else you’re doing as a publisher. I’ve spent most of this article talking about ProRata, but Perplexity is also going to share revenue. Could we see a world where publishers are getting cuts from dozens of different participants? Some will be friendlier than others, but if they’re getting traffic and money, this might become a legitimate line on the P&L.

And even then, this is not the best advice for every kind of publisher. The more niche your content, the more valuable it could be, and making it part of a flat revenue share may not be in your best interest. However, if you’re a publisher with both a broad audience as well as a more niche one—say, Politico or Adweek—it might make sense. This is how Adweek is thinking about it. Mike Beyman, Chief of Staff/Head of Strategy and Business Operations at Adweek, told AMO:

We have partnered with Perplexity, Prorata.ai (and also Tollbit, also recently announced) because, as a B2B brand with an ingrained B2C audience (we have non-industry professionals reading our creativity & Brand Coverage as well as TVNewser & TVSpy blogs), it’s a no-regret move.

This broadens our audience and content reach while not compromising our core constituency (brand marketing media and advertising leaders and those looking to reach them) and could be an excellent additional avenue to monetization. I’m seeing it as growing the overall monetization and audience pie, not dividing it differently between legacy search (e.g., Google) and others.

This is the right way to think about it. It’ll monetize its B2B audience on-site—likely with events, lead gen and subscriptions—while it can get more scaled opportunities on the AI platforms. Perhaps this is a classic win-win.