OP #308: The 2x2 of Monetizing LLMs

Some learnings from my past that may help those who are thinking about monetizing LLMs.

Author

Darren Herman
September 23, 2025

Happy Tuesday OP Community.

Last week was a blur and ended with a Bain Capital PE offsite in Newport, RI with a competitive sandcastle building challenge. I have built sandcastles with my children and some kids we’ve met on the beaches over the years, but picture 180 type A personalities in teams of 15 building sandcastles to win… yup, exactly what you are thinking.

Because I was OOO at the offsite, I didn’t read as much as I normally do, so my articles and links are limited in this OP. With that said, a fun article that I spent some time with is focused on the mathematical challenge of making a 30 team NBA schedule. I never really thought about how challenging that is… but wow. The quote of the article is: “If everyone hates you equally, you’ve done your job.”

Here’s a few articles to get us started:

The main piece of this OP is based on conversations with startups who are thinking about LLM monetization and relating it back to my time at Mozilla, where we took an existing large scale platform and developed a content and ads monetization platform and business from scratch. Hopefully some of the below is helpful - regardless of whether you are part of the ecosystem or not.

Be well, do good.

Darren

The 2x2 of Monetizing LLMs

Every new platform faces the same challenge: do you build the ad unit, or do you build the budget?

When I joined Mozilla (Firefox), there was no ad or content unit. No monetization product. Just a smart group of people with a vision.

We had to build a content and ad recommendation system from scratch—one that respected privacy. That meant no shortcuts: we weren’t going to license Outbrain, MSFT, Google, or any of the usual suspects due to their non-privacy preserving tech. We needed to zig where the industry zagged.

Building the tech was the easy part. We had a barbell team—half deeply experienced ad tech veterans, half newcomers with no baggage. That mix gave us enough grounding to know the rules, but enough naiveté to break them.

The harder part was commercializing. Marketers and agencies needed the currency of measurement. They wanted to justify buys with ROI, not just the story of “this is the hot new thing.” At first, we couldn’t give them that currency. We won some buys, but retention was tough.

Months later, we delivered measurement—our way, the Mozilla way. For some advertisers, it worked so well that five-figure IOs turned into guaranteed seven-figure deals. A few agency leads even told us privately: you’re underpricing this. They wanted open IOs to buy as much inventory as possible.

Fast-forward to today: LLM companies are beginning their own monetization experiments. Sponsored answers. Contextual integrations. Native monetization in the prompt window.

If you’re one of those companies, here’s the strategic question:

Do you create something completely new—an ad unit with zero budget allocated against it? Or do you anchor in an existing format, where budgets and buying behavior already exist?

That’s the essence of the 2x2:

  • Familiar ad units + existing budgets = easiest to scale.

  • New ad units + new budgets = hardest to scale.

There’s also the matter of time. Even with the right unit, scaling commercial monetization takes years. Look at Amazon. I was still on the agency side in 2007 when they were testing their earliest ad units with my team (client: Vanguard). Today—13+ years later—they’re still building momentum. Are articles like this giving Perplexity enough time?

Google’s AdWords scaled because it mapped neatly to familiar buying behavior (GoTo, phonebook listings). LLM monetization may follow affiliate-style paths, or something equally familiar, before it evolves into its own category.

So the question becomes: which pools of dollars are you going after—existing or new? And what level of patience do you have to scale?

The easy part will be building the tech. The hard part will be earning the currency of trust from marketers to get past their test and innovation budgets.

10x Your Productivity

If you read OP #307, you’ll likely realize I forgot to link to the main video to watch. Here it is. I apologize.

To refresh, it’s a ~20m video from Jeremy Utley, Stanford’s AI + Creativity expert discussing how to be even more effective with AI. While the video could initially feel basic to some, there’s some really good tidbits throughout for the more advanced.

Below are a few articles I came across this past week that I found interesting. While I may not agree with everything in each one, I think they're worth a read. If you stumble upon an article you think I or the Operating Partner community would enjoy, feel free to share it with me. Of course, I reserve the right to decide what gets featured in the OP.

Thank you for reading and subscribing to OP #308.

What did you think of today's OP letter?

Login or Subscribe to participate in polls.