From Measurement to Foresight

How Luminate is building the data infrastructure for the future of entertainment

Luminate sits at the center of some of the entertainment industry’s most important questions: what should be measured, how it should be measured, and what happens when the old metrics stop being enough. In this conversation, Luminate CEO, Rob Jonas, reflects on the company’s infrastructure overhaul, its expansion into other entertainment verticals, the rise of AI-generated music, and why the next phase of growth may come from helping the industry move from descriptive analytics to predictive—and eventually prescriptive—insights.

For readers who may not know Luminate well, what does the company do?

Jonas: Luminate is a data and analytics business focused entirely on the music and entertainment industry. There are data companies that work across every sector you can imagine—aviation, produce, finance—and we do the same thing but for music and entertainment. In music specifically, we are the largest or most established company of our kind, built over many years in different forms.

One of your biggest strategic moves has been modernizing Luminate’s infrastructure. Why was that necessary?

Jonas: The short answer is that the technology we had bought to create what is now Luminate was old and no longer fit for purpose. It was nearly ten years old when we started looking at it, and even then it wasn’t fully cloud-based: it was a mix of on-prem and cloud, which just wasn’t designed for where we needed to go.

We started the migration in 2022, and it took more than two-and-a-half years to complete. The reason is that a data company like ours is so deeply integrated into customer systems. We couldn’t just shut one thing down and build another from scratch. We had to run both systems in parallel for a long time, which was expensive and complicated, but it let us build the system we wanted for the next decade.

The timing turned out to be critical. We didn’t know when we started just how quickly AI would become central to the industry. If we had stayed on the old system, we would not have been able to take advantage of these new tools. The new infrastructure is designed to do exactly that.

With that infrastructure now in place, what new phase of growth does it unlock for Luminate?

Jonas: AI is the big tailwind. For the last four or five years, we’ve been very good at counting and presenting things accurately, which matters a lot. But customers increasingly want us to predict things. That means moving from descriptive analytics to predictive analytics, and then to prescriptive analytics.

That difference is important. Predictive is “we think this is going to happen.” Prescriptive is “we think this is going to happen, and this is what you need to do about it right now.” That’s a big shift for our business and for our customers. I think music will embrace it because the industry is very hungry to understand what it can do to improve performance. Film and TV will probably take a little more time.

That kind of prescriptive role depends heavily on trust. How do you think about neutrality and credibility in a business like yours?

Jonas: We have to be very careful. In music, we have massive trust with our data partners because we’ve protected those relationships over many years. That trust doesn’t yet exist in film and TV in the same way, but we think it will come. We need to show that we can be just as careful and respectful there as we have been in music.

Ownership structure matters, too. If we were heavily owned by the majors or by DSPs, that would raise very different questions about our role in the industry. Our structure is relatively neutral. We do have one shareholder with a large media and advertising portfolio, and we work closely with sister companies there, but so far that has been positive.

AI-generated music is raising major questions about charts and measurement. Should charts reflect only consumer behavior or also how the music was made?

Jonas: This debate is getting louder every week. Last year, I kept asking people two questions: when do you think an AI artist will be number one on the Hot 100, and do you care? People had very different answers. Some thought it would happen within months. Some said never. But most agreed it was coming sooner rather than later.

The answer to the first question really depends on what counts as an AI artist. I think the closest definition is probably something like Xania Monet, where there is a human behind the creation, but the human is using tools to create music from scratch. The second question gets to the heart of whether it should be allowed. One view is that AI music should never count because it is not “real” music. Another view, which I think is compelling, is that music is not how it is made; it is how it makes you feel. If it creates the same emotional response in listeners as organic music, then it should be allowed to participate in the charts.

The problem is that more and more of this music is being created by machines and listened to by machines. That’s a really big problem because it removes economics from the royalty pool, so the debate is nuanced. It comes down to whether the music creates a genuine human response, whether it respects intellectual property, and whether the audience activity around it is real.

How will the rise of AI-generated music strain the industry’s already fragile metadata and rights infrastructure?

Jonas: Metadata in music is already in bad shape. There are accuracy problems, there are huge gaps, and it was not designed for what is coming next.

We’re trying to tackle the AI question from two angles. First is whether we can identify AI artists. I think self-reporting is going to be problematic. Some DSPs are going down that route, but I think someone, or multiple parties, must determine what is and isn’t an AI artist. That’s not easy, but it’s relatively solvable. The clues often start in the metadata because the metadata coming from an AI artist or AI track is very different from that of an organic artist or track. From there, you can build models and heuristics.

The second bigger problem is attribution. If lyrics or recordings are used in an unauthorized way inside an AI-generated track, how do you attribute that properly? No one really has a great solution for that yet, but someone has to solve it because it affects royalties, market share, and the economics of the whole industry.

How do you keep legacy metrics useful while incorporating new forms of activity like short-form video or AI-driven creation?

Jonas: We’ve historically dealt with this by creating equivalencies across different types of activity and translating multiple forms of engagement into a common framework. I still think that approach makes sense, but it’s getting more complex.

Short-form video is a good example. It’s already a powerful engine for music discovery, but it’s a very different kind of activity from streaming. You’re measuring creates, likes, and shares rather than stream counts. What matters is understanding how those behaviors drive downstream activity on platforms like Spotify or Apple Music. Once you can do that, you can start to build them into the framework in a way that makes sense.

I think AI will introduce another layer of complexity because it will create new modes of both consumption and creation. But the underlying challenge is the same: how do you incorporate new forms of activity without losing the ability to compare them meaningfully? We operate at the intersection of growth and disruption, and as long as both remain manageable, there’s a lot of value to create.

How do you decide when to act and when to wait, especially in a fast-moving area like AI?

Jonas: We have a bias toward action. Sometimes that puts us in mild tension with customers or partners because we tend to move quickly. But we also try to create room for experimentation.

We have not mandated that everyone in the company use these tools. Instead, we give teams a framework: here are the tools, try them if you want, and if you do not, that is fine too. The idea is to reward experimentation without making it feel compulsory.

That approach has worked well. The early adopters test things [and] tell their colleagues, and momentum builds. What has surprised me most is just how fast AI is moving. I have been in technology for a long time, and I have never seen anything move this quickly.

Luminate has expanded internationally while also moving into film, TV, and gaming. How do you balance those priorities?

Jonas: The reason we can do both is that they have different objectives. Five years ago, our music business was mostly focused on the U.S. and Canada. We made an early decision to expand internationally, and now we’re in more than sixty countries. That matters because customers who work globally want a common data model everywhere they operate. If they want to understand how Taylor Swift performs in Japan, the UK, and the U.S., they want the same framework across markets.

The Tencent deal is particularly interesting for a different reason. Chinese music companies are looking at what happened with K-pop and J-pop, where artists built international audiences through partnerships and better visibility into overseas markets. We think C-pop is going to follow a similar path, and data will help tell that story.

On the film and TV side, it’s a completely different challenge. Music is incredibly transparent. If a recording has been around for thirty years, we can understand how it’s performed for thirty years. On the film and TV side, that kind of data simply doesn’t exist in the same way. A few years ago, I sat down with a production company after they had launched a major show on Hulu, and I asked how it was performing. Their answer was “we have no idea.” That contrast is striking. We think that lack of transparency is a major problem, and we’re trying to solve it.

As Luminate expands further into entertainment, what would make you say the company has truly become entertainment-wide?

Jonas: At the highest level, the goal is to make our film and TV business as important to our customers as our music business is today. Gaming belongs in that ambition, too. We want to play a central role in music, then have the same role in film and TV, and eventually in gaming. We also want our non-U.S. business to look as strong and relevant as our U.S. business. If we can do that across music, film, TV, gaming, and global markets, then we’ll have built something truly comprehensive.

As you look ahead, what are your biggest priorities and what keeps you up at night?

Jonas: We’ve just finished a long technology project, so in some ways it feels like we’ve finally reached a starting point. Now we have the infrastructure to build on top of. 

There are a few priorities. On the music side, we’ve been very strong in recorded music, and after acquiring the publishing side a couple of years ago, we’re now building more of a publisher-centric offering. Linking those two well is something no one is really doing yet. On the entertainment side, developing our film and TV streaming offering is important, and so is international growth.

What keeps me up at night is mostly talent. Can we keep finding the right people, keep them motivated, and keep growing at the pace we want? The good news is that a lot of smart people are naturally drawn to music and entertainment. The harder question is making sure we continue to attract and retain them. AI is another real concern because the pace of change is so fast that it is hard to predict where things are headed next.

Alexandra Zaoui (MBA ’27) is originally from London, UK. She graduated from Harvard College with a degree in Applied Mathematics. Prior to HBS, she worked in data and analytics in the music industry.