Entering the media matrix

Asset management is being transformed as even the oldest content prepares for a second life

“Binary code, plus descriptive information.”

The above quote is how David Lipsey defines an asset.

Lipsey is one of the pioneers of modern digital asset management (DAM) and co-founded the world’s only professional training programme in DAM and MAM (media asset management) at Rutgers University. He is also a co-founder of the Lab of Excellence in DAM at Toronto Metropolitan University, the only institution in the world devoted to the study of what a digital asset is and what can be done with it. But even he believes we’re only just beginning to understand what digital assets are and what they can do for rightsholders and audiences.

“Rights are the third leg of the stool,” he continues. “Being able to know whether you can use an asset and use it in a thoughtful way.” The all-digital media world means that rightsholders are continually accumulating assets – from dailies to finished shows, pre-production art and marketing materials – at an overwhelming rate. At the same time, more and more distribution channels are demanding to be fed, each with their own peculiarities and audiences.

Media companies need to find solutions for storing and managing all those files, but more importantly, new thinking about how to manage and monetise what they own – transforming data into genuine business assets.

The descriptive information Lipsey mentions – metadata – is the key to unlocking value in an asset. In the first instance, it makes content search easier, but it also allows you to see how and why your content could become monetisable in any given circumstance.

When we look at metadata, we can see much more clearly that the asset uploaded to storage in 2016 is not the same asset today. A particularly obvious recent example might be that footage of Prince Charles is now footage of King Charles III – or is it?

Librarians and archivists have wrestled with these issues for many centuries, but in a digital world, where metadata might be input by an intern or a robot at scale across huge numbers of files, it’s easy for assets to get lost or become nonsensical.

“Metadata decays,” says Lipsey. “Think about what we have seen in five years of vernacular and common language evolution, and the expense incurred to clean up metadata. If you and I were poking around museums in London or New York and asked to see the primitive art, we’d be laughed out of the room. Five years ago, it would have been: ‘Yes, we have a whole gallery of primitive art.’

“So what are we investing in metadata? It puts a whole different scale on the AI investment to look at return and time to value… maybe metadata improvement is a capital cost? This gets into interesting conversations different from ‘I can compress that file and get it to New York that much faster.’”

The zen of unstructured assets

For Chris Fournelle, director, content and marketing production at Signiant, establishing the best practices and finding the best tools requires giving up the idea that you’re going to have total control over all of your content all the time. It’s no surprise to him that many companies have moved away from a monolithic, centralised MAM for storing their content.

“Content is everywhere, and the complexity, speed and creativity the industry is trying to achieve makes things messy. It makes things disorganised. It makes things unstructured. Content is everywhere at various stages of the supply chain and it’s no longer a linear process. The idea of having everything in one place as your gold standard of truth is probably not realistic.”

With assets spread out across companies and the supply chain itself, what’s needed is a way to see into relevant assets, regardless of where they live. A layer like the Signiant Platform can enable federated search across all of your storage, whether it is in the cloud, on the premises or in any partner storage you have permission for.

It’s funny we’re still stuck with the word ‘storage’ when digital content turns storage into something very immediate, which updates instantly. The story of asset management has been one of moving out of archives and more and more into the present.

“You could shoot something while it is being immediately sent to the studio location, where it is able to be played at a high resolution without making proxies, and can be clipped, downloaded or transcoded for dailies. Legal or script supervisors or whomever can look at it while it’s still being shot. At every step of the production cycle, stakeholders with access to it do not have to wait for it to go into the final stack.”

Coordination first, tech second

Chris McMahon, senior product marketing manager at Backlight, notes that this ability to effortlessly access multiple assets in multiple places brings its own problems.

“If there’s a challenge people still experience today, it’s less of an accessibility problem, but more of a coordination problem.

“A company will come to us with a producer who needs to find a clip from a shoot that was six months ago – but that footage lives on a NAS (network-attached storage) somewhere, or maybe got moved to an S3, but nobody’s sure. There’s a folder structure, but it only makes sense to the person who set it up. So the producer is messaging three different people and waiting for somebody to dig through drives. They pull it down, then realise it’s the wrong version. These are the kind of things we hear a lot.”

In this stressful example, there is no technological reason why the producer’s clip can’t immediately be found. The stumbling block is the lack of good-old human planning and oversight – including agreed-on metadata. McMahon says the basics of storage hygiene and coordination are the first talking points when working with customers.

“Without that coordination and governance, you can have all of your assets available in one form or another, but it will fail. It will impact the economics – duplicate storage can account for 30 to 40% of media spend. All of this access can quickly become this uncontrolled variable.”

One of the – many – promises made by AI companies is that you won’t need to have a plan when it comes to working with assets and content. An AI will train itself on all of your content and processes and you’ll just have to ask it to perform tasks, which it will execute perfectly and without complaint.

The truth is AI is useful in MAM’s speech-to-text tools, which are the hero of modern media management. Face detection is also a useful AI feature. But it’s easy for companies to get enticed by AI just to overly complicate tasks and spend too much money – and carbon and water – on something that will eventually need to be done by a human.

Meaning, supplied through human-created metadata, is where business value is most likely to come from.

Getting practical with PAM

“I’ve been around thousands of MAM projects in my day and have seen many of them fail,” says Derek Barrilleaux, CEO of Projective. “One of the reasons is that what they need isn’t a MAM, it’s a PAM.”

A production asset management (PAM) system helps companies with the problem of managing assets during the production and content-publishing process, which now has a much longer lifespan than it used to.

“MAMs are great for distribution, management and repurposing of finished assets, providing a long-term library and managing rights. They are good at orchestrating the flow of these assets around a big facility, but ill-suited for the chaotic, work-in-progress part of things.”

There is much technical literature written – by David Lipsey and experts at Toronto Metropolitan University, for a start – about what constitutes an asset. But, as with the states of matter, assets can exist in different energetic forms before they finally cool off for long-term archiving. Although, as we’ve seen, companies also need to be able to publish their archived material quickly when called upon. The early, diffuse and liquid state of assets has to be managed with tools that are more process

than object focused. “In the world of MAM, there are many important, very rigid things you can do with that finished asset, but you don’t want to apply those rules and that rigour to 200 clips sitting in a project you might end up purging 85% of.”

In the framework of a production asset management system, files are grouped according to the content goals at hand – ie the project. The benefit is to make it as frictionless as possible for editors to focus only on the files they need in the most efficient workflow to get the completed project delivered.

This intense focus on process has the potential to reveal new and better ways of working, rather than reaching for a tech fix to something that isn’t really a tech problem. Barrilleaux points to the tendency to respond to filling storage by buying more storage; or getting hold of an asset problem by buying an asset manager; or solving a remote access problem by buying a remote access tool, without taking into account how the whole workflow is operating.

“When you zoom out, you start to see why some facilities have crazy Frankenstein workflows of things that are just bolted together. They would have made decisions to buy individual things that all made sense at the time, but now mean they have all these different things to manage and try to piece together.”

Automated assets, automated revenue

Machine learning is still in its infancy, but the technology is built on principles of pattern recognition, which makes it especially useful for finding things.

Human goal setting and value judgement is not going to be replaced when it comes to asset strategy, but AI can execute processes instantly and at a huge scale that could only be accomplished by armies of humans working 24/7.

Moments Lab has been a leader in developing sophisticated image analysis for content owners to monetise their libraries. The company’s AI-powered toolkit mines archives and is able to provide rich metadata about the content and subject matter.

One upside to the overwhelming proliferation of content platforms is the greater opportunity for dormant content to find new life – and new money – through content licensing, repurposing programmes for different formats or using existing content to build new shows. Big media companies can have many thousands of hours of content, built up over decades, waiting in storage.

“Fresh content is expensive,” says Philippe Petitpont, Moments Lab co- founder and CEO. “There’s now a big interest from companies saying: ‘I’m sitting on a goldmine of content. How can I transform this into revenue?’”

Moments Lab has been working with the French media multinational Banijay Entertainment to analyse its content library and get it back out into the world.

“They are particularly interested in doing this because they already have the rights,” explains Petitpont. “Before, it was too expensive to produce or repurpose these shows because there was a lot of manual logging and manual editing involved. But if AI can do this for around $20 per clip, then it makes sense.”

Another big customer in the US is using Moments Lab not just for repurposing content for social media clips, but for creating an entire FAST channel. “The viewership of a FAST channel is very small, very limited. So again, you don’t want to spend time doing the programming and scheduling. You need to scale

your content operation so there’s no human intervention to feed this beast. Both FAST and social media have to be fed all the time. That’s what we are helping with.”

When Moments Lab’s AI agent pours over an archive, it identifies and tags faces, objects, locations, text, and speech, but it can also identify narrative arcs and certain types of dramatic moments – so ‘interesting’ clips can be extracted automatically, and even published to social media, with minimal human intervention.

The recent hype around micro dramas – compact serial narratives, usually released in vertical video form – opens up myriad possibilities for repurposing old assets.

“One way of doing it is to say ‘I am not going to record a new micro drama. I’m going to take an old show, identify, extract and then publish its cliffhangers. You can turn an episode of 40 minutes into maybe ten four-minute episodes. “We’ve started this with customers – we cannot mention them, but they are all famous and it’s beginning to work very well.”

But having the tools to repurpose assets easily doesn’t always mean you can. As archives and assets become more accessible and monetisable, the biggest friction point in the supply chain will be the legal department.

A media company can repurpose an asset in a hundred different ways that would have been cut into a show once and then forever forgotten about a generation ago.

The crucial question now is do they have the rights to do so? I may own rights to a show, but does that contract also allow me to recut and republish different versions? Or can new versions be justified as marketing materials rather than new content? And if you bring GenAI in to generate or enhance content as well, that carries with it even more complications to navigate.

Tracking the rights around your assets is just as important – maybe even more important – than tracking the assets themselves.

“You can think of rights as very binary: either we have the rights or we don’t have the rights,” says Petitpont. “But actually it’s way more granular. Maybe there are only some seasons when you have the rights, or you might only have the rights to specific countries.

“It’s still a big point of pain in our industry, and I don’t see how to solve it with technology. I believe that repurposing content will need a new kind of rights that will need to be defined. The border between marketing assets and a repurposed show that is contributing revenue is going to be very thin.”

This article appeared in our NAB 2026 issue of FEED