Your browser is out-of-date!

Update your browser to view this website correctly. Update my browser now


Content discovery: The importance of going global

Just how important is ’being global’? Of course, generating scale economies and gathering international insights is important, but in media, does having a global ability to

Just how important is ’being global’? Of course, generating scale economies and gathering international insights is important, but in media, a market that has historically been limited to country borders for rights and content distribution, does having a global ability to deliver really make that big a difference?

Of course, the question is a little loaded. In recent years, we’ve seen the gradual erosion of the ‘single country focus’, first with the arrival of global multi-channels, and more latterly with the emergence of global OTTs, organisations that use the internet to deliver content across multiple territories, that are also increasingly signing global content deals for exclusivity on their platforms. For these service providers, and arguably the many others that will follow, having a ‘global ability to deliver’ is clearly a critical component of the business model.

So important, yes, but easy to do (and do well), far from it.

Sitting quite prominently amongst the list of challenges here is content discovery. For the purpose of definitions, we’re talking here about the process of building, sourcing, formatting and delivering rich datasets to power EPGs and apps, alongside the tools to deliver effective search and personalised recommendations across both linear and VoD content. And getting this right matters. It impacts directly on the quality of the user experience; the ability to attract and retain subscribers, and the potential to continually increase content viewing.

So where’s the problem?

The first point is that doing this well even on a domestic basis has, in recent years, become increasingly complex. In short, what was a world of text and schedules, with weekly or daily data deliveries to a single destination has become an ecosystem of detailed datasets, video and images, algorithms, tags and semantics, with real-time interactions on multiple devices. As the model becomes global, the issues become amplified.

On a basic level, the question is one of volume – more to source, more to format, and more interactions to deliver. This of course sits alongside the added complexity of sourcing, creating and working with multiple language sets. On the side of search and recommendation, the task becomes one of local insight. What are the subtleties of in-country ‘relevance maps’, ie the differences between territories that lead to liking (or disliking) one piece of content based on viewing or liking another? What are the specificities of language and colloquialisms that will affect meaning and intent within search engines? What commonalities or trends exist that will allow the clustering of regions in order to scale search and recommendation technologies most efficiently?

There are few set answers, but there are key areas to consider as you either start the process of building your own global model, or assessing the capabilities of a potential service provider. They include:

Data depth and data potential: Start with the basics – what depth of data do you need to power the experience? What is the depth of the data the service provider has available? Is it consistent across all territories/language sets? What resources are available to build datasets to fill the inevitable gaps?

Data structure: You can have all the data in the world, but if it’s not well structured with unique identifiers and IDs, at best you’ll be inefficient, at worst you’ll not be able to leverage the data you have.

Data provisioning: What’s your mechanism for getting data to devices? Are you using APIs? How well defined are the APIs? What type of APIs are you using? Are they continually polling or are they only making requests when needed?

What level of flexibility do you have in search and recommendation tools? Alongside considering how these tools work, it’s important to understand how you can flex them. It’s unlikely that one relevance map, one algorithm or one semantic logic will meet the needs of all countries.

What are the mechanisms that exist in recommendation tools to gather audience data? How do they encourage audiences to interact and provide data? And are there mechanisms to assess the effectiveness of recommendations served?

The list is far from exhaustive, and of course content discovery sits as only one of the challenges facing service providers seeking to build global models. But equally, it’s one of few areas that offer such high potential to create unique, powerful and sustainable competitive advantage.

By Georg Mueller-Loeffelholz, head of content discovery