IBC 2023 Recap
Another IBC has come and gone, and there was a lot to talk about this year. With 43,065 people from 170 countries — plus 1,250 exhibitors — this conference boasted a 16% increase in attendance compared to 2022. Despite not reaching pre-COVID figures, IBC chief executive Micheal Crimp hailed it a “fantastic turnout from the entire industry.”
That said, the show felt more energetic than last year with a greater number of new product and service announcements. I noticed more meeting suites and fewer (or at least, smaller) stands. There was also an air of focused intent: large companies like Microsoft and Google appeared to be targeting more scheduled interactions and allowed less access for casual visitors.
Here are my thoughts on some of the most prominent themes and conversation topics from the conference:
The Dawn of the AI Age
You can’t seem to go anywhere without hearing about AI nowadays. Many practical uses for AI were discussed at NAB earlier in the year, but it seems companies are trying to put their money where their mouths are and have started putting those uses into practice and productizing them.
For example, subtitles and transcriptions are relatively well-established capabilities in AI, but tech companies are increasingly adopting them as they become table stakes features on streaming platforms. They’re not as accurate as human-made subtitles, but their increasing proficiency is suitable for an ever-growing number of verticals.
Video indexing and metadata had a significant presence at the conference. Frédéric Petitpont, CEO and co-founder of Newsbridge, demonstrated some of his company’s MXT-1 AI video-indexing technology’s new features. Most impressive was its ability to create automatic media summaries and generate video titles, descriptions, tags, and chapters to speed publishing on digital channels. Similarly, Magnifi’s new platform can auto-generate short clips from long-form content based on facial recognition, speech, places, objects, actions, and audio clues.
Microsoft demoed its related technology, Azure Video Indexer and Cognitive Services. This tool was also capable of object recognition and generating metadata that could group content and make titles more discoverable (which is a major problem many content producers face). Plus, Microsoft showed demos from partners Evertz and Harmonic that could create clips based off voice queries using AI tooling. For instance, a user could say, “Show me all the alley-oop scores from last night’s Knicks game,” and the platform would instantly create highlights and clips.
Streaming services are starting to harness advanced algorithms and ML to analyze user preferences, viewer history, and social media interactions to offer customized content suggestions. And example is FilmBox+ SPI , which made innovative moves to enhance content distribution through personalized AI-driven channels. Plenty of other use cases made themselves known, including codec optimization, using AI to tag shoppable goods, and creating alternate audio tracks in different languages without changing an actor’s voice — not to mention advancing bespoke linear channels and ad targeting.
A Note on Generative AI
Many of the AI applications above can manipulate content in different ways and make it more discoverable — but what about generative AI that can produce it from scratch? Honestly, I believe generative AI in streaming is still in its early days as companies determine how it fits into the ecosystem. Media organizations are exploring it not to create entire new movies or shows (well, some), but personalize content for users and improve media production’s efficiency.
One problem I see is the lack of advertising inventory to support the (primarily FAST) channels coming into play. It’s possible generative AI could create advertising on the fly that’s properly targeted at scale, so I would be interested to see where that goes.
When it comes to AI overall, one lingering question in my mind is cost. Intensive computer processing isn’t free, and many applications are yet to prove their ROI. I suspect this will be resolved in time — as with all innovative technologies — but it will be a bumpy ride as organizations figure out ways to reduce AI’s expense.
Monetization
Monetization was, of course, a key point of discussion. FAST and AVOD are well-established models that aren’t going anywhere, but optimizing the advertising experience is essential due to fatigue in both. Viewers grow frustrated with too many repeated and poorly targeted ads while companies struggle with lack of inventory. Based on what I saw at IBC, I think we’ll see organizations address these problems in a few ways:
- Ad-based targeting according to either user information or video content.
- On-the-fly ad creation with generative AI to produce more relevant, cheaper ads (AWS’s demo of generative AI produced creative images with text-to-image and image-to-image prompts, alongside ad tagline generation that allows for freeform product description prompts, gender, and target age groups).
- Alternative monetization methods such as in-video shopping or gambling. These shoppable videos won’t have interruptions, though — AI could tag different objects throughout a video and make them clickable.
I was also impressed with Intertrust’s Market Maker, which showcased its abilities to integrate digital assets with blockchain-based smart contracts and Eluvio’s Content Fabric demo that proved how it can level up FAST channel offerings with hyper-personalized ad insertion. Ultimately, organizations are learning that if they want to monetize video content with ads successfully, they need to make the viewing experience less frustrating for viewers and customers — and AI can help them do so.
Cost Optimization and TCO
IBC 2023 put a great deal of emphasis on determining and optimizing TCO by reducing costs and increasing revenue. Doing so could lead companies to consider additional or unconventional revenue models, consolidating to fewer vendors with more comprehensive solutions, and evaluating on-prem, cloud, and hybrid options to optimize operating costs versus capital costs.
While the shift to the cloud is real, many companies are taking a closer look at the costs involved and ways to have the best of both worlds with on-premises resources. They’re taking particular interest in higher density hardware options from vendors like AMD, nVidia, and NetInt , which are increasingly available at reasonable prices and offer much more functionality for video encoding and processing than cloud services alone. As organizations look for more sustainable streaming options (monitoring metrics like watts per stream is critical), hybrid approaches make more sense for some. Some platform providers are even evaluating the true costs of building and maintaining their own CDNs instead of using established partners because it may be cheaper if their audiences’ locations and viewing habits are consistent and easily identifiable.
Closing Thoughts
Aside from the wide array of promising technologies displayed at the conference, IBC is always an excellent time for me to get a pulse for what’s happening amongst the participating organizations themselves. The streaming market is saturated with lots of small players and startups that are creating remarkably interesting technologies, but they’re struggling to go to market or productize their ideas. Larger companies with more resources will want to gobble these startups up to turn their IP into something sellable.
There are also many technologies that are difficult to differentiate from their competitors — instant FAST channel creator, anyone? I predict we’ll see a lot of consolidation as companies looking to fill a product gap or accelerate growth will acquire smaller organizations to make their existing products more unique. Not to mention that as customers look to migrate to single-vendor, end-to-end products, vendors will need to resort to acquisitions to create those solutions. It will be fascinating to see which consolidations happen and which AI tools take flight by next year’s IBC.