Personalisation special: Under the magnifying glass
Posted on Apr 3, 2025 by FEED Staff
As viewers drown in an ocean of content and advertisers despair of hooking the right fish, personalisation offers a lifeline to both. What does this mean for privacy and the user experience? FEED finds out how broadcasters, streamers and advertisers are putting viewers and content under the microscope, and how their customers feel about it
Words Paul Bray
When I was a lad, ‘being spoilt for choice’ was considered a good thing. Not any more.
“Today’s consumers are drowning in a sea of content, whether that’s from linear TV, streaming or social media, and there’s only more to come,” says Gabriel Cosgrave, EMEA general manager at Xperi. “Without personalisation technologies to guide them, you’ll find people aimlessly scrolling through their televisions. And when you throw non-targeted ads into the mix, it’s easy to see why many develop a case of content fatigue and simply disengage.”
According to a Harris Poll survey for Google Cloud, two-fifths of viewers don’t have a specific programme in mind when they turn on the TV. Searching for something to watch can take them an average of 24 minutes per session – do this daily and that’s three hours of your life every week you’re not going to get back. Not surprisingly, almost half of respondents were willing to cancel a subscription if they failed to find something to watch – if they had a subscription at all.
The cost-of-living crisis, as well as perceived value issues, have pushed many viewers towards pay-per-view and FAST (free ad-supported streaming TV) channels. US research from the Video Advertising Bureau found two-thirds of 18- to 34-year-olds now prefer streaming video content with commercials to ad-free subscriptions.
If viewers don’t like the idea of wallowing in an endless ocean of undifferentiated content, neither do advertisers.
“The reason advertisers are shifting investment from live broadcast to social media and OTT providers is because they know they’re going to reach their target audience, instead of casting their line into an incredibly wide pool and hoping it’ll hook the right fish,” explains Helen Matthews, senior market analyst at Futuresource. “Companies have milliseconds to catch consumers’ attention before they swipe to the next video or change app, so content needs to be targeted to the right people to maintain engagement, otherwise advertisers aren’t going to see any conversion from ad to purchase.”
What most people seem to want is personalisation. Four-fifths of viewers from the Google Cloud/Harris Poll survey said they expect streaming services to provide a highly personalised experience, for example. Personalisation is equally important to advertisers, as Paul Davies, head of marketing at Yospace, details.
“Digital advertisers expect accurate and real-time data on ad views. A bar has been set by other digital platforms, such as social media, that broadcasters must match in order to grow advertising revenues in the streaming age. The power of personalisation is in one-to-one session measurement: knowing exactly how many people watched each ad and for how long.”
Personalisation today spans user interfaces, advertising and content, each driven by AI and data analytics. “Platforms like Netflix and YouTube refine recommendations based on viewing habits, while social media algorithms – such as TikTok’s For You page – deliver individualised experiences,” highlights Ian Blaine, SVP of innovation and insights at Brightcove. “Advertising has evolved, with platforms like Hulu and Disney+ offering dynamically inserted ads tailored to viewer demographics and interests, while services like Spotify can adjust layouts based on individual listening patterns.”
Social media feeds are almost entirely personalised in real time, based on users’ profiles, behaviour and preferences.
‘Hyper-personalised’ engagement could be the next big thing, believes Mrugesh Desai, VP North America at Accedo. “The video industry is starting to take cues from platforms such as Spotify, which uses curated content feeds like Discover Weekly, Daylists, Release Radar and Spotify DJ to continually present users with personalised recommendations. This approach could be taken even further, for example serving content being watched by listeners’ friends and family.”
Documentaries, drama series, movies and reality TV are among the genres that lend themselves most readily to personalisation, whether on the simple premise of ‘you’ve watched this sort of thing before’, or a more sophisticated ‘this could chime with your other interests, or even match your current mood’.
To achieve this in a sophisticated way requires solid metadata, believes Cosgrave. “A good personalisation engine will take everything known about a programme and start to build associations and profiles to better understand a viewer’s interests. This must go beyond standard descriptors, such as genre, to deep descriptors related to the content’s emotional and artistic qualities.
“For example, someone might like romantic comedies. A basic personalisation engine will simply offer one romcom after another. But did they actually like Sex and the City or Shakespeare in Love? The best personalisation engines understand what specifically is driving someone’s choices and viewing habits. Maybe the viewer likes historical settings – or maybe it’s the actors? With good metadata, interesting connections can be drawn.”
Adaptive content could take this even further, highlights Tanya Easterman, chief marketing officer at Dio. “Imagine trailers that adjusted based on viewer emotions, or even interactive storytelling that shifted based on engagement cues.”
Other segments, like live sport, are also adopting personalisation successfully. “Expanding your international reach, as well as fan engagement, is critical in sports, with leagues and rightsholders needing to reach tens or hundreds of countries with tailored live event coverage,” notes Roger Franklin, chief strategy officer at LTN. “They’re utilising live event versioning tools with highly tailored graphics, local-language commentary integration, region-specific frame rates, virtual on-screen advertising, as well as focused content versions relevant to specific markets to provide more personalised, valuable experiences. For example, they might take a live event stream and spin up a Lionel Messi-focused feed for one dedicated fan group, and another centring on Cristiano Ronaldo.”
What we hear can be as tailored towards what we see, according to Cosgrave. “Machine-learning models can be trained to ‘unmix’ audio by identifying acoustic feature patterns. Audio tracks can be separated into dialogue, music and effects, allowing users to boost dialogue based on their own hearing requirements or viewing environment.”
If this allows me to cancel the mindless background music that ruins so much documentary and current affairs output, then roll on the day.
A fascinating prospect is AI-driven biometric personalisation, enabling content to be adjusted on the basis of physical and emotional cues. “Facial expressions, heart rate and other behavioural inputs could influence what a user sees next, creating a feedback loop that constantly optimises the experience,” remarks Desai. “Imagine watching a thriller that subtly adjusted the plot pacing based on your heart rate, or an interface that changed colour schemes or text size based on your current mood.”
Personalisation could extend beyond a single device, as David Dembowski, senior VP of global sales at Operative, points out – for example, through watch parties that allow friends and family to view the same show together, even from different locations.
“Multiplatform providers such as Disney are building infrastructure to unify all their offerings, having just brought Hulu metadata under the Disney+ umbrella,” highlights Dembowski. “This sets the stage for personalised experiences across apps and devices, such as knowing someone was watching a series on their phone when they log in through their TV.”
Another potential area is the car. “There’s an argument that personalisation in-vehicle is more important to reduce time-to-content and keep eyes on the road,” says Cosgrave. “And with operating systems like Tivo available for both TVs and cars, if you watched a war documentary one evening, your car might recommend a similar podcast for the next morning’s commute.”
Personalisation certainly seems to have a positive effect. According to the Google Cloud/Harris Poll research, 79% of viewers had kept a subscription after discovering new content, while 85% of upgrades to paid subscriptions were influenced by the ability to find content.
“Customers using our AI-based Think360 personalised content discovery solution have achieved massive increases in engagement,” says Peter Docherty, founder and CTO of Think Analytics. “Some have seen a doubling of VOD consumption and a 35% increase in the time spent watching content.”
Most viewers seem to be positive about personalisation. In a survey conducted for Xperi by Opinium Research last year, a whole three-quarters of respondents said it was important for entertainment content to be personalised to their individual preferences. This rises to 84% for younger viewers.
In a way, this is hardly surprising. Who wouldn’t want something that saved them time and helped them find programmes they enjoyed watching? What’s more revealing is that the same enthusiasm often extends to the ads.
“Close to three-quarters of viewers prefer advertising that relates to the content they’re viewing, according to research from Integral Ad Science,” points out Kjetil Horneland, CEO of Ease Live.
“I certainly remember, in the pre-streaming days, sometimes noticing that all the ads during a TV programme were completely irrelevant to me, which made it clear that the programme itself probably wasn’t for me either,” says Christopher Bell, global media and entertainment strategy and business development leader at Amazon Web Services (AWS). “Good personalisation is essential commercially and, when handled correctly, sharing private data offers a clear benefit to the consumer.”
Basic-level targeting can be achieved without data on individual viewers. “Time of day or year, the weather or even the user’s general location can provide contextual signals that can make sensible suggestions about which ads to push and when – cold drinks on a hot summer’s day, or food delivery on a rainy night,” notes Alice Beecroft, senior director for global strategy and partnerships at Yahoo Advertising.
Another approach is to match the ad more closely to the content. “An upcoming challenge is Google’s plan to phase out third-party cookies for Chrome, which will require new solutions that can deliver relevant ads while respecting user privacy,” says Stefan Lederer, CEO of Bitmovin.
“Our AI Contextual Advertising,” he adds, “lets service providers deliver hyper-personalised ads to audiences based on the content they are watching by analysing the video, audio and text data during encoding, and passing this information to the ad server. This ensures maximum impact and increased conversions, making it both privacy-friendly and highly effective.”
But what about matching ads to the individual viewer? As broadcasters, streamers, social media sites and their advertisers have more information that enables them to tailor content to us, how concerned should we be about where that information comes from, and where our digital footprints lead?
“Viewers are increasingly sensitive about how their personal and private data is used,” expresses Olivier Karra, cloud solutions marketing director at Broadpeak. “What data is being used – on which service, when and in what context – and who it is being shared with are all concerns.
“For targeted advertising, users have a choice between opting in and opting out. Opting out doesn’t mean advertising won’t be personalised, just that the user hasn’t given their consent to their private data being used, so the experience won’t be as highly targeted. On some free-to-access platforms, publishers push users to create an account and log in on the premise that, ‘if you don’t tell us who you are, you only get a subset of the content or services’. Sometimes, it’s only trailers.”
Advertising can be highly targeted, believes Tim Labelle, head of sales at Streamguys. “There are companies that collect digital buying signals on millions of households. This allows advertisers to home in on mobile advertising IDs or IP addresses that are showing interest in specific topics. That valuable data can then be used in other platforms, such as audio ad insertion or social media, to target specific users with a multiplatform marketing campaign.”
This kind of third-party data – as opposed to first-party data knowingly supplied by the user – must be handled with care. “Viewer data belongs to the broadcaster or streaming platform and is their most-prized asset when it comes to advertising and personalisation,” says Davies. “A number of broadcasters will complement their data with external sources, but this will be with carefully chosen partners that won’t expose their viewers to the prospect of re-targeting by a third party. They also have a legal obligation to be GDPR compliant.
“A big factor here is trust. Many broadcasters are trusted brands that have been part of people’s lives for years, and they can make other digital video platforms such as social media look like the Wild West.”
Transparency is key, believes Chris Kleinschmidt, VP of connected TV advertising sales at Xperi. “Users interested in how personalised recommendations work shouldn’t have a hard time finding out, and platforms now provide choices for managing personalisation preferences, providing users with greater control over their digital footprint. Explaining why certain content is being promoted – whether this is because it’s a trending show or purely based on users’ personal data – can also help users understand the advantages of personalisation.”
For João Tocha, CEO of Digital Azul, it is a matter of education in schools and in society at large about how the media works. “We need to create a better understanding of how and why content reaches us, how advertising works and how media is produced and financed. It’s about creating media literacy in society.
“On a more practical level, users need to be savvy about their data privacy settings and regularly review their online permissions.”
Whatever the hurdles, the dictates of economics and efficiency mean that personalised advertising, like personalised content, is here to stay.
“Typically, a targeted digital CPM is worth three to five times more than a linear broadcast equivalent, although the actual ratio varies a lot, sometimes depending on the amount of user targeting metadata available,” explains Karra.
Personalisation is transforming advertising from a spray-and-pray model into a precision instrument, believes Cory Zachman, SVP of engineering and services at Mediakind. “Dynamic ad insertion technology is enabling broadcasters to tailor ad experiences in real time, meaning every ad and sponsorship placement can be matched with audiences who actually care. The impact of integrating ads into personalised content itself could be even greater, making advertising feel less like an annoying interruption and more like a reasonable extension of the viewing experience.”
However, in its enthusiasm to embrace new technology, the media industry must not forget the most important person: the viewer.
“Perhaps the main misconception is that personalisation is just a data arms race,” admits Zachman. “In reality, it’s an empowerment tool. The best solutions give viewers real control over their experience, while allowing technology to do the heavy lifting behind the scenes. Overload them with clumsy AI guesses, and they’ll tune out.”
He concludes: “Everyone wins when audiences feel like the overall experience is theirs to shape, and not a forced feed of content. Data might be the new currency, but it’s how you spend it that ultimately determines your success.”
Scrapper
Dio believes audience reactions to a programme or movie can be digitised through a mix of pre- and post-viewing questionnaires and advanced biometric analysis.
Cameras track 38 facial data points and emotional states, while on-body sensors record more than 200 points, measuring heart rate and electrodermal response.
On the British comedy drama Scrapper, Dio was able to plot a precise graph of the audience’s emotional responses – including the exact moments when the heartstrings twanged loudest – and discover the scenes that appealed most to different demographics, such as people within precise age ranges. It also pinpointed some less-resonant aspects, such as fantasy moments, so these could be dialled down in the movie’s marketing.
“Using the emotional response of the audience via Dio’s reporting helped us refine our strategy, ensuring we optimised our marketing budget and reached the right audience, ultimately helping us exceed expectations at the box office,” says Sara Frain, director of marketing and distribution at Picturehouse Entertainment.
Subscription and payment
Personalisation can even extend beyond advertising and content to optimising pricing and subscriptions.
“NBA League Pass, the NBA’s direct-to-consumer service, provides personalised subscription, pricing tiers, match-choice flexibility and pause-resume capabilities,” says Paolo Cuttorelli, SVP of global sales at Evergent. “This approach helped achieve 50% growth in active subscribers and a 48% increase in viewing time within a single season.
“Personalisation helps manage churn by using real-time data analytics to monitor the subscriber experience, detecting at-risk viewers – due to buffering or poor streaming quality – and offering them incentives like discounts.”
He concludes: “Customised bundles of streaming services, based on viewers’ individual preferences, can be created by highly automated payment and billing systems, simplifying subscription management and payments for customers.”
Channel 4
Channel 4 (C4), which runs the UK’s largest free streaming service, alongside 11 television channels, has integrated Think Analytics’ tailored content discovery software Think360 with its customer engagement platform, enabling content recommendations to be personalised to a user.
This has enabled C4 to target viewers one-to-one, using real-time behavioural data to maximise open, conversion and click-through rates, deliver relevant and personalised recommendations and thus increase streaming growth and loyalty. C4 also deploys Think360 to personalise the presentation of various rails across the user interface, combining editorially curated content with personalised recommendations.
“We have been able to drive significant growth in streaming uptake through a combination of editorial and AI-powered curation,” says Dave Cameron, C4’s former chief product officer.
Another tech partnership has seen C4 using data from Acxiom Infobase to gain insights into the lifestyle, demographics and purchasing behaviours of its 29 million subscribers, creating roughly 150 highly targeted audience segments that can then be matched to the profile of advertising brands.
This has enabled C4 to increase the accuracy of audience targeting, ensuring households see content more relevant to the products and services most likely to interest them. “Channel 4 has significantly enhanced the user experience, while advertisers have seen an improved performance in campaigns and a maximised ROI,” explains Acxiom’s business development director, Andrew Hooper.
This feature was first published in the Spring 2025 issue of FEED.
