A recurring discussion at NAB concerned whether it is getting harder to make money from selling content personalization and recommendations software?
Some executives were adamant that this is a fledgling field with a long runway for innovation and riches in streaming media, while others approached this subject with more caution, when speaking with Faultline in Las Vegas last week.
Clues to support an argument in either direction were dotted around the event. ContentWise, for instance, had all-but dissipated into obscurity with only a claustrophobic meeting box, located in the very back corner of the new West Hall, to show for its NAB presence. Others, like ThinkAnalytics and XroadMedia, continue to report relentless growth.
We have previously voiced concerns about ContentWise virtually giving its software away for free in recent years, sliding into customer accounts that have been deemed untenable by other recommendation engines, as clients in some parts of the world, with certain budgets and unique targets, appear to pay less per successful recommendation served or percentage point of churn reduced.
We believe there is an element of complacency creeping into this sector from the service provider perspective, under the misconception that something as seemingly simple as searching for content is solved – when the reality is that discovering content is a headache that has only just begun.
Meanwhile, 24i Media was busy boasting of its recent acquisition of UK-based recommendations firm The Filter, away from the show floor from a private hotel suite in Las Vegas. We learned that The Filter’s brand will be killed off to become part of the burgeoning 24i, which is owned by Aferian, but with its own sales channel, we understand.
24i’s CEO, Dr. Neale Foster, alongside The Filter’s CEO, Damien Read (now 24i’s SVP of Data Products), believe this market is not slowing down.
“Every product manager wants personalization,” they declared. Read is particularly enthusiastic about Dr. Foster’s vision for pushing the team to think bigger – with The Filter’s technology becoming part of a more multi-faceted churn-focused analytics product.
No cuts have been made following the takeover, we are assured, but instead the in-house data analytics team is expanding fast. At the same time, there is a roadmap towards automation where fewer teams of people are required, but the consensus seems to be that this is about reducing the need for video customers to have in-house data teams, providing data analytics and personalization as a service.
The same is true of our trips to Gracenote and ThinkAnalytics at NAB. The former being acquired by Nielsen in 2017 has produced some eyebrow-raising product developments, while the latter has been the subject of its own selling-up rumors on the grapevine, fueled by recent activity such as The Filter deal.
Everyone has a price, of course, but – as it stands – ThinkAnalytics is not in the process of a sale. The recommendations market frontrunner was also flexing its churn-busting and eyeball-boosting muscles at NAB, claiming that one especially large US customer has successfully increased views by 300% using ThinkAnalytics’ search and discovery software.
Think’s founding Glaswegian duo, CEO Eddie Young and CTO Peter Docherty, mention that some customers are using in-house search tools that are not fit for purpose (some slightly more colorful language may have been used). This echoes a common theme about the greatest competition coming from in-house teams, not from the competition.
We note that XroadMedia recently published figures that place it closer to ThinkAnalytics in terms of subscribers than we would have guessed, at 250 million across 50 customers compared to Think’s 400 million from 80 deployments. Young and Docherty are familiarly unfazed, reiterating that Think has the whole value understanding of this marketplace – from actual viewing data, to content, UI, and advertising.
Guards are dropped when we regurgitate XroadMedia’s claims that no rivals can match its software suite for minimal integration and development effort required for PoCs. “It has minimum integration effort because it is has minimal functionality,” declares Young.
Part of the elation about being back in-person is that such quotes encapsulating competition in business, with reactive honesty, are harder to come by through the medium of Zoom.
Being NAB, ThinkAdvertising is generating positive leads at the show. This has been a long time coming, with ThinkAnalytics first exploring targeted ads back in 2009. Only recently, however, has the company started to see the pieces fall into place for its first-party-driven ThinkAdvertising suite – blending first-party behavioral data with a broadened set of enriched metadata including financial and census data.
The big selling point is about building rich user profiles of great value to TV advertisers seeking to target highly segmented audience niches that they can measure. Young and Docherty reveal that ThinkAdvertising has been shown to increase ad revenues by between 1% to 5% – which is huge in the context of advertising powerhouses.
With the rise of ad-funded services, and Think’s existing established billing relationships with many of the world’s largest operators, as well as its long experience in advertising, it is difficult to see how ThinkAdvertising cannot be anything other than a success story. However, we reserve caution that the world of advertising is turbulent and unpredictable.
Last but not least on our NAB personalization and recommendations rollercoaster, the Gracenote team tells us of tremendous room for expansion in discovery within a broader battle for information.
The content metadata specialist has recently waded into content analytics, with its new Audience Predict feature getting some good early traction.
Audience Predict uses a combination of Gracenote’s own metadata expertise, combined with the audience measurement data and machine learning capabilities of its parent company. The aim of Gracenote Audience Predict is to maximize return on investments made in content programming – by equipping content owners and distributors with insights to make strategic decisions.
An episode of SpongeBob Square Pants would not be a good fit for ESPN, is the basic example given of Audience Predict’s matchmaking capabilities. A large group of data scientists have been working on Audience Predict for a long time, we are told, so clearly there are more difficult decisions ahead than this example. Besides, this is more about long-term predictions, not timelines of up to 6 months, which are much harder to nail an accurate prediction for.
With Gracenote claiming that it dominates the metadata discovery market with something like 70% share in most of the world’s top 80 economies, we have to question how much of an opportunity this represents for others wanting a slice of the pie? What we are seeing is that personalization software specialists are finding a sweet spot between Gracenote and clients – building on top of this metadata with fuller analytics packages, with more automated tools and AI-based data training models.