Outcomes are in vogue with broadcasters at the moment. From ‘Project Lantern’ in the UK, a joint project between ITV, Sky, and Channel 4 aimed at demonstrating TV’s immediate sales impact, to a slew of recent shoppable ad launches in the US, TV businesses on both sides of the Atlantic are investing in outcome-friendly formats and capabilities.
But broadcasters often stop short of talking extensively about performance. While performance has always had a role to play in TV, brand building remains TV’s bread and butter. Broadcasters often feel their high quality environments and tuned-in viewer experience gives them brand building capabilities which tech platforms struggle to emulate. On the flip side, competing with the performance offerings of the data-fuelled, mass reach tech offerings can be a daunting prospect, offering unfavourable comparisons.
French broadcaster TF1, however, is firmly setting its stall behind its performance capabilities. Two months ago, the company announced its ambition for its streaming service TF1+ to become advertisers’ “go-to full-funnel marketing platform”, unveiling a slew of new ad tools to help it fulfil this brief. François-Xavier Pierrel, chief data and adtech officer at TF1 Group, says this isn’t just marketing speak – the broadcaster has made significant changes to its streaming offering in order to cater to advertisers’ performance objectives, helping it compete for different budgets from a wider variety of brands.
For TF1, the push into performance is partly driven by necessity. The reality for TV businesses is that lots of brands – even large multinationals which have weighted TV heavily in their media plans for decades – are increasingly looking to drive short-term results. “It’s well known that for lots of big brands like L’Oréal and Renault, the branding part is shrinking like hell,” said Pierrel. “They all want to know how many people you can drive into their stores.”
But it’s also driven by opportunity. The launch of TF1+ earlier this year, replacing predecessor MyTF1, changed TF1’s streaming service from essentially a catch-up hub to a destination in its own right. This in turn transformed the platform into much more of a data-driven product, while also opening up space for new formats – the two pillars of TF1’s performance strategy.
Lowering barriers to data activation
On the data front, Pierrel says that the shift away from catch-up required the broadcaster to rework its data strategy. “When you’re just a replay platform, data is interesting, but it’s not at the centre of the user experience as a streamer,” he said. “Algorithms aren’t at the heart of your proposition, and on the other side, you’ve got traditional segmentation for media activation.” For a destination streaming service however, users expect personalised content recommendations and a tailored user experience. That requires an identity infrastructure, which can then also be used to deliver personalised, performance-driving advertising.
At the heart of TF1+’s data strategy is ‘Graph:ID’, an audience graph developed in partnership with Snowflake. The graph combines data collected at the point of sign-up (such as name, post code, and birthday) with viewing data, all pegged to a stable identifier (usually their email address). This is key for content recommendation, and to personalise TF1+’s service. But it provides a strong bank of advertiser-friendly first-party data – TF1 says it has around 25 million user profiles signed up to the service. This is enhanced through data integrations with retail and data partners including Infinity Advertising, Unlimitail, and Cdiscount.
TF1’s data offering has been further strengthened by a new partnership with LiveRamp-owned data collaboration technology Habu, allowing advertisers to use their own data to inform and measure campaigns on TF1+. Data collaboration isn’t new to TF1, but Pierrel says that previously these integrations were previously quite onerous on brands, and required them to have mature first-party data capabilities. This limited the number of partners these capabilities were realistically open to.
The Habu partnership has lowered these barriers to entry, and seen the launch of a self-serve system. “Now, you can come with a CSV file, or a cloud, or a bucket, we don’t care,” said Pierrel. “We can connect, ingest your data, match it in a very safe way with ours, and build up learnings and segments.”
“We’re trying to make sure that data collaboration is not limited to premium advertisers, otherwise we’ll end up with the same usual suspects who have the money, the teams, and the know-how – I’m not sure how we build growth there,” he added. “We need to explore collaboration with brands who are less mature, but they still have their CRM systems and their conversion data. That’s why we were so interested in Habu – we wanted to make it simpler for more companies to come into our system, match their data and explore what we can do together.”
Moving beyond pre-roll
Data alone however isn’t enough – new formats are also key to TF1+’s performance proposition. MyTF1 was limited to pre-roll and mid-roll ads, and as Pierrel puts it, “if you only do pre-roll, you can’t claim you do performance”.
In September, TF1 brought on Magnite as a second ad server alongside FreeWheel, which has enabled it to develop and run a wider range of ad formats, including display ads, shoppable ads, and playable ads. These formats better suit advertisers with pure lower-funnel objectives, who usually would spend with either YouTube or Meta.
Pierrel said the challenge for TF1+ is to change the way clients plan campaigns with the broadcaster. Brands will often run a much larger number of performance campaigns across the year on social platforms, but are more used to running a few big brand campaigns on TV.
“We need to make sure we can answer each type of objective across the year, and that means we’re not running five campaigns for a client, but twenty campaigns,” said Pierrel. “There might be a lower budget per campaign, but at the end of the year it’s more revenue because you’re answering more of their challenges. And this is where new formats are really necessary. We need to find ways to push-to-mobile, to run discount campaigns, to have QR codes – those things help us become much better at addressing every step of a client’s campaigns and their objectives.”
Meeting expectations
TF1’s ambition is that all advertising run through TF1+ will be data-driven (with the caveat that users have to consent to use of their data through Graph:ID – Pierrel believes that providing a high quality data-driven experience is key for achieving high opt-in rates). That doesn’t mean that all campaigns delivered through TF1+ will be highly targeted. Brand building of course remains a strength for TF1, and is obviously a core part of any full-funnel offering. “We aim to be data-driven, but that doesn’t mean you need to target everything. But you do need some data, to report back on the quality you’re delivering,” said Pierrel.
While targeting will be a big part to TF1+’s performance offering, pre-campaign insights and measurement are also crucial.
“Where we weren’t so good in the past was the preliminary phase of campaign planning, the insights,” said Pierrel. “What do you know about my target audience, about the customers I want to reach? Now, with Habu, we can expose our Graph:ID to advertisers, so they can explore what we know about our audiences. And if they want to, they can bring their own data on board in a much smoother way. We are now better with insights, better with activation, and better in measurement thanks to the retail partnerships we have.”
One concern for broadcasters when investing in performance capabilities is whether they can match up to the metrics offered by the big tech platforms. It’s one thing to deliver short-term results, but another to be able to prove it – something which the tech giants are adept at.
Pierrel says that’s a reality that TF1+ has to deal with. “We have to be better at reporting results,” he said. “If you want to do performance, what advertisers ask you is ‘how many sales did you drive, how many people did you bring to my store?’ If you can’t deliver on those types of KPIs, you can’t claim you are good at performance. So we’re working with our data team to deliver that.”
Another reality of performance advertising is that sometimes results will fall short of expectations. Here, Pierrel says TF1+ must have a “learning agenda” with its clients.
“Sometimes we’ll perform very well, and like any advertising business in the world, sometimes results will be disappointing. So we have to learn from every campaign – what worked very well and what didn’t work, and make sure we’re getting better all the time. That means that data should have a role in meetings between our customers and our sales teams, so they can learn and build better campaigns together.”
