Sky Media brings outcomes ‘certainty’ to addressable TV planning
Sky Media has launched Norman, a databank of 2,400 addressable TV advertising campaigns aired since 2020, spanning 535 advertisers, which can be used to improve pre-campaign planning and execution.
The comprehensive databank has been analysed to understand the characteristics of the best performing addressable campaigns run through Sky Media, and the worst performing.
Norman will become a significant tool for forecasting brand and business outcomes and optimising campaigns inflight to maximise these.
Analysis of Norman has shown, for example, that a small increase in the number of targeting attributes used has a disproportionate impact on campaign effectiveness.
The top performing addressable TV branding campaigns use an average of 3.4 targeting attributes, with mid-performers using 3.2 on average and the lowest performers using 2.7, for example.
Confirming the fine margins involved, the most successful performance campaigns (judged by incremental website visits) use an average of 3.5 attributes, with mid-performers at 3.2 and lower performers at 3.0.
The use of first-party advertiser data (matched to Sky Media data) is the biggest campaign accelerator.
Campaigns leveraging first-party data drive up to 8x higher sales uplift and 6x higher incremental visits than addressable campaigns without first-party data.
Sky Media unveiled Norman this week and simultaneously published a study underpinned by its data, called ‘Addressable Advantage’.
Paul Dyson is co-founder at effectiveness consultancy Accelero, which was tasked with unravelling the Norman data to show effectiveness drivers and help build a model for outcomes predictions.
First-party data boost
He observed: “First-party data gives a significant improvement in effectiveness for those advertisers that use it. It is a key driver of effectiveness.”
Many brands that have first-party data have only used it for 30% of their campaigns. Dyson found that if they had used it for all their addressable campaigns, they would have achieved a 25% uplift in incremental website visits.
Analysis of the Norman databank shows that addressable TV campaign performance in Week 1 is a strong indicator of how the whole campaign will perform, so Sky Media is focusing on detailed Week 1 tracking and inflight optimisations to drive total campaign effectiveness.
Dyson noted: “Within the next few months it will be possible to change that campaign in almost real-time.”
Inflight optimisation is a key benefit from using the tool. The 2,400 campaigns in Norman have been plotted, and average performance demonstrated.
This provides a benchmark to work against, showing whether an addressable campaign is over or under-performing, based on category, and how long the campaign has been running, for example.
The Addressable Advantage study provides more evidence that addressable TV advertising works.
Compared with linear TV alone, addressable campaigns deliver a 2.1x uplift in awareness, with modelling showing a +121% increase in aided ad recall. Addressable campaigns also provide a 3x uplift in incremental web traffic vs linear alone.
Matt Hill, director of insight & effectiveness at Sky Media (pictured above at The Future of TV Advertising Global) says Addressable Advantage fills a key knowledge gap on the efficacy of addressable TV for small-medium businesses through to major household brands.
He added: “Norman will enable Sky Media to provide advertisers greater certainty around TV’s outcomes, helping TV compete more effectively for lower funnel or data‑driven budgets.
Creating outcomes certainty
“Lack of certainty has been one of the key barriers to spending on television. It can feel like a punt in the dark – you are not sure what the outcomes will be for your business or how to tell that to your finance director.
“Norman moves TV into a new place. We have the power to predict, with more certainty, the impact a TV campaign will have.
“Norman also shows us the levers you can pull to improve effectiveness.”
This last point is key to how Norman will be used to boost addressable TV effectiveness in future.
Accelero was commissioned by Sky Media to analyse all the campaigns in Norman and identify the factors most strongly associated with campaign success.
Criteria like category and business size, and variable campaign elements like budget, frequency and targeting strategy, were assessed. Elements already mentioned, like use of first-party data and the number of targeting attributes, were also analysed as contributors to success (or lack of).
Dyson explained: “If you are an automotive brand, Norman can tell you whether it would be better to use current car owners or people who are due to change car soon, or first-party data, as your chosen targeting attributes.
“Is it better to use addressable TV only, or combine that with addressable VOD? That is the kind of question it answers. It helps with decisions when designing your campaign.”
Eunice Demetrius, insight manager at Sky Media, revealed that a theme park brand originally used Mastercard targeting (to only those who had spent on theme parks previously) then added Age and Mosaic attributes. Those adjustments led to a 160% increase in campaign efficiency.
Mazda added automotive targeting to its TV campaigns and saw a 27% increase in efficiency, and more than 4,000 additional website visits driven by TV activity (vs an unexposed control group).
“Small targeting enhancements can have a big impact” Demetrius noted.
Marginal gains
“This is all about marginal gains,” Hill confirmed.
The initial focus for Sky Media, when using Norman, is addressable TV campaign outcomes forecasting and optimisation. However, this databank will be used to understand and improve sponsorship and linear campaigns, too.
The Norman to-do list includes a deeper analysis on how linear and addressable campaigns work together. Creative has also been analysed and the databank will become a tool for creative effectiveness forecasting and optimisation, too.
Norman is focused on brand outcomes (e.g. brand uplift), by harnessing post-exposure surveys, and business outcomes. Incremental website visits are the current focus for the latter.
The databank includes large and medium brands and small-medium businesses – with large brands being the largest group.
Notably, Norman contains the highest and lowest performing campaigns, including some that achieved zero uplift from their addressable campaign (compared to the unexposed control group).
That differentiates this dataset from the IPA Effectiveness Databank, which harnesses awards data and so skews to successful campaigns.
Norman is a living databank, and is being fed 250 new client studies a year.
The Norman data is derived from household-level outcomes, harnessing Sky Media’s 5 million home panel for exposure data. Specialist web attribution tools correlate exposure to traffic.
Hill emphasised that the non-exposed control groups used in the effectiveness studies that populate Norman are a good representation of the exposed audience. They share the same demographics and behavioural characteristics as an exposed group, for example.
“This is rigorous,” he declares. “By comparing the brand or web outcomes between the two groups we can confidently measure the incremental impact of being exposed to the campaign. It is true AB testing.”
Dyson confirmed: “Norman is a really strong database, with a very robust sample size so you can build robust models around the data. You can cut into the data without compromising that sample size, too.”
Free tool for brands
Norman now becomes a standard tool that Sky Media will use with agencies and clients to maximise the effectiveness of an addressable TV campaign. That analysis is free.
Sky is already a member of Lantern, the UK broadcaster initiative to demonstrate how TV drives online responses and actions. That is ‘Powered by Thinkbox’ – the advocacy body for ad-funded TV – and also backed by ITV and Channel 4.
Norman is not part of Lantern, and today is a proprietary Sky Media tool.
Returning to the Addressable Advantage study, Hill says it shows that every advertiser should be considering addressable TV in their media plan.
Demetrius stated that addressable TV is suitable for all advertisers, regardless of their size, objective or data maturity.
“You can run an addressable TV campaign for as little as £3,000,” she revealed. “We had a leisure brand that spent £8,000 and achieved a 41% uplift in website visits [vs an unexposed control group].”
Hill is adamant that TV buying should still include linear, and that addressable spend growth will come from new-to-TV budget.
“It is not about dumping emotive, brand-building reach campaigns in favour of addressable TV or hyper-targeting.
Instead, “Buyers should consider the potential for television to work at the bottom of the funnel.”
He believes it could be beneficial for some advertisers to switch money from social media as it nears the top of the diminishing returns curve, when new spend is not working as hard.
“We would advise using that money to test addressable TV, which works in a different way to drive low funnel results.
“This is where extra money [for addressable] could come from – not from the linear campaign.”