Location Data and Machine Learning Form the Hilt of New-Age Cross-Screen Advertising Attribution

Location Data and Machine Learning Form the Hilt of New-Age Cross-Screen Advertising Attribution

Interview with TVadSync and Unacast

Walle and McNichols Said That By Taking the Human Element out of the Attribution Model, the Machine Learning-Based Approach Was the Only Way to Measure Cross-Channel Media

Is TV advertising industry truly entitled? We keep hearing that TV advertising is not up to the mark in delivering results proportional to the media budget. A part of the problem lies in the gaps in TV performance analytics. The latest partnership between TVadSync and Unacast changes this dynamic, combining cutting-edge cross-screen media and transparent location data to measure footfall traffic and help advertisers fully quantify the impact of their combined TV and digital advertising. To better understand how TV advertising is set to change the multi-touch attribution model for advertising with data and analytics, we spoke to TVadSync’s Head of Product, John McNicholas and Unacast CEO, Thomas Walle.

What is the state of location data in cross-screen advertising? How do you enable customers to benefit from location data technology?

Thomas Walle: The use of location data in cross-screen advertising is picking up steam – after several years of “cord-cutting” left advertisers uncertain of where their spend is really going, a partnership like the one between TVadSync and Unacast provides the kind of full circle customer journey measurement that advertisers need. With location data assisting in the ability to quantify the combined effect of digital and TV advertising, advertisers can be confident in their campaign metrics, once again bringing the TV to the forefront of their minds.

At Unacast, we’re excited about the possibilities of location data with the TV space – in the past, although advertisers could get a good idea of how many sets of eyes were seeing a commercial, they couldn’t connect that directly with in-store traffic. Location data makes that possible.

Would you define ‘transparency’ in location data optimization? How does Unacast deliver committed transparency?

Thomas Walle: We believe that transparent location data is the foundation of building better products and making smarter decisions. If you can’t trust your foundation – or worse, your foundation is shaky and built with questionable construction – you’ll have a hard time building anything worthwhile on top of it. That’s why transparency in datasets is so important.

At Unacast, we define transparency as a clear understanding of the origins of any given location data set and which attributes are included in the data. We’re committed to delivering radical transparency. That means we provide the source IDs, latitude, and longitude, dwell time, timestamp, venue, category, and the number of nearby venues in our datasets, so partners can feel confident about the authenticity of the data.

What are the key takeaways from your recent partnership with Unacast and how does it benefit your adtech customers?

 John McNicholas: The partnership helps prove out how effective TV and digital are –when combined, at driving the store and location visits. It offers total clarity for advertisers as to which channels and creatives provide the best ROI in terms of foot-fall, and it forges a real, attributable connection between upper and lower funnel activity, re-connecting the marketing funnel from end-to-end.

What is your roadmap to build readily-adaptable location data adtech platforms? How do you leverage/ work with audience data platforms?

Thomas: Unacast provides transparent, contextualized location data that helps marketers build out their audiences – in essence, a vital tool in the targeting and attribution toolkit. Our team of engineers and data scientists consolidates billions of disconnected data points and raw signals, vetting and analyzing each point before grouping it with like points and ultimately delivering a Visit – a clear understanding of a device’s activity, location and time spent at that location, among other attributes. Armed with this information, marketers can more easily and accurately build out audiences.

In the GDPR era, what changes to location data strategy have you made? How does the post-GDPR impact the ecosystem?

Thomas: As a company founded in Norway, where the transparent sharing of information is part of the fabric of our culture, Unacast has always worked to provide data transparency to our partners. In fact, “Trust through Transparency” is one of our core company values.

We’ve always felt the importance of an equal focus on privacy as well (that’s where the “trust” part of our company value comes in), which meant that GDPR didn’t necessitate any major changes to our strategy or the principles that govern our business. That said, although we currently collect data in North America only, we plan to expand into Europe soon, and still made the choice to become GDPR-compliant.

We believe that GDPR can and should have a positive impact on the location data ecosystem – wiping out smaller players that don’t treat privacy and transparency in the right way and leading to higher-quality data, even if we see a temporary decrease in quantity.

How do you work with AI/Machine learning in making Location Data-based TV advertising and adtech personalization more effective?

John: TVadSync’s Multi-Touch Attribution model uses machine learning to correctly attribute performance value to cross-channel media. The attribution model is not trained with any historical or market data. It is totally blind as to whether it’s viewing a digital ad or a TV ad, and so no bias exists. It learns from mapping millions of touch-points across thousands of user journeys across TV and Digital media and calculating using a probabilistic method, which channel/placement/creative, or TV spot were most influential at driving conversions.

Traditional attribution models were based on pre-determined, often position based biases which are prone to human error, and positively skew lower funnel advertising such as digital media, leading to inaccurate attribution and huge optimization inefficiencies. By taking the human element out of the attribution model, the machine learning-based approach is the only way to measure cross-channel media and is providing huge gains to TVadSync’s clients.

What are the opportunities and risks you foresee in the way location data is shaping for 2020-2025? How do you prepare for these disruptions?

Thomas: The focus of location data is moving away from individuals and more toward trends-based data. There is a major opportunity there to take location data from micro to macro, opening up the industry to solve big picture challenges and better plan for the future of the human movement.

We’re already working on that shift in data collection and analysis. It’s also an exciting time to move from tracking the 30% of time people spend online, which up until recently has been the main focus of marketing technology, to tracking the 70% of the time they spend in the real world through location data. Moving forward with that in mind, it will be key for companies within our space will need to dig deeper into client needs to make sure we’re able to deliver products that help them achieve their goals.

Thank you, Thomas and John, for chatting with us about your recent partnership and the impact it could have on the TV advertising industry. 
Picture of Sudipto Ghosh

Sudipto Ghosh

Sudipto Ghosh is a former Director of Content at iTech Series.

You Might Also Like