Interviews, insight & analysis on digital media & marketing

Why contextual data is the key to effective ad targeting in a post-cookie world

By Carl Carter, Marketing Strategy & Effectiveness Director, IRI

For obvious reasons, people did more of their Christmas shopping online this year than ever before. Advertisers saw this time as an opportunity to claw back revenue lost over 2020, which has made the landscape exceptionally competitive, with ads at risk of being lost amongst the noise.

At the same time, the ecommerce experience remains blighted by the same old issues: we’re still interrupted by ‘accept all cookies’ pop-ups, and followed around by ads for things we’ve already bought.

This festive season highlighted how important effective and efficient targeting is – and how unfit for purpose third-party cookies are in an age of privacy and trust. They’re being phased out, which is a positive step, but the annoyance, brand damage and wasted ad spend created by this once-valuable tool suggest it’s time to retire.

With a post-cookie world on the horizon, and as we strive to recover from the rigours of 2020, the pressure on advertising spend is huge. Brands can’t risk missing the mark, alienating potential customers, or targeting audiences that have already been saturated.

To ensure they serve relevant content to the right people at the right time, advertisers and media owners must find a new approach to reaching and influencing target consumers with the optimal mix of channels, while upholding the need to protect personal information.

Many have staked their hopes on the concept of first-party identity resolution.

Finding growth beyond the cookie

ID resolution involves taking first-party data, such as an email or house address – which the data subject has given their permission to be used – and matching that person across multiple partners or publishers to create a single, omnichannel view of the individual.

First-party data has its advantages. It’s permissioned, which ticks the privacy box. The high levels of quality and detail it can provide opens up the doors to a direct, one-to-one personalised relationship with the consumer, built on a basis of trust. It also supports measurement of the impact of advertising through closed loop measurement techniques.

However, it also has its limitations – most notably the difficulty of obtaining the scale and reach required to drive incremental sales. The less data you have, the fewer the matches you get; and seed data could just be an email address. Even in strong cases you might only be able to deliver a match rate of 20% across your first-party dataset. In addition, with media plans typically consisting of 10 or more partners, the need to connect all the separate datasets, and put in place commercial agreements for each party, is a challenge.

Layering data to provide accuracy, relevance and scale

On its own, first-party data cannot resolve all the challenges ahead. However, when it’s combined with contextual data and predictive modelling, it becomes the foundation for a unified dataset that enables advertisers to identify clear, untapped sales opportunities.

The approach strikes a balance between effectiveness – driving high volume of sales, and efficiency – minimising wastage and achieving strong ROI.

It starts with the high-quality seed data on people who are already buying from the brand, which provides the ‘accuracy’ part of the equation. Contextual data is then layered on top, including sales data, and details on the consumer such as location, household, purchasing behaviour and values.

Scale and reach are then built through predictive modelling, to reveal similar ‘look-alike’ consumers who have something in common with those who’ve already bought, and who might also be likely to purchase, along with details such as where they might buy.

This provides a brand with a granular view of where to spend its media budget, identifying high opportunity areas for driving significant penetration and sales, as well as low opportunity areas where, for example, sales are saturated, and therefore advertising would be wasted. There’s no infringement on consumers’ privacy, because the layered approach highlights cohorts of people to target, rather than individuals.

Advertisers and media planners should be working right now to ensure they have less need to rely on cookies to target audiences. Blending first-party data with contextual data and predictive modelling allows them to understand customer journeys and grow their audience, while maximising media spend – and all without creating any trust issues.

Alongside driving high volume sales with high efficiency, this more respectful approach to targeting will result in stronger open relationships between the consumer and the brand.