A devastating pandemic crippled the world, the only solution was mass immunization, and fortunately, the global science community worked wonders.

The task for Walgreens was to ensure those vaccines were injected into America’s arms as quickly and equitably as possible.

But, people were hesitant about receiving the COVID-19 vaccine and supply and geographical distribution of doses was unpredictable.

Objectives:

1. Build brand trust with Walgreens for the vaccine.

2. Drive vaccine Appointments; and

3. Drive 30M COVID-19 shots in arms by August 31st 2021

Walgreens needed a targeting solution to identify vaccine prospects effectively and efficiently.

We created a Vaccine-Readiness Model to predict individuals most likely to want a COVID-19 vaccine a week before actually seeking one out.

By dynamically identifying and scoring prospects for vaccine readiness - regardless of their previous level of Walgreens engagement – we would take into account the wide range of behavioural, cultural, and mindset factors affecting vaccine uptake across the U.S.

We would identify the degree of vaccine readiness and vaccine hesitancy and create a model that allowed us to target not just the best prospects. But also, by linking our model to different messaging and the programmatic buying ecosystem, we would move people from ‘hesitant to have a vaccine’ to ‘ready to have a vaccine.’

Other factors had to be considered, including predicting market-level vaccination fluctuations to help local optimization and strategic and efficient media allocation.

We established a detailed methodology based on machine learning to create and rank audiences by Designated Market Area according to their propensity to take action.

Our machine learning process allowed us to isolate independent demand variables across more than 35 search queries.

We clustered each DMA by likely demand to create seven groups across the U.S.

We developed a forecasting model for each cluster, leveraging a gradient boosted decision trees algorithm that explains key drivers in COVID-19 vaccination readiness.

This model was converted into media audiences, using 124 factors based on health and purchase behavior to identify audiences within each DMA with a high-affinity score.

We layered store location information and pulled in audiences within the 5-mile radius of Walgreens stores, to deliver targeted messages that varied according to vaccine readiness.

The vaccine-ready groups being asked to 'book an appointment now,' the vaccine-hesitant to 'learn more about the vaccine here'.

We’ve exceeded our vaccination goal by over 4.5M.

We've identified, targeted, and driven highly qualified consumers to book appointments, delivering over 34.5M vaccines – all in a highly efficient manner, significantly outperforming standard audiences offered by buying platforms.

Media efficiency statistics are hugely impressive. Comparing audiences generated by our Vaccine-Readiness Model with the standard third-party audiences offered by the buying platforms, our custom model drove:

+47% higher engagement rate

+43% higher appointment rate

at a 20% cheaper cost per appointment.

In fact, our model successfully drove 78% of all vaccination appointments within our programmatic buy.

We've integrated data insight, phenomenal modelling performance, and media message delivery via programmatic platforms to deliver an incredible result: millions of Americans now have protection against COVID-19, and we've done more than deliver on the objectives; we've undoubtedly saved lives as well. And that's the best result of all.