Publish your content on The Drum

The Starbucks data strategy that tempted customers back in-store for a Pumpkin Spice Latte

With the UK entering a second lockdown, in-store and out-of-home campaigns were no longer effective in reaching customers

MiQ and Havas Media Group won the ‘Most Effective Use of Data’ category at The Drum Awards for the Digital Industries 2021. Here, the team behind the winning entry reveal the secrets of this successful project...

The challenge

In September 2020, Starbucks set a commercial aim to return to pre-Covid sales by March 2021. Pumpkin Spice Latte season was fast approaching and the business chose this seasonal moment to achieve short-term commercial success to kick-start the longer term objective.

With the UK entering a second lockdown, in-store and out-of-home campaigns were no longer effective in reaching customers. Starbucks briefed media agency partner Havas and programmatic partner MiQ to identify its ‘at-home’ audience and reach them online.

The strategy

The big idea and creativity in approach to this challenge was inspired by leveraging new-to-market anonymized mobile data, exclusive to MiQ. Identifying who Starbucks’ new and potential customers were and where they were located would be the game-changer.

This rich mobile data from 20m+ UK handsets included locations, interests, demography and app usage.

MiQ’s data scientists then leveraged these insights to identify audience clusters at postcode level and built highly targeted audience segments for groundbreaking targeting.

Using this data, audience clusters were built by integrating Starbucks store postcodes with anonymized mobile data. Layering on MiQ’s existing datasets and expertise enabled the build of high converting audience segments in close proximity of a Starbucks store including: medium to heavy Starbucks app users (hero buyers), light Starbucks app users and Starbucks’ competitor app users.

This innovative research and groundwork to identify Starbucks’ highest converting audiences uncovered the over-indexed audience areas and shifted targeting away from blanket awareness.

The execution

MiQ then mapped the three audiences built using the anonymized mobile data to identify opportunities for the greatest engagement in the areas where the share of voice was below average for Starbucks’ app usage. To ensure the highest performance from the targeting, and higher than Starbucks’ TV audiences, MiQ removed postcodes where Starbucks was not easily accessible either by foot or within a 10 minute drive.

Finally, MiQ then targeted around the radius of stores where the audiences were identified to index highly for coffee drinkers or coffee app users.

MiQ then surveyed the audience through a YouGov panel to measure the uplift of the campaign and anonymized mobile data to measure the increase in share of voice in the geo locations the campaign was delivered.

In addition, MiQ mapped the audience data back in an addressable way to its media buys by indexing the top 1,000 publisher domains the target audience was using. Creating an overlap of the top 1,000 domains that Starbucks’ target audience profile were using meant MiQ was even better positioned to promote Pumpkin Spice Lattes to the Starbucks’ app users on the specific domains they were active on at the right time of day.

Once indexed, MiQ’s data scientists and data analysts built out custom modelling and custom bidding scripts. This ensured MiQ’s campaign planning optimized spend in the most effective way, with accuracy and at scale – ultimately driving campaign success. MiQ’s specialist data team calculated a unique probability for the likelihood of each ad impression available belonging to a user who is a Starbucks’ app user. This resulted in 168,000 rows of data – derived from 1,000 domains, seven days a week, 24 hours a day – giving the value for any given impression by the hour. By syncing this data, MiQ’s trading team, powered by proprietary tech, used this data to inform bidding on any given impressions.

The results

Campaign effectiveness was measured by adding a pixel to campaign creatives to build an exposed user group, matching this group to a YouGov panel to survey them and to a control group of panelists to compare campaign exposure uplift. MiQ secured positive uplift across all brand metrics measured.

  • +17% intent to visit a store in the next three months when exposed to campaign.

  • +4% intent to purchase a Pumpkin Spice Latte after exposure to campaign.

  • +62% over-delivery on Pumpkin Spice Latte sales versus business targets during campaign period.

“Achieving uplift for a global brand that’s already so well-known is difficult; to achieve positive results across all the metrics was a significant win for Starbucks thanks to the sophistication of MiQ’s audience buying and activation. The unique capability of MiQ to utilize mobile handset, location and demographic data signals to reach our audience across multiple platforms had a direct impact on the success of this campaign.” Alex Boniface, client partner, Havas Media Group London

This project was a winner at The Drum Awards for the Digital Industries 2021. Find out which of The Drum Awards are currently open for entry.

By continuing to use The Drum, I accept the use of cookies as per The Drum's privacy policy