The beating heart of GAP’s strategy to personalise its marketing efforts comes in the form of a proprietary customer data platform (CDP), which allows it to create a single view of its customer base. Key to personalisation is the art of segmenting its first-party data and balancing its lower and upper funnel.

In addition to this, GAP augments its first-party data with third-party data such as demographics, interests and life stages to further segment audiences before syndicating them to advertising platforms like Google and Facebook. This approach encourages scale and makes its data segmentation actionable across various activation channels. Proper analysis of customers' intentions allow GAP to achieve much better math rates and higher degrees of accuracy. Rather than serving an irrelevant and possibly confusing message, this approach allows it to adapt its marketing strategy in real-time and change messages accordingly, depending on where in the funnel the customer is.

Painting a complete picture of audiences

But past behaviour and customers’ lifecycle stage are just part of the story. By overlaying first-party data with third-party data segments, GAP’s data science team can predict propensities and behaviours before a campaign even starts. These hybrid audiences are an excellent seed to build prospecting audiences from, and to reach prospects much more efficiently.

But what happens once the first step is done? Where can we go from here? Once audiences are created, the next step is to build corresponding strategies and a library of assets, which would include messages and creatives. When the campaign launches, the results are treated with machine learning algorithms to determine the winning combination.

Using AI to optimise communications

On top of that, GAP uses Persado - an artificial intelligence (AI) solution that optimises text copy against a customer’s preferred emotional sentiment or ‘voice’. Initially, it was used for email subject lines, and to then expand the use of this tool to other channels, including personalised messaging in select use cases on its own websites.

With such a large data set at its fingertips, GAP’s manual segmentation would be nothing less than a complete nightmare, so it uses AI and automation to help with customer resolution, segmentation and clustering. Going down this route allows for much higher granularity when building and testing audiences, but there’s also the risk of over-targeting – when too much targeting yields to diminishing returns. This is handled with meticulous planning, hypothesis development, and testing.

In short, this activity can be grouped into two main steps: examine the results of targeting efforts and refine strategy. If there isn’t any meaningful lift from an audience sub-segment, it would be folded back into a broader audience. This process is 'rinsed and repeated' until the expected results are identified.