Home › Resources › Case Studies ›
case study
RoofID: Using AI and aerial property imagery to target media more effectively.
Project Description
By training an AI model to identify degraded roofs using freely available satellite imagery of Auckland, Together developed an automated end to end targeting solution for COLORSTEEL® that identifies properties needing a new roof and turns these priority locations into media buying audiences.
KEY Outcomes
Quantification of COLORSTEEL® target market
Prioritisation and management of geo-targeted demand generation activity
Improved demand generation activity ROMI
The Challenge
COLORSTEEL® tasked us with increasing the volume and efficiency of their digital lead generation in Auckland. Most re-roofing happens when people are renovating. However with strong economic headwinds and high interest rates, renovation rates were falling. Our strategy was to identify and target the roofs that most needed replacing - the challenge was how to do this and do it cost effectively.
The Solution
We developed a unique solution, dubbed Roof ID, that had three primary components:
- Sophisticated geo-spatial analytics to identify and clip the images of roofs from freely available satellite imagery, associate it with location data and create an input data set.
- An image recognition AI model, trained in-house, to score the level of degradation of each roof in the data set.
- Location based targeting and ad delivery system that allowed us to target advertising content within a 10m radius (essentially to individual property level).
The Result
Roof ID identifies and reaches an audience with a qualified need, at scale. By generating lists of locations that need new roofs, Roof ID has also provided Colorsteel with a precise understanding of where to focus marketing and on-the-ground efforts.
