Revenue-generating models, enabled by Predactiv data
Lift in models compared to a random data sample
Improvement in model performance by adding Predactiv data to the marketing consultancy’s models
Increase in model usage by the consultancy’s clients, driving significant revenue growth
A leading marketing consultancy had two primary objectives: to develop new models, and to enhance the performance of existing models. Both initiatives would unlock greater revenue opportunities for the consultancy.
For the first objective, Predactiv provided its proprietary, deterministic data asset as seed audiences. These audiences consisted of data assets that were otherwise non-existent at the marketing consultancy. The seeds were essentially “dependent variables” of the models, which the marketing consultancy used to build look-alike models—scoring large universes of consumers to achieve specific audiences (e.g., luxury auto brand owners).
The second objective was to improve the marketing consultancy’s existing models. The client had thousands of pre-built models in their arsenal and aimed to improve their performance by incorporating an additional data asset. In this case, Predactiv’s digital intent data was used as “independent variables” in the models to enhance lift. Even a marginal improvement would have a significant impact for the client.
All of the models—the newly built ones, as well as the models with Predactiv Data (as independent variables)—outperformed expectations.
Predactiv “seeds” were initially used in about 100 models, but after strong results, an additional 100 models were built with new “seeds” from Predactiv, thus improving their revenue opportunity significantly.
Across the additional 100 models, Predactiv drove lifts ranging from 1.5x to 4.5x compared to a random sample of data.
The below charts are an example of one model, representing luxury auto buyers:
Model vs Random Sample
With Predactiv Data
as the independent variable in the model:
Double-tap to Zoom
Model vs Random Sample
With Predactiv Data
as the independent variable in the model:
Double-tap to Zoom
The left chart above demonstrates the lift when using Predactiv data as independent variables, in combination with the client’s internal data. Predactiv Data drove a significant lift of 4.5x over a random sample.
The right chart above shows the lift with only the client’s internal data, where only a small lift was seen.
Return on Ad Spend (ROAS), nearly 5x the benchmark
Return on Ad Spend (ROAS), nearly 5x the benchmark
Cost Per Action/Conversion (CPA), 65% lower than the client’s $25 CPA goal and 83% lower than the $50 CPA average across prior campaigns
Cost Per Action/Conversion (CPA), 65% lower than the client’s $25 CPA goal and 83% lower than the $50 CPA average across prior campaigns
Booking Rate, 75% higher than Meta’s benchmark
Cost Per Booking, 20% lower than Meta’s benchmark
Booking Rate, 75% higher than Meta’s benchmark
Cost Per Booking, 20% lower than Meta’s benchmark
Be the first to know about our latest product announcements, company news, industry insights, and expert content—delivered straight to your inbox.