How using Synthetic Data increased Statflo’s customer targeting accuracy and ultimate sales by 16%
Hypertargeting tablet users with maximized model accuracy
Company Profile
- Founded in 2012
- Employs over 80 professionals
- Serves over 150 premier enterprise-grade customers
- Manages 3 million plus monthly customer interactions
Summary
Statflo is a leading Canadian startup specializing in developing compliant one-to-one business text messaging technology, enabling businesses to have meaningful, two-way conversations with their customers. With no Personal Identifiable Information (PII) required from Statflo, Arima can consolidate both parties’ relevant data and allow Statflo to drastically improve the performance of their customer propensity models, which ultimately increases sales by 16%.
Statflo works with Arima to segment and understand the targeted audience at the most granular level to deliver an effective texting campaign.
Results
- Improved customer segmentation and model accuracy to achieve hyper-targeting and drive texting campaign ROI
- Sales Boost: 16%
- F1 Score Increase: 2–6 Times
- Privacy Compliant: 100%
The Background
Statflo aims to cross-sell tablets to existing customers while ensuring it doesn’t hinder the primary objective of plan renewal. Given limited customer touchpoints, accurately identifying high-propensity buyers is crucial.
The Challenge
Statflo is committed to developing a ML model to discover high potential targets who should be included in campaigns for the cross-selling of a consumer electronic product. With limited first-party data, Statflo requires a strong and credible third-party database to make the model more accurate and explainable for better campaign performance.
The Approach
- Arima joins the client’s 1st party data, the customer wireless usage, with the Synthetic Society database.
- From a list of 4,000+ consumer attributes, Arima brings a list of 20 variables with the highest correlation to tablet purchases, including income, number of kids, housing type (rent/own, condo/house) and buyer psychology, to help Statflo understand customers’ motivation, attitudes and behaviours leading to purchases.
- Statflo develops ML models with rich and valuable insights into the audiences and crafts the campaign accordingly with confidence.
- Statflo doesn’t need to release customer informations to Arima. With no Personal Identifiable Information (PII) needed, this approach is 100% privacy compliant.
The Sucess
Wearables XGBoost Tablet XGBoost
- Including data from the Synthetic Society by Arima significantly improved Statflo’s model performance and resulted in 16% sales increase for their client.
- Arima helped Statflo improve the model’s explainability and identifying correlations between variables and final sales. We discovered, for instance, that people with children and homeowners are more likely to purchase tablets.
- With no PII required, Statflo doesn’t need to worry about violating privacy policies while being able to seamlessly integrate useful data points from Arima.
“Arima’s Synthetic Society database provides us with credible and impactful data signals and continually helps us to increase model accuracy so that we can give our clients better predictive guidance driving strong business results.”
— Doug Creighton, Data Science & Business Systems Lead
Book a demo today and learn how you can harness the best source of 3rd party data and build better models with Arima.
